MyArxiv
Robotics 36
☆ Enhancing Underwater Navigation through Cross-Correlation-Aware Deep INS/DVL Fusion
The accurate navigation of autonomous underwater vehicles critically depends on the precision of Doppler velocity log (DVL) velocity measurements. Recent advancements in deep learning have demonstrated significant potential in improving DVL outputs by leveraging spatiotemporal dependencies across multiple sensor modalities. However, integrating these estimates into model-based filters, such as the extended Kalman filter, introduces statistical inconsistencies, most notably, cross-correlations between process and measurement noise. This paper addresses this challenge by proposing a cross-correlation-aware deep INS/DVL fusion framework. Building upon BeamsNet, a convolutional neural network designed to estimate AUV velocity using DVL and inertial data, we integrate its output into a navigation filter that explicitly accounts for the cross-correlation induced between the noise sources. This approach improves filter consistency and better reflects the underlying sensor error structure. Evaluated on two real-world underwater trajectories, the proposed method outperforms both least squares and cross-correlation-neglecting approaches in terms of state uncertainty. Notably, improvements exceed 10% in velocity and misalignment angle confidence metrics. Beyond demonstrating empirical performance, this framework provides a theoretically principled mechanism for embedding deep learning outputs within stochastic filters.
☆ Dataset and Analysis of Long-Term Skill Acquisition in Robot-Assisted Minimally Invasive Surgery
Objective: We aim to investigate long-term robotic surgical skill acquisition among surgical residents and the effects of training intervals and fatigue on performance. Methods: For six months, surgical residents participated in three training sessions once a month, surrounding a single 26-hour hospital shift. In each shift, they participated in training sessions scheduled before, during, and after the shift. In each training session, they performed three dry-lab training tasks: Ring Tower Transfer, Knot-Tying, and Suturing. We collected a comprehensive dataset, including videos synchronized with kinematic data, activity tracking, and scans of the suturing pads. Results: We collected a dataset of 972 trials performed by 18 residents of different surgical specializations. Participants demonstrated consistent performance improvement across all tasks. In addition, we found variations in between-shift learning and forgetting across metrics and tasks, and hints for possible effects of fatigue. Conclusion: The findings from our first analysis shed light on the long-term learning processes of robotic surgical skills with extended intervals and varying levels of fatigue. Significance: This study lays the groundwork for future research aimed at optimizing training protocols and enhancing AI applications in surgery, ultimately contributing to improved patient outcomes. The dataset will be made available upon acceptance of our journal submission.
comment: 12 pages, 8 figures
☆ Cooking Task Planning using LLM and Verified by Graph Network
Cooking tasks remain a challenging problem for robotics due to their complexity. Videos of people cooking are a valuable source of information for such task, but introduces a lot of variability in terms of how to translate this data to a robotic environment. This research aims to streamline this process, focusing on the task plan generation step, by using a Large Language Model (LLM)-based Task and Motion Planning (TAMP) framework to autonomously generate cooking task plans from videos with subtitles, and execute them. Conventional LLM-based task planning methods are not well-suited for interpreting the cooking video data due to uncertainty in the videos, and the risk of hallucination in its output. To address both of these problems, we explore using LLMs in combination with Functional Object-Oriented Networks (FOON), to validate the plan and provide feedback in case of failure. This combination can generate task sequences with manipulation motions that are logically correct and executable by a robot. We compare the execution of the generated plans for 5 cooking recipes from our approach against the plans generated by a few-shot LLM-only approach for a dual-arm robot setup. It could successfully execute 4 of the plans generated by our approach, whereas only 1 of the plans generated by solely using the LLM could be executed.
☆ Data-Driven Contact-Aware Control Method for Real-Time Deformable Tool Manipulation: A Case Study in the Environmental Swabbing
Deformable Object Manipulation (DOM) remains a critical challenge in robotics due to the complexities of developing suitable model-based control strategies. Deformable Tool Manipulation (DTM) further complicates this task by introducing additional uncertainties between the robot and its environment. While humans effortlessly manipulate deformable tools using touch and experience, robotic systems struggle to maintain stability and precision. To address these challenges, we present a novel State-Adaptive Koopman LQR (SA-KLQR) control framework for real-time deformable tool manipulation, demonstrated through a case study in environmental swab sampling for food safety. This method leverages Koopman operator-based control to linearize nonlinear dynamics while adapting to state-dependent variations in tool deformation and contact forces. A tactile-based feedback system dynamically estimates and regulates the swab tool's angle, contact pressure, and surface coverage, ensuring compliance with food safety standards. Additionally, a sensor-embedded contact pad monitors force distribution to mitigate tool pivoting and deformation, improving stability during dynamic interactions. Experimental results validate the SA-KLQR approach, demonstrating accurate contact angle estimation, robust trajectory tracking, and reliable force regulation. The proposed framework enhances precision, adaptability, and real-time control in deformable tool manipulation, bridging the gap between data-driven learning and optimal control in robotic interaction tasks.
comment: Submitted for Journal Review
☆ STAMICS: Splat, Track And Map with Integrated Consistency and Semantics for Dense RGB-D SLAM
Simultaneous Localization and Mapping (SLAM) is a critical task in robotics, enabling systems to autonomously navigate and understand complex environments. Current SLAM approaches predominantly rely on geometric cues for mapping and localization, but they often fail to ensure semantic consistency, particularly in dynamic or densely populated scenes. To address this limitation, we introduce STAMICS, a novel method that integrates semantic information with 3D Gaussian representations to enhance both localization and mapping accuracy. STAMICS consists of three key components: a 3D Gaussian-based scene representation for high-fidelity reconstruction, a graph-based clustering technique that enforces temporal semantic consistency, and an open-vocabulary system that allows for the classification of unseen objects. Extensive experiments show that STAMICS significantly improves camera pose estimation and map quality, outperforming state-of-the-art methods while reducing reconstruction errors. Code will be public available.
☆ Neuro-Symbolic Imitation Learning: Discovering Symbolic Abstractions for Skill Learning ICRA
Imitation learning is a popular method for teaching robots new behaviors. However, most existing methods focus on teaching short, isolated skills rather than long, multi-step tasks. To bridge this gap, imitation learning algorithms must not only learn individual skills but also an abstract understanding of how to sequence these skills to perform extended tasks effectively. This paper addresses this challenge by proposing a neuro-symbolic imitation learning framework. Using task demonstrations, the system first learns a symbolic representation that abstracts the low-level state-action space. The learned representation decomposes a task into easier subtasks and allows the system to leverage symbolic planning to generate abstract plans. Subsequently, the system utilizes this task decomposition to learn a set of neural skills capable of refining abstract plans into actionable robot commands. Experimental results in three simulated robotic environments demonstrate that, compared to baselines, our neuro-symbolic approach increases data efficiency, improves generalization capabilities, and facilitates interpretability.
comment: IEEE International Conference on Robotics and Automation (ICRA) 2025
☆ AcL: Action Learner for Fault-Tolerant Quadruped Locomotion Control
Quadrupedal robots can learn versatile locomotion skills but remain vulnerable when one or more joints lose power. In contrast, dogs and cats can adopt limping gaits when injured, demonstrating their remarkable ability to adapt to physical conditions. Inspired by such adaptability, this paper presents Action Learner (AcL), a novel teacher-student reinforcement learning framework that enables quadrupeds to autonomously adapt their gait for stable walking under multiple joint faults. Unlike conventional teacher-student approaches that enforce strict imitation, AcL leverages teacher policies to generate style rewards, guiding the student policy without requiring precise replication. We train multiple teacher policies, each corresponding to a different fault condition, and subsequently distill them into a single student policy with an encoder-decoder architecture. While prior works primarily address single-joint faults, AcL enables quadrupeds to walk with up to four faulty joints across one or two legs, autonomously switching between different limping gaits when faults occur. We validate AcL on a real Go2 quadruped robot under single- and double-joint faults, demonstrating fault-tolerant, stable walking, smooth gait transitions between normal and lamb gaits, and robustness against external disturbances.
☆ A Data-Driven Method for INS/DVL Alignment
Autonomous underwater vehicles (AUVs) are sophisticated robotic platforms crucial for a wide range of applications. The accuracy of AUV navigation systems is critical to their success. Inertial sensors and Doppler velocity logs (DVL) fusion is a promising solution for long-range underwater navigation. However, the effectiveness of this fusion depends heavily on an accurate alignment between the inertial sensors and the DVL. While current alignment methods show promise, there remains significant room for improvement in terms of accuracy, convergence time, and alignment trajectory efficiency. In this research we propose an end-to-end deep learning framework for the alignment process. By leveraging deep-learning capabilities, such as noise reduction and capture of nonlinearities in the data, we show using simulative data, that our proposed approach enhances both alignment accuracy and reduces convergence time beyond current model-based methods.
☆ UGNA-VPR: A Novel Training Paradigm for Visual Place Recognition Based on Uncertainty-Guided NeRF Augmentation
Visual place recognition (VPR) is crucial for robots to identify previously visited locations, playing an important role in autonomous navigation in both indoor and outdoor environments. However, most existing VPR datasets are limited to single-viewpoint scenarios, leading to reduced recognition accuracy, particularly in multi-directional driving or feature-sparse scenes. Moreover, obtaining additional data to mitigate these limitations is often expensive. This paper introduces a novel training paradigm to improve the performance of existing VPR networks by enhancing multi-view diversity within current datasets through uncertainty estimation and NeRF-based data augmentation. Specifically, we initially train NeRF using the existing VPR dataset. Then, our devised self-supervised uncertainty estimation network identifies places with high uncertainty. The poses of these uncertain places are input into NeRF to generate new synthetic observations for further training of VPR networks. Additionally, we propose an improved storage method for efficient organization of augmented and original training data. We conducted extensive experiments on three datasets and tested three different VPR backbone networks. The results demonstrate that our proposed training paradigm significantly improves VPR performance by fully utilizing existing data, outperforming other training approaches. We further validated the effectiveness of our approach on self-recorded indoor and outdoor datasets, consistently demonstrating superior results. Our dataset and code have been released at \href{https://github.com/nubot-nudt/UGNA-VPR}{https://github.com/nubot-nudt/UGNA-VPR}.
comment: Accepted to IEEE Robotics and Automation Letters (RA-L)
☆ Lidar-only Odometry based on Multiple Scan-to-Scan Alignments over a Moving Window
Lidar-only odometry considers the pose estimation of a mobile robot based on the accumulation of motion increments extracted from consecutive lidar scans. Many existing approaches to the problem use a scan-to-map registration, which neglects the accumulation of errors within the maintained map due to drift. Other methods use a refinement step that jointly optimizes the local map on a feature basis. We propose a solution that avoids this by using multiple independent scan-to-scan Iterative Closest Points (ICP) registrations to previous scans in order to derive constraints for a pose graph. The optimization of the pose graph then not only yields an accurate estimate for the latest pose, but also enables the refinement of previous scans in the optimization window. By avoiding the need to recompute the scan-to-scan alignments, the computational load is minimized. Extensive evaluation on the public KITTI and MulRan datasets as well as on a custom automotive lidar dataset is carried out. Results show that the proposed approach achieves state-of-the-art estimation accuracy, while alleviating the mentioned issues.
☆ An analysis of higher-order kinematics formalisms for an innovative surgical parallel robot
The paper presents a novel modular hybrid parallel robot for pancreatic surgery and its higher-order kinematics derived based on various formalisms. The classical vector, homogeneous transformation matrices and dual quaternion approaches are studied for the kinematic functions using both classical differentiation and multidual algebra. The algorithms for inverse kinematics for all three studied formalisms are presented for both differentiation and multidual algebra approaches. Furthermore, these algorithms are compared based on numerical stability, execution times and number and type of mathematical functions and operators contained in each algorithm. A statistical analysis shows that there is significant improvement in execution time for the algorithms implemented using multidual algebra, while the numerical stability is appropriate for all algorithms derived based on differentiation and multidual algebra. While the implementation of the kinematic algorithms using multidual algebra shows positive results when benchmarked on a standard PC, further work is required to evaluate the multidual algorithms on hardware/software used for the modular parallel robot command and control.
☆ Haptic bilateral teleoperation system for free-hand dental procedures
Free-hand dental procedures are typically repetitive, time-consuming and require high precision and manual dexterity. Dental robots can play a key role in improving procedural accuracy and safety, enhancing patient comfort, and reducing operator workload. However, robotic solutions for free-hand procedures remain limited or completely lacking, and their acceptance is still low. To address this gap, we develop a haptic bilateral teleoperation system (HBTS) for free-hand dental procedures. The system includes a dedicated mechanical end-effector, compatible with standard clinical tools, and equipped with an endoscopic camera for improved visibility of the intervention site. By ensuring motion and force correspondence between the operator's actions and the robot's movements, monitored through visual feedback, we enhance the operator's sensory awareness and motor accuracy. Furthermore, recognizing the need to ensure procedural safety, we limit interaction forces by scaling the motion references provided to the admittance controller based solely on measured contact forces. This ensures effective force limitation in all contact states without requiring prior knowledge of the environment. The proposed HBTS is validated in a dental scaling procedure using a dental phantom. The results show that the system improves the naturalness, safety, and accuracy of teleoperation, highlighting its potential to enhance free-hand dental procedures.
comment: 12 pages, 12 figures
☆ Output-Feedback Boundary Control of Thermally and Flow-Induced Vibrations in Slender Timoshenko Beams
This work is motivated by the engineering challenge of suppressing vibrations in turbine blades of aero engines, which often operate under extreme thermal conditions and high-Mach aerodynamic environments that give rise to complex vibration phenomena, commonly referred to as thermally-induced and flow-induced vibrations. Using Hamilton's variational principle, the system is modeled as a rotating slender Timoshenko beam under thermal and aerodynamic loads, described by a mixed hyperbolic-parabolic PDE system where instabilities occur both within the PDE domain and at the uncontrolled boundary, and the two types of PDEs are cascaded in the domain. For such a system, we present the state-feedback control design based on the PDE backstepping method. Recognizing that the distributed temperature gradients and structural vibrations in the Timoshenko beam are typically unmeasurable in practice, we design a state observer for the mixed hyperbolic-parabolic PDE system. Based on this observer, an output-feedback controller is then built to regulate the overall system using only available boundary measurements. In the closed-loop system, the state of the uncontrolled boundary, i.e., the furthest state from the control input, is proved to be exponentially convergent to zero, and all signals are proved as uniformly ultimately bounded. The proposed control design is validated on an aero-engine flexible blade under extreme thermal and aerodynamic conditions.
☆ OminiAdapt: Learning Cross-Task Invariance for Robust and Environment-Aware Robotic Manipulation
With the rapid development of embodied intelligence, leveraging large-scale human data for high-level imitation learning on humanoid robots has become a focal point of interest in both academia and industry. However, applying humanoid robots to precision operation domains remains challenging due to the complexities they face in perception and control processes, the long-standing physical differences in morphology and actuation mechanisms between humanoid robots and humans, and the lack of task-relevant features obtained from egocentric vision. To address the issue of covariate shift in imitation learning, this paper proposes an imitation learning algorithm tailored for humanoid robots. By focusing on the primary task objectives, filtering out background information, and incorporating channel feature fusion with spatial attention mechanisms, the proposed algorithm suppresses environmental disturbances and utilizes a dynamic weight update strategy to significantly improve the success rate of humanoid robots in accomplishing target tasks. Experimental results demonstrate that the proposed method exhibits robustness and scalability across various typical task scenarios, providing new ideas and approaches for autonomous learning and control in humanoid robots. The project will be open-sourced on GitHub.
☆ Dimensional optimization of single-DOF planar rigid link-flapping mechanisms for high lift and low power
Rigid link flapping mechanisms remain the most practical choice for flapping wing micro-aerial vehicles (MAVs) to carry useful payloads and onboard batteries for free flight due to their long-term durability and reliability. However, to achieve high agility and maneuverability-like insects-MAVs with these mechanisms require significant weight reduction. One approach involves using single-DOF planar rigid linkages, which are rarely optimized dimensionally for high lift and low power so that smaller motors and batteries could be used. We integrated a mechanism simulator based on a quasistatic nonlinear finite element method with an unsteady vortex lattice method-based aerodynamic analysis tool within an optimization routine. We optimized three different mechanism topologies from the literature. As a result, significant power savings were observed up to 42% in some cases, due to increased amplitude and higher lift coefficients resulting from optimized asymmetric sweeping velocity profiles. We also conducted an uncertainty analysis that revealed the need for high manufacturing tolerances to ensure reliable mechanism performance. The presented unified computational tool also facilitates the optimal selection of MAV components based on the payload and flight time requirements.
☆ TAGA: A Tangent-Based Reactive Approach for Socially Compliant Robot Navigation Around Human Groups IROS
Robot navigation in densely populated environments presents significant challenges, particularly regarding the interplay between individual and group dynamics. Current navigation models predominantly address interactions with individual pedestrians while failing to account for human groups that naturally form in real-world settings. Conversely, the limited models implementing group-aware navigation typically prioritize group dynamics at the expense of individual interactions, both of which are essential for socially appropriate navigation. This research extends an existing simulation framework to incorporate both individual pedestrians and human groups. We present Tangent Action for Group Avoidance (TAGA), a modular reactive mechanism that can be integrated with existing navigation frameworks to enhance their group-awareness capabilities. TAGA dynamically modifies robot trajectories using tangent action-based avoidance strategies while preserving the underlying model's capacity to navigate around individuals. Additionally, we introduce Group Collision Rate (GCR), a novel metric to quantitatively assess how effectively robots maintain group integrity during navigation. Through comprehensive simulation-based benchmarking, we demonstrate that integrating TAGA with state-of-the-art navigation models (ORCA, Social Force, DS-RNN, and AG-RL) reduces group intrusions by 45.7-78.6% while maintaining comparable success rates and navigation efficiency. Future work will focus on real-world implementation and validation of this approach.
comment: 6 pages, 3 figures. Submitted as a conference paper in IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2025
☆ Safe Human Robot Navigation in Warehouse Scenario
The integration of autonomous mobile robots (AMRs) in industrial environments, particularly warehouses, has revolutionized logistics and operational efficiency. However, ensuring the safety of human workers in dynamic, shared spaces remains a critical challenge. This work proposes a novel methodology that leverages control barrier functions (CBFs) to enhance safety in warehouse navigation. By integrating learning-based CBFs with the Open Robotics Middleware Framework (OpenRMF), the system achieves adaptive and safety-enhanced controls in multi-robot, multi-agent scenarios. Experiments conducted using various robot platforms demonstrate the efficacy of the proposed approach in avoiding static and dynamic obstacles, including human pedestrians. Our experiments evaluate different scenarios in which the number of robots, robot platforms, speed, and number of obstacles are varied, from which we achieve promising performance.
☆ Fuzzy-Logic-based model predictive control: A paradigm integrating optimal and common-sense decision making
This paper introduces a novel concept, fuzzy-logic-based model predictive control (FLMPC), along with a multi-robot control approach for exploring unknown environments and locating targets. Traditional model predictive control (MPC) methods rely on Bayesian theory to represent environmental knowledge and optimize a stochastic cost function, often leading to high computational costs and lack of effectiveness in locating all the targets. Our approach instead leverages FLMPC and extends it to a bi-level parent-child architecture for enhanced coordination and extended decision making horizon. Extracting high-level information from probability distributions and local observations, FLMPC simplifies the optimization problem and significantly extends its operational horizon compared to other MPC methods. We conducted extensive simulations in unknown 2-dimensional environments with randomly placed obstacles and humans. We compared the performance and computation time of FLMPC against MPC with a stochastic cost function, then evaluated the impact of integrating the high-level parent FLMPC layer. The results indicate that our approaches significantly improve both performance and computation time, enhancing coordination of robots and reducing the impact of uncertainty in large-scale search and rescue environments.
comment: 50 Pages, 8 figures, 3 tables
♻ ☆ Immersive and Wearable Thermal Rendering for Augmented Reality
In augmented reality (AR), where digital content is overlaid onto the real world, realistic thermal feedback has been shown to enhance immersion. Yet current thermal feedback devices, heavily influenced by the needs of virtual reality, often hinder physical interactions and are ineffective for immersion in AR. To bridge this gap, we have identified three design considerations relevant for AR thermal feedback: indirect feedback to maintain dexterity, thermal passthrough to preserve real-world temperature perception, and spatiotemporal rendering for dynamic sensations. We then created a unique and innovative thermal feedback device that satisfies these criteria. Human subject experiments assessing perceptual sensitivity, object temperature matching, spatial pattern recognition, and moving thermal stimuli demonstrated the impact of our design, enabling realistic temperature discrimination, virtual object perception, and enhanced immersion. These findings demonstrate that carefully designed thermal feedback systems can bridge the sensory gap between physical and virtual interactions, enhancing AR realism and usability.
♻ ☆ Model-Predictive Trajectory Generation for Aerial Search and Coverage
This paper introduces a trajectory planning algorithm for search and coverage missions with an Unmanned Aerial Vehicle (UAV) based on an uncertainty map that represents prior knowledge of the target region, modeled by a Gaussian Mixture Model (GMM). The trajectory planning problem is formulated as an Optimal Control Problem (OCP), which aims to maximize the uncertainty reduction within a specified mission duration. However, this results in an intractable OCP whose objective functional cannot be expressed in closed form. To address this, we propose a Model Predictive Control (MPC) algorithm based on a relaxed formulation of the objective function to approximate the optimal solutions. This relaxation promotes efficient map exploration by penalizing overlaps in the UAV's visibility regions along the trajectory. The algorithm can produce efficient and smooth trajectories, and it can be efficiently implemented using standard Nonlinear Programming solvers, being suitable for real-time planning. Unlike traditional methods, which often rely on discretizing the mission space and using complex mixed-integer formulations, our approach is computationally efficient and easier to implement. The MPC algorithm is initially assessed in MATLAB, followed by Gazebo simulations and actual experimental tests conducted in an outdoor environment. The results demonstrate that the proposed strategy can generate efficient and smooth trajectories for search and coverage missions.
♻ ☆ Towards Optimizing a Convex Cover of Collision-Free Space for Trajectory Generation
We propose an online iterative algorithm to optimize a convex cover to under-approximate the free space for autonomous navigation to delineate Safe Flight Corridors (SFC). The convex cover consists of a set of polytopes such that the union of the polytopes represents obstacle-free space, allowing us to find trajectories for robots that lie within the convex cover. In order to find the SFC that facilitates trajectory optimization, we iteratively find overlapping polytopes of maximum volumes that include specified waypoints initialized by a geometric or kinematic planner. Constraints at waypoints appear in two alternating stages of a joint optimization problem, which is solved by a novel heuristic-based iterative algorithm with partially distributed variables. We validate the effectiveness of our proposed algorithm using a range of parameterized environments and show its applications for two-stage motion planning.
♻ ☆ How NeRFs and 3D Gaussian Splatting are Reshaping SLAM: a Survey
Over the past two decades, research in the field of Simultaneous Localization and Mapping (SLAM) has undergone a significant evolution, highlighting its critical role in enabling autonomous exploration of unknown environments. This evolution ranges from hand-crafted methods, through the era of deep learning, to more recent developments focused on Neural Radiance Fields (NeRFs) and 3D Gaussian Splatting (3DGS) representations. Recognizing the growing body of research and the absence of a comprehensive survey on the topic, this paper aims to provide the first comprehensive overview of SLAM progress through the lens of the latest advancements in radiance fields. It sheds light on the background, evolutionary path, inherent strengths and limitations, and serves as a fundamental reference to highlight the dynamic progress and specific challenges.
comment: Updated to November 2024
♻ ☆ Efficient Continual Adaptation of Pretrained Robotic Policy with Online Meta-Learned Adapters
Continual adaptation is essential for general autonomous agents. For example, a household robot pretrained with a repertoire of skills must still adapt to unseen tasks specific to each household. Motivated by this, building upon parameter-efficient fine-tuning in language models, prior works have explored lightweight adapters to adapt pretrained policies, which can preserve learned features from the pretraining phase and demonstrate good adaptation performances. However, these approaches treat task learning separately, limiting knowledge transfer between tasks. In this paper, we propose Online Meta-Learned adapters (OMLA). Instead of applying adapters directly, OMLA can facilitate knowledge transfer from previously learned tasks to current learning tasks through a novel meta-learning objective. Extensive experiments in both simulated and real-world environments demonstrate that OMLA can lead to better adaptation performances compared to the baseline methods. The project link: https://ricky-zhu.github.io/OMLA/.
comment: Project link: https://ricky-zhu.github.io/OMLA/
♻ ☆ Integrating Naturalistic Insights in Objective Multi-Vehicle Safety Framework
As autonomous vehicle technology advances, the precise assessment of safety in complex traffic scenarios becomes crucial, especially in mixed-vehicle environments where human perception of safety must be taken into account. This paper presents a framework designed for assessing traffic safety in multi-vehicle situations, facilitating the simultaneous utilization of diverse objective safety metrics. Additionally, it allows the integration of subjective perception of safety by adjusting model parameters. The framework was applied to evaluate various model configurations in car-following scenarios on a highway, utilizing naturalistic driving datasets. The evaluation of the model showed an outstanding performance, particularly when integrating multiple objective safety measures. Furthermore, the performance was significantly enhanced when considering all surrounding vehicles.
♻ ☆ Online POMDP Planning with Anytime Deterministic Guarantees
Decision-making under uncertainty is a critical aspect of many practical autonomous systems due to incomplete information. Partially Observable Markov Decision Processes (POMDPs) offer a mathematically principled framework for formulating decision-making problems under such conditions. However, finding an optimal solution for a POMDP is generally intractable. In recent years, there has been a significant progress of scaling approximate solvers from small to moderately sized problems, using online tree search solvers. Often, such approximate solvers are limited to probabilistic or asymptotic guarantees towards the optimal solution. In this paper, we derive a deterministic relationship for discrete POMDPs between an approximated and the optimal solution. We show that at any time, we can derive bounds that relate between the existing solution and the optimal one. We show that our derivations provide an avenue for a new set of algorithms and can be attached to existing algorithms that have a certain structure to provide them with deterministic guarantees with marginal computational overhead. In return, not only do we certify the solution quality, but we demonstrate that making a decision based on the deterministic guarantee may result in superior performance compared to the original algorithm without the deterministic certification.
♻ ☆ Risk-Aware Reinforcement Learning for Autonomous Driving: Improving Safety When Driving through Intersection
Applying reinforcement learning to autonomous driving has garnered widespread attention. However, classical reinforcement learning methods optimize policies by maximizing expected rewards but lack sufficient safety considerations, often putting agents in hazardous situations. This paper proposes a risk-aware reinforcement learning approach for autonomous driving to improve the safety performance when crossing the intersection. Safe critics are constructed to evaluate driving risk and work in conjunction with the reward critic to update the actor. Based on this, a Lagrangian relaxation method and cyclic gradient iteration are combined to project actions into a feasible safe region. Furthermore, a Multi-hop and Multi-layer perception (MLP) mixed Attention Mechanism (MMAM) is incorporated into the actor-critic network, enabling the policy to adapt to dynamic traffic and overcome permutation sensitivity challenges. This allows the policy to focus more effectively on surrounding potential risks while enhancing the identification of passing opportunities. Simulation tests are conducted on different tasks at unsignalized intersections. The results show that the proposed approach effectively reduces collision rates and improves crossing efficiency in comparison to baseline algorithms. Additionally, our ablation experiments demonstrate the benefits of incorporating risk-awareness and MMAM into RL.
comment: 11 pages, 10 figures
♻ ☆ Constrained Nonlinear Kaczmarz Projection on Intersections of Manifolds for Coordinated Multi-Robot Mobile Manipulation ICRA
Cooperative manipulation tasks impose various structure-, task-, and robot-specific constraints on mobile manipulators. However, current methods struggle to model and solve these myriad constraints simultaneously. We propose a twofold solution: first, we model constraints as a family of manifolds amenable to simultaneous solving. Second, we introduce the constrained nonlinear Kaczmarz (cNKZ) projection technique to produce constraint-satisfying solutions. Experiments show that cNKZ dramatically outperforms baseline approaches, which cannot find solutions at all. We integrate cNKZ with a sampling-based motion planning algorithm to generate complex, coordinated motions for 3 to 6 mobile manipulators (18--36 DoF), with cNKZ solving up to 80 nonlinear constraints simultaneously and achieving up to a 92% success rate in cluttered environments. We also demonstrate our approach on hardware using three Turtlebot3 Waffle Pi robots with OpenMANIPULATOR-X arms.
comment: Accepted for publication at IEEE International Conference on Robotics and Automation (ICRA) 2025
♻ ☆ MUSE: A Real-Time Multi-Sensor State Estimator for Quadruped Robots
This paper introduces an innovative state estimator, MUSE (MUlti-sensor State Estimator), designed to enhance state estimation's accuracy and real-time performance in quadruped robot navigation. The proposed state estimator builds upon our previous work presented in [1]. It integrates data from a range of onboard sensors, including IMUs, encoders, cameras, and LiDARs, to deliver a comprehensive and reliable estimation of the robot's pose and motion, even in slippery scenarios. We tested MUSE on a Unitree Aliengo robot, successfully closing the locomotion control loop in difficult scenarios, including slippery and uneven terrain. Benchmarking against Pronto [2] and VILENS [3] showed 67.6% and 26.7% reductions in translational errors, respectively. Additionally, MUSE outperformed DLIO [4], a LiDAR-inertial odometry system in rotational errors and frequency, while the proprioceptive version of MUSE (P-MUSE) outperformed TSIF [5], with a 45.9% reduction in absolute trajectory error (ATE).
comment: Accepted for publication in IEEE Robotics and Automation Letters
♻ ☆ Mirroring the Parking Target: An Optimal-Control-Based Parking Motion Planner with Strengthened Parking Reliability and Faster Parking Completion
Automated Parking Assist (APA) systems are now facing great challenges of low adoption in applications, due to users' concerns about parking capability, reliability, and completion efficiency. To upgrade the conventional APA planners and enhance user's acceptance, this research proposes an optimal-control-based parking motion planner. Its highlight lies in its control logic: planning trajectories by mirroring the parking target. This method enables: i) parking capability in narrow spaces; ii) better parking reliability by expanding Operation Design Domain (ODD); iii) faster completion of parking process; iv) enhanced computational efficiency; v) universal to all types of parking. A comprehensive evaluation is conducted. Results demonstrate the proposed planner does enhance parking success rate by 40.6%, improve parking completion efficiency by 18.0%, and expand ODD by 86.1%. It shows its superiority in difficult parking cases, such as the parallel parking scenario and narrow spaces. Moreover, the average computation time of the proposed planner is 74 milliseconds. Results indicate that the proposed planner is ready for real-time commercial applications.
comment: IEEE Transactions on Intelligent Transportation Systems (2024)
♻ ☆ Safety-Aware Human-Lead Vehicle Platooning by Proactively Reacting to Uncertain Human Behaving
Human-Lead Cooperative Adaptive Cruise Control (HL-CACC) is regarded as a promising vehicle platooning technology in real-world implementation. By utilizing a Human-driven Vehicle (HV) as the platoon leader, HL-CACC reduces the cost and enhances the reliability of perception and decision-making. However, state-of-the-art HL-CACC technology still has a great limitation on driving safety due to the lack of considering the leading human driver's uncertain behavior. In this study, a HL-CACC controller is designed based on Stochastic Model Predictive Control (SMPC). It is enabled to predict the driving intention of the leading Connected Human-Driven Vehicle (CHV). The proposed controller has the following features: i) enhanced perceived safety in oscillating traffic; ii) guaranteed safety against hard brakes; iii) computational efficiency for real-time implementation. The proposed controller is evaluated on a PreScan&Simulink simulation platform. Real vehicle trajectory data is collected for the calibration of the simulation. Results reveal that the proposed controller: i) improves perceived safety by 19.17% in oscillating traffic; ii) enhances actual safety by 7.76% against hard brakes; iii) is confirmed with string stability. The computation time is approximately 3.2 milliseconds when running on a laptop equipped with an Intel i5-13500H CPU. This indicates the proposed controller is ready for real-time implementation.
♻ ☆ AlphaSpace: Enabling Robotic Actions through Semantic Tokenization and Symbolic Reasoning
This paper presents AlphaSpace, a novel methodology designed to enhance the spatial reasoning capabilities of language models for robotic manipulation in 3D Cartesian space. AlphaSpace employs a hierarchical semantics-based tokenization strategy that encodes spatial information at both coarse and fine-grained levels. Our approach represents objects with their attributes, positions, and height information through structured tokens, enabling precise spatial reasoning without relying on traditional vision-based embeddings. This approach enables LLMs to accurately manipulate objects by positioning them at specific (x, y, z) coordinates. Experimental results suggest that AlphaSpace demonstrates promising potential for improving manipulation tasks, achieving a total accuracy of 66.67%, compared to 37.5% for GPT-4o and 29.17% for Claude 3.5 Sonnet. These results demonstrate the potential of structured spatial encoding for manipulation tasks and warrant further exploration.
♻ ☆ LaMOuR: Leveraging Language Models for Out-of-Distribution Recovery in Reinforcement Learning
Deep Reinforcement Learning (DRL) has demonstrated strong performance in robotic control but remains susceptible to out-of-distribution (OOD) states, often resulting in unreliable actions and task failure. While previous methods have focused on minimizing or preventing OOD occurrences, they largely neglect recovery once an agent encounters such states. Although the latest research has attempted to address this by guiding agents back to in-distribution states, their reliance on uncertainty estimation hinders scalability in complex environments. To overcome this limitation, we introduce Language Models for Out-of-Distribution Recovery (LaMOuR), which enables recovery learning without relying on uncertainty estimation. LaMOuR generates dense reward codes that guide the agent back to a state where it can successfully perform its original task, leveraging the capabilities of LVLMs in image description, logical reasoning, and code generation. Experimental results show that LaMOuR substantially enhances recovery efficiency across diverse locomotion tasks and even generalizes effectively to complex environments, including humanoid locomotion and mobile manipulation, where existing methods struggle. The code and supplementary materials are available at https://lamour-rl.github.io/.
comment: This paper is currently under security review and will be re-released once the review is complete
♻ ☆ DexForce: Extracting Force-informed Actions from Kinesthetic Demonstrations for Dexterous Manipulation
Imitation learning requires high-quality demonstrations consisting of sequences of state-action pairs. For contact-rich dexterous manipulation tasks that require dexterity, the actions in these state-action pairs must produce the right forces. Current widely-used methods for collecting dexterous manipulation demonstrations are difficult to use for demonstrating contact-rich tasks due to unintuitive human-to-robot motion retargeting and the lack of direct haptic feedback. Motivated by these concerns, we propose DexForce. DexForce leverages contact forces, measured during kinesthetic demonstrations, to compute force-informed actions for policy learning. We collect demonstrations for six tasks and show that policies trained on our force-informed actions achieve an average success rate of 76% across all tasks. In contrast, policies trained directly on actions that do not account for contact forces have near-zero success rates. We also conduct a study ablating the inclusion of force data in policy observations. We find that while using force data never hurts policy performance, it helps most for tasks that require advanced levels of precision and coordination, like opening an AirPods case and unscrewing a nut.
comment: Videos can be found here: https://clairelc.github.io/dexforce.github.io/
GR00T N1: An Open Foundation Model for Generalist Humanoid Robots
General-purpose robots need a versatile body and an intelligent mind. Recent advancements in humanoid robots have shown great promise as a hardware platform for building generalist autonomy in the human world. A robot foundation model, trained on massive and diverse data sources, is essential for enabling the robots to reason about novel situations, robustly handle real-world variability, and rapidly learn new tasks. To this end, we introduce GR00T N1, an open foundation model for humanoid robots. GR00T N1 is a Vision-Language-Action (VLA) model with a dual-system architecture. The vision-language module (System 2) interprets the environment through vision and language instructions. The subsequent diffusion transformer module (System 1) generates fluid motor actions in real time. Both modules are tightly coupled and jointly trained end-to-end. We train GR00T N1 with a heterogeneous mixture of real-robot trajectories, human videos, and synthetically generated datasets. We show that our generalist robot model GR00T N1 outperforms the state-of-the-art imitation learning baselines on standard simulation benchmarks across multiple robot embodiments. Furthermore, we deploy our model on the Fourier GR-1 humanoid robot for language-conditioned bimanual manipulation tasks, achieving strong performance with high data efficiency.
comment: Authors are listed alphabetically. Project leads are Linxi "Jim" Fan and Yuke Zhu. For more information, see https://developer.nvidia.com/isaac/gr00t
♻ ☆ AnyBimanual: Transferring Unimanual Policy for General Bimanual Manipulation
Performing general language-conditioned bimanual manipulation tasks is of great importance for many applications ranging from household service to industrial assembly. However, collecting bimanual manipulation data is expensive due to the high-dimensional action space, which poses challenges for conventional methods to handle general bimanual manipulation tasks. In contrast, unimanual policy has recently demonstrated impressive generalizability across a wide range of tasks because of scaled model parameters and training data, which can provide sharable manipulation knowledge for bimanual systems. To this end, we propose a plug-and-play method named AnyBimanual, which transfers pre-trained unimanual policy to general bimanual manipulation policy with few bimanual demonstrations. Specifically, we first introduce a skill manager to dynamically schedule the skill representations discovered from pre-trained unimanual policy for bimanual manipulation tasks, which linearly combines skill primitives with task-oriented compensation to represent the bimanual manipulation instruction. To mitigate the observation discrepancy between unimanual and bimanual systems, we present a visual aligner to generate soft masks for visual embedding of the workspace, which aims to align visual input of unimanual policy model for each arm with those during pretraining stage. AnyBimanual shows superiority on 12 simulated tasks from RLBench2 with a sizable 12.67% improvement in success rate over previous methods. Experiments on 9 real-world tasks further verify its practicality with an average success rate of 84.62%.
comment: Project page: https://anybimanual.github.io/
♻ ☆ SyncDiff: Synchronized Motion Diffusion for Multi-Body Human-Object Interaction Synthesis
Synthesizing realistic human-object interaction motions is a critical problem in VR/AR and human animation. Unlike the commonly studied scenarios involving a single human or hand interacting with one object, we address a more generic multi-body setting with arbitrary numbers of humans, hands, and objects. This complexity introduces significant challenges in synchronizing motions due to the high correlations and mutual influences among bodies. To address these challenges, we introduce SyncDiff, a novel method for multi-body interaction synthesis using a synchronized motion diffusion strategy. SyncDiff employs a single diffusion model to capture the joint distribution of multi-body motions. To enhance motion fidelity, we propose a frequency-domain motion decomposition scheme. Additionally, we introduce a new set of alignment scores to emphasize the synchronization of different body motions. SyncDiff jointly optimizes both data sample likelihood and alignment likelihood through an explicit synchronization strategy. Extensive experiments across four datasets with various multi-body configurations demonstrate the superiority of SyncDiff over existing state-of-the-art motion synthesis methods.
comment: 26 pages, 10 figures
Artificial Intelligence 148
☆ StyleMotif: Multi-Modal Motion Stylization using Style-Content Cross Fusion
We present StyleMotif, a novel Stylized Motion Latent Diffusion model, generating motion conditioned on both content and style from multiple modalities. Unlike existing approaches that either focus on generating diverse motion content or transferring style from sequences, StyleMotif seamlessly synthesizes motion across a wide range of content while incorporating stylistic cues from multi-modal inputs, including motion, text, image, video, and audio. To achieve this, we introduce a style-content cross fusion mechanism and align a style encoder with a pre-trained multi-modal model, ensuring that the generated motion accurately captures the reference style while preserving realism. Extensive experiments demonstrate that our framework surpasses existing methods in stylized motion generation and exhibits emergent capabilities for multi-modal motion stylization, enabling more nuanced motion synthesis. Source code and pre-trained models will be released upon acceptance. Project Page: https://stylemotif.github.io
comment: Project Page: https://stylemotif.github.io
☆ Stable-SCore: A Stable Registration-based Framework for 3D Shape Correspondence CVPR 2025
Establishing character shape correspondence is a critical and fundamental task in computer vision and graphics, with diverse applications including re-topology, attribute transfer, and shape interpolation. Current dominant functional map methods, while effective in controlled scenarios, struggle in real situations with more complex challenges such as non-isometric shape discrepancies. In response, we revisit registration-for-correspondence methods and tap their potential for more stable shape correspondence estimation. To overcome their common issues including unstable deformations and the necessity for careful pre-alignment or high-quality initial 3D correspondences, we introduce Stable-SCore: A Stable Registration-based Framework for 3D Shape Correspondence. We first re-purpose a foundation model for 2D character correspondence that ensures reliable and stable 2D mappings. Crucially, we propose a novel Semantic Flow Guided Registration approach that leverages 2D correspondence to guide mesh deformations. Our framework significantly surpasses existing methods in challenging scenarios, and brings possibilities for a wide array of real applications, as demonstrated in our results.
comment: Accepted by CVPR 2025. Homepage: https://haolinliu97.github.io/Stable-Score/
☆ Uni4D: Unifying Visual Foundation Models for 4D Modeling from a Single Video CVPR 2025
This paper presents a unified approach to understanding dynamic scenes from casual videos. Large pretrained vision foundation models, such as vision-language, video depth prediction, motion tracking, and segmentation models, offer promising capabilities. However, training a single model for comprehensive 4D understanding remains challenging. We introduce Uni4D, a multi-stage optimization framework that harnesses multiple pretrained models to advance dynamic 3D modeling, including static/dynamic reconstruction, camera pose estimation, and dense 3D motion tracking. Our results show state-of-the-art performance in dynamic 4D modeling with superior visual quality. Notably, Uni4D requires no retraining or fine-tuning, highlighting the effectiveness of repurposing visual foundation models for 4D understanding.
comment: CVPR 2025. Project page (with code): https://davidyao99.github.io/uni4d
☆ Fwd2Bot: LVLM Visual Token Compression with Double Forward Bottleneck
In this work, we aim to compress the vision tokens of a Large Vision Language Model (LVLM) into a representation that is simultaneously suitable for (a) generative and (b) discriminative tasks, (c) is nearly lossless, and (d) is storage-efficient. We propose a novel compression approach, called Fwd2Bot, that uses the LVLM itself to compress the visual information in a task-agnostic manner. At the core of Fwd2bot there exists a "double-forward pass" training strategy, whereby, during the first forward pass, the LLM (of the LVLM) creates a bottleneck by condensing the visual information into a small number of summary tokens. Then, using the same LLM, the second forward pass processes the language instruction(s) alongside the summary tokens, used as a direct replacement for the image ones. The training signal is provided by two losses: an autoregressive one applied after the second pass that provides a direct optimization objective for compression, and a contrastive loss, applied after the first pass, that further boosts the representation strength, especially for discriminative tasks. The training is further enhanced by stage-specific adapters. We accompany the proposed method by an in-depth ablation study. Overall, Fwd2Bot results in highly-informative compressed representations suitable for both generative and discriminative tasks. For generative tasks, we offer a 2x higher compression rate without compromising the generative capabilities, setting a new state-of-the-art result. For discriminative tasks, we set a new state-of-the-art on image retrieval and compositionality.
☆ CTRL-O: Language-Controllable Object-Centric Visual Representation Learning CVPR 2025
Object-centric representation learning aims to decompose visual scenes into fixed-size vectors called "slots" or "object files", where each slot captures a distinct object. Current state-of-the-art object-centric models have shown remarkable success in object discovery in diverse domains, including complex real-world scenes. However, these models suffer from a key limitation: they lack controllability. Specifically, current object-centric models learn representations based on their preconceived understanding of objects, without allowing user input to guide which objects are represented. Introducing controllability into object-centric models could unlock a range of useful capabilities, such as the ability to extract instance-specific representations from a scene. In this work, we propose a novel approach for user-directed control over slot representations by conditioning slots on language descriptions. The proposed ConTRoLlable Object-centric representation learning approach, which we term CTRL-O, achieves targeted object-language binding in complex real-world scenes without requiring mask supervision. Next, we apply these controllable slot representations on two downstream vision language tasks: text-to-image generation and visual question answering. The proposed approach enables instance-specific text-to-image generation and also achieves strong performance on visual question answering.
comment: Accepted at CVPR 2025
☆ GateLens: A Reasoning-Enhanced LLM Agent for Automotive Software Release Analytics
Ensuring the reliability and effectiveness of software release decisions is critical, particularly in safety-critical domains like automotive systems. Precise analysis of release validation data, often presented in tabular form, plays a pivotal role in this process. However, traditional methods that rely on manual analysis of extensive test datasets and validation metrics are prone to delays and high costs. Large Language Models (LLMs) offer a promising alternative but face challenges in analytical reasoning, contextual understanding, handling out-of-scope queries, and processing structured test data consistently; limitations that hinder their direct application in safety-critical scenarios. This paper introduces GateLens, an LLM-based tool for analyzing tabular data in the automotive domain. GateLens translates natural language queries into Relational Algebra (RA) expressions and then generates optimized Python code. It outperforms the baseline system on benchmarking datasets, achieving higher F1 scores and handling complex and ambiguous queries with greater robustness. Ablation studies confirm the critical role of the RA module, with performance dropping sharply when omitted. Industrial evaluations reveal that GateLens reduces analysis time by over 80% while maintaining high accuracy and reliability. As demonstrated by presented results, GateLens achieved high performance without relying on few-shot examples, showcasing strong generalization across various query types from diverse company roles. Insights from deploying GateLens with a partner automotive company offer practical guidance for integrating AI into critical workflows such as release validation. Results show that by automating test result analysis, GateLens enables faster, more informed, and dependable release decisions, and can thus advance software scalability and reliability in automotive systems.
☆ ReaRAG: Knowledge-guided Reasoning Enhances Factuality of Large Reasoning Models with Iterative Retrieval Augmented Generation
Large Reasoning Models (LRMs) exhibit remarkable reasoning abilities but rely primarily on parametric knowledge, limiting factual accuracy. While recent works equip reinforcement learning (RL)-based LRMs with retrieval capabilities, they suffer from overthinking and lack robustness in reasoning, reducing their effectiveness in question answering (QA) tasks. To address this, we propose ReaRAG, a factuality-enhanced reasoning model that explores diverse queries without excessive iterations. Our solution includes a novel data construction framework with an upper bound on the reasoning chain length. Specifically, we first leverage an LRM to generate deliberate thinking, then select an action from a predefined action space (Search and Finish). For Search action, a query is executed against the RAG engine, where the result is returned as observation to guide reasoning steps later. This process iterates until a Finish action is chosen. Benefiting from ReaRAG's strong reasoning capabilities, our approach outperforms existing baselines on multi-hop QA. Further analysis highlights its strong reflective ability to recognize errors and refine its reasoning trajectory. Our study enhances LRMs' factuality while effectively integrating robust reasoning for Retrieval-Augmented Generation (RAG).
☆ Collab: Controlled Decoding using Mixture of Agents for LLM Alignment ICLR 2025
Alignment of Large Language models (LLMs) is crucial for safe and trustworthy deployment in applications. Reinforcement learning from human feedback (RLHF) has emerged as an effective technique to align LLMs to human preferences and broader utilities, but it requires updating billions of model parameters, which is computationally expensive. Controlled Decoding, by contrast, provides a mechanism for aligning a model at inference time without retraining. However, single-agent decoding approaches often struggle to adapt to diverse tasks due to the complexity and variability inherent in these tasks. To strengthen the test-time performance w.r.t the target task, we propose a mixture of agent-based decoding strategies leveraging the existing off-the-shelf aligned LLM policies. Treating each prior policy as an agent in the spirit of mixture of agent collaboration, we develop a decoding method that allows for inference-time alignment through a token-level selection strategy among multiple agents. For each token, the most suitable LLM is dynamically chosen from a pool of models based on a long-term utility metric. This policy-switching mechanism ensures optimal model selection at each step, enabling efficient collaboration and alignment among LLMs during decoding. Theoretical analysis of our proposed algorithm establishes optimal performance with respect to the target task represented via a target reward for the given off-the-shelf models. We conduct comprehensive empirical evaluations with open-source aligned models on diverse tasks and preferences, which demonstrates the merits of this approach over single-agent decoding baselines. Notably, Collab surpasses the current SoTA decoding strategy, achieving an improvement of up to 1.56x in average reward and 71.89% in GPT-4 based win-tie rate.
comment: Accepted to ICLR 2025
☆ Outlier dimensions favor frequent tokens in language model
We study last-layer outlier dimensions, i.e.dimensions that display extreme activations for the majority of inputs. We show that outlier dimensions arise in many different modern language models, and trace their function back to the heuristic of constantly predicting frequent words. We further show how a model can block this heuristic when it is not contextually appropriate, by assigning a counterbalancing weight mass to the remaining dimensions, and we investigate which model parameters boost outlier dimensions and when they arise during training. We conclude that outlier dimensions are a specialized mechanism discovered by many distinct models to implement a useful token prediction heuristic.
comment: 9 pages, 4 figures
☆ Elementwise Layer Normalization
A recent paper proposed Dynamic Tanh (DyT) as a drop-in replacement for Layer Normalization. Although the method is empirically well-motivated and appealing from a practical point of view, it lacks a theoretical foundation. In this work, we derive DyT mathematically and show that a well-defined approximation is needed to do so. By dropping said approximation, an alternative element-wise transformation is obtained, which we call Elementwise Layer Normalization (ELN). We demonstrate that ELN resembles Layer Normalization more accurately than DyT does.
comment: 11 pages, 3 figures
☆ MAVERIX: Multimodal Audio-Visual Evaluation Reasoning IndeX
Frontier models have either been language-only or have primarily focused on vision and language modalities. Although recent advancements in models with vision and audio understanding capabilities have shown substantial progress, the field lacks a standardized evaluation framework for thoroughly assessing their cross-modality perception performance. We introduce MAVERIX~(Multimodal Audio-Visual Evaluation Reasoning IndeX), a novel benchmark with 700 videos and 2,556 questions explicitly designed to evaluate multimodal models through tasks that necessitate close integration of video and audio information. MAVERIX uniquely provides models with audiovisual tasks, closely mimicking the multimodal perceptual experiences available to humans during inference and decision-making processes. To our knowledge, MAVERIX is the first benchmark aimed explicitly at assessing comprehensive audiovisual integration. Experiments with state-of-the-art models, including Gemini 1.5 Pro and o1, show performance approaching human levels (around 70% accuracy), while human experts reach near-ceiling performance (95.1%). With standardized evaluation protocols, a rigorously annotated pipeline, and a public toolkit, MAVERIX establishes a challenging testbed for advancing audiovisual multimodal intelligence.
☆ AMA-SAM: Adversarial Multi-Domain Alignment of Segment Anything Model for High-Fidelity Histology Nuclei Segmentation
Accurate segmentation of cell nuclei in histopathology images is essential for numerous biomedical research and clinical applications. However, existing cell nucleus segmentation methods only consider a single dataset (i.e., primary domain), while neglecting to leverage supplementary data from diverse sources (i.e., auxiliary domains) to reduce overfitting and enhance the performance. Although incorporating multiple datasets could alleviate overfitting, it often exacerbates performance drops caused by domain shifts. In this work, we introduce Adversarial Multi-domain Alignment of Segment Anything Model (AMA-SAM) that extends the Segment Anything Model (SAM) to overcome these obstacles through two key innovations. First, we propose a Conditional Gradient Reversal Layer (CGRL), a multi-domain alignment module that harmonizes features from diverse domains to promote domain-invariant representation learning while preserving crucial discriminative features for the primary dataset. Second, we address SAM's inherent low-resolution output by designing a High-Resolution Decoder (HR-Decoder), which directly produces fine-grained segmentation maps in order to capture intricate nuclei boundaries in high-resolution histology images. To the best of our knowledge, this is the first attempt to adapt SAM for multi-dataset learning with application to histology nuclei segmentation. We validate our method on several publicly available datasets, demonstrating consistent and significant improvements over state-of-the-art approaches.
comment: 13 pages, 4 tables, 2 figures
☆ Progressive Rendering Distillation: Adapting Stable Diffusion for Instant Text-to-Mesh Generation without 3D Data CVPR 2025
It is highly desirable to obtain a model that can generate high-quality 3D meshes from text prompts in just seconds. While recent attempts have adapted pre-trained text-to-image diffusion models, such as Stable Diffusion (SD), into generators of 3D representations (e.g., Triplane), they often suffer from poor quality due to the lack of sufficient high-quality 3D training data. Aiming at overcoming the data shortage, we propose a novel training scheme, termed as Progressive Rendering Distillation (PRD), eliminating the need for 3D ground-truths by distilling multi-view diffusion models and adapting SD into a native 3D generator. In each iteration of training, PRD uses the U-Net to progressively denoise the latent from random noise for a few steps, and in each step it decodes the denoised latent into 3D output. Multi-view diffusion models, including MVDream and RichDreamer, are used in joint with SD to distill text-consistent textures and geometries into the 3D outputs through score distillation. Since PRD supports training without 3D ground-truths, we can easily scale up the training data and improve generation quality for challenging text prompts with creative concepts. Meanwhile, PRD can accelerate the inference speed of the generation model in just a few steps. With PRD, we train a Triplane generator, namely TriplaneTurbo, which adds only $2.5\%$ trainable parameters to adapt SD for Triplane generation. TriplaneTurbo outperforms previous text-to-3D generators in both efficiency and quality. Specifically, it can produce high-quality 3D meshes in 1.2 seconds and generalize well for challenging text input. The code is available at https://github.com/theEricMa/TriplaneTurbo.
comment: Accepted to CVPR 2025. Code:https://github.com/theEricMa/TriplaneTurbo. Demo:https://huggingface.co/spaces/ZhiyuanthePony/TriplaneTurbo
☆ LLM-Gomoku: A Large Language Model-Based System for Strategic Gomoku with Self-Play and Reinforcement Learning
In recent years, large language models (LLMs) have shown significant advancements in natural language processing (NLP), with strong capa-bilities in generation, comprehension, and rea-soning. These models have found applications in education, intelligent decision-making, and gaming. However, effectively utilizing LLMs for strategic planning and decision-making in the game of Gomoku remains a challenge. This study aims to develop a Gomoku AI system based on LLMs, simulating the human learning process of playing chess. The system is de-signed to understand and apply Gomoku strat-egies and logic to make rational decisions. The research methods include enabling the model to "read the board," "understand the rules," "select strategies," and "evaluate positions," while en-hancing its abilities through self-play and rein-forcement learning. The results demonstrate that this approach significantly improves the se-lection of move positions, resolves the issue of generating illegal positions, and reduces pro-cess time through parallel position evaluation. After extensive self-play training, the model's Gomoku-playing capabilities have been notably enhanced.
☆ Intelligent IoT Attack Detection Design via ODLLM with Feature Ranking-based Knowledge Base
The widespread adoption of Internet of Things (IoT) devices has introduced significant cybersecurity challenges, particularly with the increasing frequency and sophistication of Distributed Denial of Service (DDoS) attacks. Traditional machine learning (ML) techniques often fall short in detecting such attacks due to the complexity of blended and evolving patterns. To address this, we propose a novel framework leveraging On-Device Large Language Models (ODLLMs) augmented with fine-tuning and knowledge base (KB) integration for intelligent IoT network attack detection. By implementing feature ranking techniques and constructing both long and short KBs tailored to model capacities, the proposed framework ensures efficient and accurate detection of DDoS attacks while overcoming computational and privacy limitations. Simulation results demonstrate that the optimized framework achieves superior accuracy across diverse attack types, especially when using compact models in edge computing environments. This work provides a scalable and secure solution for real-time IoT security, advancing the applicability of edge intelligence in cybersecurity.
☆ COMI-LINGUA: Expert Annotated Large-Scale Dataset for Multitask NLP in Hindi-English Code-Mixing
The rapid growth of digital communication has driven the widespread use of code-mixing, particularly Hindi-English, in multilingual communities. Existing datasets often focus on romanized text, have limited scope, or rely on synthetic data, which fails to capture realworld language nuances. Human annotations are crucial for assessing the naturalness and acceptability of code-mixed text. To address these challenges, We introduce COMI-LINGUA, the largest manually annotated dataset for code-mixed text, comprising 100,970 instances evaluated by three expert annotators in both Devanagari and Roman scripts. The dataset supports five fundamental NLP tasks: Language Identification, Matrix Language Identification, Part-of-Speech Tagging, Named Entity Recognition, and Translation. We evaluate LLMs on these tasks using COMILINGUA, revealing limitations in current multilingual modeling strategies and emphasizing the need for improved code-mixed text processing capabilities. COMI-LINGUA is publically availabe at: https://huggingface.co/datasets/LingoIITGN/COMI-LINGUA.
☆ Cognitive Science-Inspired Evaluation of Core Capabilities for Object Understanding in AI
One of the core components of our world models is 'intuitive physics' - an understanding of objects, space, and causality. This capability enables us to predict events, plan action and navigate environments, all of which rely on a composite sense of objecthood. Despite its importance, there is no single, unified account of objecthood, though multiple theoretical frameworks provide insights. In the first part of this paper, we present a comprehensive overview of the main theoretical frameworks in objecthood research - Gestalt psychology, enactive cognition, and developmental psychology - and identify the core capabilities each framework attributes to object understanding, as well as what functional roles they play in shaping world models in biological agents. Given the foundational role of objecthood in world modelling, understanding objecthood is also essential in AI. In the second part of the paper, we evaluate how current AI paradigms approach and test objecthood capabilities compared to those in cognitive science. We define an AI paradigm as a combination of how objecthood is conceptualised, the methods used for studying objecthood, the data utilised, and the evaluation techniques. We find that, whilst benchmarks can detect that AI systems model isolated aspects of objecthood, the benchmarks cannot detect when AI systems lack functional integration across these capabilities, not solving the objecthood challenge fully. Finally, we explore novel evaluation approaches that align with the integrated vision of objecthood outlined in this paper. These methods are promising candidates for advancing from isolated object capabilities toward general-purpose AI with genuine object understanding in real-world contexts.
☆ Model Assembly Learning with Heterogeneous Layer Weight Merging ICLR 2025
Model merging acquires general capabilities without extra data or training by combining multiple models' parameters. Previous approaches achieve linear mode connectivity by aligning parameters into the same loss basin using permutation invariance. In this paper, we introduce Model Assembly Learning (MAL), a novel paradigm for model merging that iteratively integrates parameters from diverse models in an open-ended model zoo to enhance the base model's capabilities. Unlike previous works that require identical architectures, MAL allows the merging of heterogeneous architectures and selective parameters across layers. Specifically, the base model can incorporate parameters from different layers of multiple pre-trained models. We systematically investigate the conditions and fundamental settings of heterogeneous parameter merging, addressing all possible mismatches in layer widths between the base and target models. Furthermore, we establish key laws and provide practical guidelines for effectively implementing MAL.
comment: ICLR 2025 Workshop on Neural Network Weights as a New Data Modality
☆ Unlocking the Potential of Past Research: Using Generative AI to Reconstruct Healthcare Simulation Models
Discrete-event simulation (DES) is widely used in healthcare Operations Research, but the models themselves are rarely shared. This limits their potential for reuse and long-term impact in the modelling and healthcare communities. This study explores the feasibility of using generative artificial intelligence (AI) to recreate published models using Free and Open Source Software (FOSS), based on the descriptions provided in an academic journal. Using a structured methodology, we successfully generated, tested and internally reproduced two DES models, including user interfaces. The reported results were replicated for one model, but not the other, likely due to missing information on distributions. These models are substantially more complex than AI-generated DES models published to date. Given the challenges we faced in prompt engineering, code generation, and model testing, we conclude that our iterative approach to model development, systematic comparison and testing, and the expertise of our team were necessary to the success of our recreated simulation models.
☆ Towards Fully Automated Decision-Making Systems for Greenhouse Control: Challenges and Opportunities
Machine learning has been successful in building control policies to drive a complex system to desired states in various applications (e.g. games, robotics, etc.). To be specific, a number of parameters of policy can be automatically optimized from the observations of environment to be able to generate a sequence of decisions leading to the best performance. In this survey paper, we particularly explore such policy-learning techniques for another unique, practical use-case scenario--farming, in which critical decisions (e.g., water supply, heating, etc.) must be made in a timely manner to minimize risks (e.g., damage to plants) while maximizing the revenue (e.g., healthy crops) in the end. We first provide a broad overview of latest studies on it to identify not only domain-specific challenges but opportunities with potential solutions, some of which are suggested as promising directions for future research. Also, we then introduce our successful approach to being ranked second among 46 teams at the ''3rd Autonomous Greenhouse Challenge'' to use this specific example to discuss the lessons learned about important considerations for design to create autonomous farm-management systems.
☆ When Astronomy Meets AI: Manazel For Crescent Visibility Prediction in Morocco
The accurate determination of the beginning of each Hijri month is essential for religious, cultural, and administrative purposes. Manazel (The code and datasets are available at https://github.com/lairgiyassir/manazel) addresses this challenge in Morocco by leveraging 13 years of crescent visibility data to refine the ODEH criterion, a widely used standard for lunar crescent visibility prediction. The study integrates two key features, the Arc of Vision (ARCV) and the total width of the crescent (W), to enhance the accuracy of lunar visibility assessments. A machine learning approach utilizing the Logistic Regression algorithm is employed to classify crescent visibility conditions, achieving a predictive accuracy of 98.83%. This data-driven methodology offers a robust and reliable framework for determining the start of the Hijri month, comparing different data classification tools, and improving the consistency of lunar calendar calculations in Morocco. The findings demonstrate the effectiveness of machine learning in astronomical applications and highlight the potential for further enhancements in the modeling of crescent visibility.
☆ UI-R1: Enhancing Action Prediction of GUI Agents by Reinforcement Learning
The recent DeepSeek-R1 has showcased the emergence of reasoning capabilities in LLMs through reinforcement learning (RL) with rule-based rewards. Building on this idea, we are the first to explore how rule-based RL can enhance the reasoning capabilities of multimodal large language models (MLLMs) for graphic user interface (GUI) action prediction tasks. To this end, we curate a small yet high-quality dataset of 136 challenging tasks, encompassing five common action types on mobile devices. We also introduce a unified rule-based action reward, enabling model optimization via policy-based algorithms such as Group Relative Policy Optimization (GRPO). Experimental results demonstrate that our proposed data-efficient model, UI-R1-3B, achieves substantial improvements on both in-domain (ID) and out-of-domain (OOD) tasks. Specifically, on the ID benchmark AndroidControl, the action type accuracy improves by 15%, while grounding accuracy increases by 10.3%, compared with the base model (i.e. Qwen2.5-VL-3B). On the OOD GUI grounding benchmark ScreenSpot-Pro, our model surpasses the base model by 6.0% and achieves competitive performance with larger models (e.g., OS-Atlas-7B), which are trained via supervised fine-tuning (SFT) on 76K data. These results underscore the potential of rule-based reinforcement learning to advance GUI understanding and control, paving the way for future research in this domain.
☆ A Measure Based Generalizable Approach to Understandability
Successful agent-human partnerships require that any agent generated information is understandable to the human, and that the human can easily steer the agent towards a goal. Such effective communication requires the agent to develop a finer-level notion of what is understandable to the human. State-of-the-art agents, including LLMs, lack this detailed notion of understandability because they only capture average human sensibilities from the training data, and therefore afford limited steerability (e.g., requiring non-trivial prompt engineering). In this paper, instead of only relying on data, we argue for developing generalizable, domain-agnostic measures of understandability that can be used as directives for these agents. Existing research on understandability measures is fragmented, we survey various such efforts across domains, and lay a cognitive-science-rooted groundwork for more coherent and domain-agnostic research investigations in future.
comment: 6 pages
☆ GenEdit: Compounding Operators and Continuous Improvement to Tackle Text-to-SQL in the Enterprise
Recent advancements in Text-to-SQL, driven by large language models, are democratizing data access. Despite these advancements, enterprise deployments remain challenging due to the need to capture business-specific knowledge, handle complex queries, and meet expectations of continuous improvements. To address these issues, we designed and implemented GenEdit: our Text-to-SQL generation system that improves with user feedback. GenEdit builds and maintains a company-specific knowledge set, employs a pipeline of operators decomposing SQL generation, and uses feedback to update its knowledge set to improve future SQL generations. We describe GenEdit's architecture made of two core modules: (i) decomposed SQL generation; and (ii) knowledge set edits based on user feedback. For generation, GenEdit leverages compounding operators to improve knowledge retrieval and to create a plan as chain-of-thought steps that guides generation. GenEdit first retrieves relevant examples in an initial retrieval stage where original SQL queries are decomposed into sub-statements, clauses or sub-queries. It then also retrieves instructions and schema elements. Using the retrieved contextual information, GenEdit then generates step-by-step plan in natural language on how to produce the query. Finally, GenEdit uses the plan to generate SQL, minimizing the need for model reasoning, which enhances complex SQL generation. If necessary, GenEdit regenerates the query based on syntactic and semantic errors. The knowledge set edits are recommended through an interactive copilot, allowing users to iterate on their feedback and to regenerate SQL queries as needed. Each generation uses staged edits which update the generation prompt. Once the feedback is submitted, it gets merged after passing regression testing and obtaining an approval, improving future generations.
☆ Prompt, Divide, and Conquer: Bypassing Large Language Model Safety Filters via Segmented and Distributed Prompt Processing
Large Language Models (LLMs) have transformed task automation and content generation across various domains while incorporating safety filters to prevent misuse. We introduce a novel jailbreaking framework that employs distributed prompt processing combined with iterative refinements to bypass these safety measures, particularly in generating malicious code. Our architecture consists of four key modules: prompt segmentation, parallel processing, response aggregation, and LLM-based jury evaluation. Tested on 500 malicious prompts across 10 cybersecurity categories, the framework achieves a 73.2% Success Rate (SR) in generating malicious code. Notably, our comparative analysis reveals that traditional single-LLM judge evaluation overestimates SRs (93.8%) compared to our LLM jury system (73.2%), with manual verification confirming that single-judge assessments often accept incomplete implementations. Moreover, we demonstrate that our distributed architecture improves SRs by 12% over the non-distributed approach in an ablation study, highlighting both the effectiveness of distributed prompt processing and the importance of robust evaluation methodologies in assessing jailbreak attempts.
comment: 22 pages; 26 figures
☆ Critical Iterative Denoising: A Discrete Generative Model Applied to Graphs
Discrete Diffusion and Flow Matching models have significantly advanced generative modeling for discrete structures, including graphs. However, the time dependencies in the noising process of these models lead to error accumulation and propagation during the backward process. This issue, particularly pronounced in mask diffusion, is a known limitation in sequence modeling and, as we demonstrate, also impacts discrete diffusion models for graphs. To address this problem, we propose a novel framework called Iterative Denoising, which simplifies discrete diffusion and circumvents the issue by assuming conditional independence across time. Additionally, we enhance our model by incorporating a Critic, which during generation selectively retains or corrupts elements in an instance based on their likelihood under the data distribution. Our empirical evaluations demonstrate that the proposed method significantly outperforms existing discrete diffusion baselines in graph generation tasks.
☆ AlignDiff: Learning Physically-Grounded Camera Alignment via Diffusion
Accurate camera calibration is a fundamental task for 3D perception, especially when dealing with real-world, in-the-wild environments where complex optical distortions are common. Existing methods often rely on pre-rectified images or calibration patterns, which limits their applicability and flexibility. In this work, we introduce a novel framework that addresses these challenges by jointly modeling camera intrinsic and extrinsic parameters using a generic ray camera model. Unlike previous approaches, AlignDiff shifts focus from semantic to geometric features, enabling more accurate modeling of local distortions. We propose AlignDiff, a diffusion model conditioned on geometric priors, enabling the simultaneous estimation of camera distortions and scene geometry. To enhance distortion prediction, we incorporate edge-aware attention, focusing the model on geometric features around image edges, rather than semantic content. Furthermore, to enhance generalizability to real-world captures, we incorporate a large database of ray-traced lenses containing over three thousand samples. This database characterizes the distortion inherent in a diverse variety of lens forms. Our experiments demonstrate that the proposed method significantly reduces the angular error of estimated ray bundles by ~8.2 degrees and overall calibration accuracy, outperforming existing approaches on challenging, real-world datasets.
☆ Magnitude-Phase Dual-Path Speech Enhancement Network based on Self-Supervised Embedding and Perceptual Contrast Stretch Boosting ICME 2025
Speech self-supervised learning (SSL) has made great progress in various speech processing tasks, but there is still room for improvement in speech enhancement (SE). This paper presents BSP-MPNet, a dual-path framework that combines self-supervised features with magnitude-phase information for SE. The approach starts by applying the perceptual contrast stretching (PCS) algorithm to enhance the magnitude-phase spectrum. A magnitude-phase 2D coarse (MP-2DC) encoder then extracts coarse features from the enhanced spectrum. Next, a feature-separating self-supervised learning (FS-SSL) model generates self-supervised embeddings for the magnitude and phase components separately. These embeddings are fused to create cross-domain feature representations. Finally, two parallel RNN-enhanced multi-attention (REMA) mask decoders refine the features, apply them to the mask, and reconstruct the speech signal. We evaluate BSP-MPNet on the VoiceBank+DEMAND and WHAMR! datasets. Experimental results show that BSP-MPNet outperforms existing methods under various noise conditions, providing new directions for self-supervised speech enhancement research. The implementation of the BSP-MPNet code is available online\footnote[2]{https://github.com/AlimMat/BSP-MPNet. \label{s1}}
comment: Main paper (6 pages). Accepted for publication by ICME 2025
☆ A Local Perspective-based Model for Overlapping Community Detection
Community detection, which identifies densely connected node clusters with sparse between-group links, is vital for analyzing network structure and function in real-world systems. Most existing community detection methods based on GCNs primarily focus on node-level information while overlooking community-level features, leading to performance limitations on large-scale networks. To address this issue, we propose LQ-GCN, an overlapping community detection model from a local community perspective. LQ-GCN employs a Bernoulli-Poisson model to construct a community affiliation matrix and form an end-to-end detection framework. By adopting local modularity as the objective function, the model incorporates local community information to enhance the quality and accuracy of clustering results. Additionally, the conventional GCNs architecture is optimized to improve the model capability in identifying overlapping communities in large-scale networks. Experimental results demonstrate that LQ-GCN achieves up to a 33% improvement in Normalized Mutual Information (NMI) and a 26.3% improvement in Recall compared to baseline models across multiple real-world benchmark datasets.
comment: 10 pages, 3 figures, 3 tables
☆ debug-gym: A Text-Based Environment for Interactive Debugging
Large Language Models (LLMs) are increasingly relied upon for coding tasks, yet in most scenarios it is assumed that all relevant information can be either accessed in context or matches their training data. We posit that LLMs can benefit from the ability to interactively explore a codebase to gather the information relevant to their task. To achieve this, we present a textual environment, namely debug-gym, for developing LLM-based agents in an interactive coding setting. Our environment is lightweight and provides a preset of useful tools, such as a Python debugger (pdb), designed to facilitate an LLM-based agent's interactive debugging. Beyond coding and debugging tasks, this approach can be generalized to other tasks that would benefit from information-seeking behavior by an LLM agent.
☆ SWI: Speaking with Intent in Large Language Models
Intent, typically clearly formulated and planned, functions as a cognitive framework for reasoning and problem-solving. This paper introduces the concept of Speaking with Intent (SWI) in large language models (LLMs), where the explicitly generated intent encapsulates the model's underlying intention and provides high-level planning to guide subsequent analysis and communication. By emulating deliberate and purposeful thoughts in the human mind, SWI is hypothesized to enhance the reasoning capabilities and generation quality of LLMs. Extensive experiments on mathematical reasoning benchmarks consistently demonstrate the superiority of Speaking with Intent over Baseline (i.e., generation without explicit intent). Moreover, SWI outperforms answer-trigger prompting methods Chain-of-Thought and Plan-and-Solve and maintains competitive performance with the strong method ARR (Analyzing, Retrieving, and Reasoning). Additionally, the effectiveness and generalizability of SWI are solidified on reasoning-intensive question answering (QA) and text summarization benchmarks, where SWI brings consistent improvement to the Baseline generation. In text summarization, SWI-generated summaries exhibit greater accuracy, conciseness, and factual correctness, with fewer hallucinations. Furthermore, human evaluations verify the coherence, effectiveness, and interpretability of the intent produced by SWI. This proof-of-concept study creates a novel avenue for enhancing LLMs' reasoning abilities with cognitive notions.
comment: 24 pages. Code: https://github.com/YuweiYin/SWI
☆ LOCATEdit: Graph Laplacian Optimized Cross Attention for Localized Text-Guided Image Editing
Text-guided image editing aims to modify specific regions of an image according to natural language instructions while maintaining the general structure and the background fidelity. Existing methods utilize masks derived from cross-attention maps generated from diffusion models to identify the target regions for modification. However, since cross-attention mechanisms focus on semantic relevance, they struggle to maintain the image integrity. As a result, these methods often lack spatial consistency, leading to editing artifacts and distortions. In this work, we address these limitations and introduce LOCATEdit, which enhances cross-attention maps through a graph-based approach utilizing self-attention-derived patch relationships to maintain smooth, coherent attention across image regions, ensuring that alterations are limited to the designated items while retaining the surrounding structure. \method consistently and substantially outperforms existing baselines on PIE-Bench, demonstrating its state-of-the-art performance and effectiveness on various editing tasks. Code can be found on https://github.com/LOCATEdit/LOCATEdit/
☆ Low-Resource Transliteration for Roman-Urdu and Urdu Using Transformer-Based Models
As the Information Retrieval (IR) field increasingly recognizes the importance of inclusivity, addressing the needs of low-resource languages remains a significant challenge. Transliteration between Urdu and its Romanized form, Roman Urdu, remains underexplored despite the widespread use of both scripts in South Asia. Prior work using RNNs on the Roman-Urdu-Parl dataset showed promising results but suffered from poor domain adaptability and limited evaluation. We propose a transformer-based approach using the m2m100 multilingual translation model, enhanced with masked language modeling (MLM) pretraining and fine-tuning on both Roman-Urdu-Parl and the domain-diverse Dakshina dataset. To address previous evaluation flaws, we introduce rigorous dataset splits and assess performance using BLEU, character-level BLEU, and CHRF. Our model achieves strong transliteration performance, with Char-BLEU scores of 96.37 for Urdu->Roman-Urdu and 97.44 for Roman-Urdu->Urdu. These results outperform both RNN baselines and GPT-4o Mini and demonstrate the effectiveness of multilingual transfer learning for low-resource transliteration tasks.
☆ MONO2REST: Identifying and Exposing Microservices: a Reusable RESTification Approach
The microservices architectural style has become the de facto standard for large-scale cloud applications, offering numerous benefits in scalability, maintainability, and deployment flexibility. Many organizations are pursuing the migration of legacy monolithic systems to a microservices architecture. However, this process is challenging, risky, time-intensive, and prone-to-failure while several organizations lack necessary financial resources, time, or expertise to set up this migration process. So, rather than trying to migrate a legacy system where migration is risky or not feasible, we suggest exposing it as a microservice application without without having to migrate it. In this paper, we present a reusable, automated, two-phase approach that combines evolutionary algorithms with machine learning techniques. In the first phase, we identify microservices at the method level using a multi-objective genetic algorithm that considers both structural and semantic dependencies between methods. In the second phase, we generate REST APIs for each identified microservice using a classification algorithm to assign HTTP methods and endpoints. We evaluated our approach with a case study on the Spring PetClinic application, which has both monolithic and microservices implementations that serve as ground truth for comparison. Results demonstrate that our approach successfully aligns identified microservices with those in the reference microservices implementation, highlighting its effectiveness in service identification and API generation.
☆ Quantitative Evaluation of Quantum/Classical Neural Network Using a Game Solver Metric
To evaluate the performance of quantum computing systems relative to classical counterparts and explore the potential for quantum advantage, we propose a game-solving benchmark based on Elo ratings in the game of tic-tac-toe. We compare classical convolutional neural networks (CNNs), quantum convolutional neural networks (QCNNs), and hybrid classical-quantum models by assessing their performance against a random-move agent in automated matches. Additionally, we implement a QCNN integrated with quantum communication and evaluate its performance to quantify the overhead introduced by noisy quantum channels. Our results show that the classical-quantum hybrid model achieves Elo ratings comparable to those of classical CNNs, while the standalone QCNN underperforms under current hardware constraints. The communication overhead was found to be modest. These findings demonstrate the viability of using game-based benchmarks for evaluating quantum computing systems and suggest that quantum communication can be incorporated with limited impact on performance, providing a foundation for future hybrid quantum applications.
comment: 11 pages, 16 figures
☆ Keyword-Oriented Multimodal Modeling for Euphemism Identification
Euphemism identification deciphers the true meaning of euphemisms, such as linking "weed" (euphemism) to "marijuana" (target keyword) in illicit texts, aiding content moderation and combating underground markets. While existing methods are primarily text-based, the rise of social media highlights the need for multimodal analysis, incorporating text, images, and audio. However, the lack of multimodal datasets for euphemisms limits further research. To address this, we regard euphemisms and their corresponding target keywords as keywords and first introduce a keyword-oriented multimodal corpus of euphemisms (KOM-Euph), involving three datasets (Drug, Weapon, and Sexuality), including text, images, and speech. We further propose a keyword-oriented multimodal euphemism identification method (KOM-EI), which uses cross-modal feature alignment and dynamic fusion modules to explicitly utilize the visual and audio features of the keywords for efficient euphemism identification. Extensive experiments demonstrate that KOM-EI outperforms state-of-the-art models and large language models, and show the importance of our multimodal datasets.
☆ Adaptive Resampling with Bootstrap for Noisy Multi-Objective Optimization Problems
The challenge of noisy multi-objective optimization lies in the constant trade-off between exploring new decision points and improving the precision of known points through resampling. This decision should take into account both the variability of the objective functions and the current estimate of a point in relation to the Pareto front. Since the amount and distribution of noise are generally unknown, it is desirable for a decision function to be highly adaptive to the properties of the optimization problem. This paper presents a resampling decision function that incorporates the stochastic nature of the optimization problem by using bootstrapping and the probability of dominance. The distribution-free estimation of the probability of dominance is achieved using bootstrap estimates of the means. To make the procedure applicable even with very few observations, we transfer the distribution observed at other decision points. The efficiency of this resampling approach is demonstrated by applying it in the NSGA-II algorithm with a sequential resampling procedure under multiple noise variations.
comment: 14 pages. 5 figures
☆ The Procedural Content Generation Benchmark: An Open-source Testbed for Generative Challenges in Games
This paper introduces the Procedural Content Generation Benchmark for evaluating generative algorithms on different game content creation tasks. The benchmark comes with 12 game-related problems with multiple variants on each problem. Problems vary from creating levels of different kinds to creating rule sets for simple arcade games. Each problem has its own content representation, control parameters, and evaluation metrics for quality, diversity, and controllability. This benchmark is intended as a first step towards a standardized way of comparing generative algorithms. We use the benchmark to score three baseline algorithms: a random generator, an evolution strategy, and a genetic algorithm. Results show that some problems are easier to solve than others, as well as the impact the chosen objective has on quality, diversity, and controllability of the generated artifacts.
comment: 12 pages, 4 figures, 2 tables, published at FDG2025
☆ Retinal Fundus Multi-Disease Image Classification using Hybrid CNN-Transformer-Ensemble Architectures
Our research is motivated by the urgent global issue of a large population affected by retinal diseases, which are evenly distributed but underserved by specialized medical expertise, particularly in non-urban areas. Our primary objective is to bridge this healthcare gap by developing a comprehensive diagnostic system capable of accurately predicting retinal diseases solely from fundus images. However, we faced significant challenges due to limited, diverse datasets and imbalanced class distributions. To overcome these issues, we have devised innovative strategies. Our research introduces novel approaches, utilizing hybrid models combining deeper Convolutional Neural Networks (CNNs), Transformer encoders, and ensemble architectures sequentially and in parallel to classify retinal fundus images into 20 disease labels. Our overarching goal is to assess these advanced models' potential in practical applications, with a strong focus on enhancing retinal disease diagnosis accuracy across a broader spectrum of conditions. Importantly, our efforts have surpassed baseline model results, with the C-Tran ensemble model emerging as the leader, achieving a remarkable model score of 0.9166, surpassing the baseline score of 0.9. Additionally, experiments with the IEViT model showcased equally promising outcomes with improved computational efficiency. We've also demonstrated the effectiveness of dynamic patch extraction and the integration of domain knowledge in computer vision tasks. In summary, our research strives to contribute significantly to retinal disease diagnosis, addressing the critical need for accessible healthcare solutions in underserved regions while aiming for comprehensive and accurate disease prediction.
comment: 17 pages, 3 figures, 7 tables. Conference paper presented at the International Health Informatics Conference (IHIC 2023)
☆ Harnessing Chain-of-Thought Metadata for Task Routing and Adversarial Prompt Detection
In this work, we propose a metric called Number of Thoughts (NofT) to determine the difficulty of tasks pre-prompting and support Large Language Models (LLMs) in production contexts. By setting thresholds based on the number of thoughts, this metric can discern the difficulty of prompts and support more effective prompt routing. A 2% decrease in latency is achieved when routing prompts from the MathInstruct dataset through quantized, distilled versions of Deepseek with 1.7 billion, 7 billion, and 14 billion parameters. Moreover, this metric can be used to detect adversarial prompts used in prompt injection attacks with high efficacy. The Number of Thoughts can inform a classifier that achieves 95% accuracy in adversarial prompt detection. Our experiments ad datasets used are available on our GitHub page: https://github.com/rymarinelli/Number_Of_Thoughts/tree/main.
☆ Unveiling Latent Information in Transaction Hashes: Hypergraph Learning for Ethereum Ponzi Scheme Detection
With the widespread adoption of Ethereum, financial frauds such as Ponzi schemes have become increasingly rampant in the blockchain ecosystem, posing significant threats to the security of account assets. Existing Ethereum fraud detection methods typically model account transactions as graphs, but this approach primarily focuses on binary transactional relationships between accounts, failing to adequately capture the complex multi-party interaction patterns inherent in Ethereum. To address this, we propose a hypergraph modeling method for the Ponzi scheme detection method in Ethereum, called HyperDet. Specifically, we treat transaction hashes as hyperedges that connect all the relevant accounts involved in a transaction. Additionally, we design a two-step hypergraph sampling strategy to significantly reduce computational complexity. Furthermore, we introduce a dual-channel detection module, including the hypergraph detection channel and the hyper-homo graph detection channel, to be compatible with existing detection methods. Experimental results show that, compared to traditional homogeneous graph-based methods, the hyper-homo graph detection channel achieves significant performance improvements, demonstrating the superiority of hypergraph in Ponzi scheme detection. This research offers innovations for modeling complex relationships in blockchain data.
☆ Graph-to-Vision: Multi-graph Understanding and Reasoning using Vision-Language Models
Graph Neural Networks (GNNs), as the dominant paradigm for graph-structured learning, have long faced dual challenges of exponentially escalating computational complexity and inadequate cross-scenario generalization capability. With the rapid advancement of multimodal learning, Vision-Language Models (VLMs) have demonstrated exceptional cross-modal relational reasoning capabilities and generalization capacities, thereby opening up novel pathways for overcoming the inherent limitations of conventional graph learning paradigms. However, current research predominantly concentrates on investigating the single-graph reasoning capabilities of VLMs, which fundamentally fails to address the critical requirement for coordinated reasoning across multiple heterogeneous graph data in real-world application scenarios. To address these limitations, we propose the first multi-graph joint reasoning benchmark for VLMs. Our benchmark encompasses four graph categories: knowledge graphs, flowcharts, mind maps, and route maps,with each graph group accompanied by three progressively challenging instruction-response pairs. Leveraging this benchmark, we conducted comprehensive capability assessments of state-of-the-art VLMs and performed fine-tuning on open-source models. This study not only addresses the underexplored evaluation gap in multi-graph reasoning for VLMs but also empirically validates their generalization superiority in graph-structured learning.
☆ Neuroplasticity in Artificial Intelligence -- An Overview and Inspirations on Drop In \& Out Learning
Artificial Intelligence (AI) has achieved new levels of performance and spread in public usage with the rise of deep neural networks (DNNs). Initially inspired by human neurons and their connections, NNs have become the foundation of AI models for many advanced architectures. However, some of the most integral processes in the human brain, particularly neurogenesis and neuroplasticity in addition to the more spread neuroapoptosis have largely been ignored in DNN architecture design. Instead, contemporary AI development predominantly focuses on constructing advanced frameworks, such as large language models, which retain a static structure of neural connections during training and inference. In this light, we explore how neurogenesis, neuroapoptosis, and neuroplasticity can inspire future AI advances. Specifically, we examine analogous activities in artificial NNs, introducing the concepts of ``dropin'' for neurogenesis and revisiting ``dropout'' and structural pruning for neuroapoptosis. We additionally suggest neuroplasticity combining the two for future large NNs in ``life-long learning'' settings following the biological inspiration. We conclude by advocating for greater research efforts in this interdisciplinary domain and identifying promising directions for future exploration.
☆ Federated Intelligence: When Large AI Models Meet Federated Fine-Tuning and Collaborative Reasoning at the Network Edge
Large artificial intelligence (AI) models exhibit remarkable capabilities in various application scenarios, but deploying them at the network edge poses significant challenges due to issues such as data privacy, computational resources, and latency. In this paper, we explore federated fine-tuning and collaborative reasoning techniques to facilitate the implementation of large AI models in resource-constrained wireless networks. Firstly, promising applications of large AI models within specific domains are discussed. Subsequently, federated fine-tuning methods are proposed to adapt large AI models to specific tasks or environments at the network edge, effectively addressing the challenges associated with communication overhead and enhancing communication efficiency. These methodologies follow clustered, hierarchical, and asynchronous paradigms to effectively tackle privacy issues and eliminate data silos. Furthermore, to enhance operational efficiency and reduce latency, efficient frameworks for model collaborative reasoning are developed, which include decentralized horizontal collaboration, cloud-edge-end vertical collaboration, and multi-access collaboration. Next, simulation results demonstrate the effectiveness of our proposed methods in reducing the fine-tuning loss of large AI models across various downstream tasks. Finally, several open challenges and research opportunities are outlined.
comment: 8 pages, 6 figures
☆ Exploring the Roles of Large Language Models in Reshaping Transportation Systems: A Survey, Framework, and Roadmap
Modern transportation systems face pressing challenges due to increasing demand, dynamic environments, and heterogeneous information integration. The rapid evolution of Large Language Models (LLMs) offers transformative potential to address these challenges. Extensive knowledge and high-level capabilities derived from pretraining evolve the default role of LLMs as text generators to become versatile, knowledge-driven task solvers for intelligent transportation systems. This survey first presents LLM4TR, a novel conceptual framework that systematically categorizes the roles of LLMs in transportation into four synergetic dimensions: information processors, knowledge encoders, component generators, and decision facilitators. Through a unified taxonomy, we systematically elucidate how LLMs bridge fragmented data pipelines, enhance predictive analytics, simulate human-like reasoning, and enable closed-loop interactions across sensing, learning, modeling, and managing tasks in transportation systems. For each role, our review spans diverse applications, from traffic prediction and autonomous driving to safety analytics and urban mobility optimization, highlighting how emergent capabilities of LLMs such as in-context learning and step-by-step reasoning can enhance the operation and management of transportation systems. We further curate practical guidance, including available resources and computational guidelines, to support real-world deployment. By identifying challenges in existing LLM-based solutions, this survey charts a roadmap for advancing LLM-driven transportation research, positioning LLMs as central actors in the next generation of cyber-physical-social mobility ecosystems. Online resources can be found in the project page: https://github.com/tongnie/awesome-llm4tr.
☆ Neuro-Symbolic Imitation Learning: Discovering Symbolic Abstractions for Skill Learning ICRA
Imitation learning is a popular method for teaching robots new behaviors. However, most existing methods focus on teaching short, isolated skills rather than long, multi-step tasks. To bridge this gap, imitation learning algorithms must not only learn individual skills but also an abstract understanding of how to sequence these skills to perform extended tasks effectively. This paper addresses this challenge by proposing a neuro-symbolic imitation learning framework. Using task demonstrations, the system first learns a symbolic representation that abstracts the low-level state-action space. The learned representation decomposes a task into easier subtasks and allows the system to leverage symbolic planning to generate abstract plans. Subsequently, the system utilizes this task decomposition to learn a set of neural skills capable of refining abstract plans into actionable robot commands. Experimental results in three simulated robotic environments demonstrate that, compared to baselines, our neuro-symbolic approach increases data efficiency, improves generalization capabilities, and facilitates interpretability.
comment: IEEE International Conference on Robotics and Automation (ICRA) 2025
☆ An evaluation of LLMs and Google Translate for translation of selected Indian languages via sentiment and semantic analyses
Large Language models (LLMs) have been prominent for language translation, including low-resource languages. There has been limited study about the assessment of the quality of translations generated by LLMs, including Gemini, GPT and Google Translate. In this study, we address this limitation by using semantic and sentiment analysis of selected LLMs for Indian languages, including Sanskrit, Telugu and Hindi. We select prominent texts that have been well translated by experts and use LLMs to generate their translations to English, and then we provide a comparison with selected expert (human) translations. Our findings suggest that while LLMs have made significant progress in translation accuracy, challenges remain in preserving sentiment and semantic integrity, especially in figurative and philosophical contexts. The sentiment analysis revealed that GPT-4o and GPT-3.5 are better at preserving the sentiments for the Bhagavad Gita (Sanskrit-English) translations when compared to Google Translate. We observed a similar trend for the case of Tamas (Hindi-English) and Maha P (Telugu-English) translations. GPT-4o performs similarly to GPT-3.5 in the translation in terms of sentiments for the three languages. We found that LLMs are generally better at translation for capturing sentiments when compared to Google Translate.
☆ HybridoNet-Adapt: A Domain-Adapted Framework for Accurate Lithium-Ion Battery RUL Prediction
Accurate prediction of the remaining useful life (RUL) in Lithium-ion battery (LIB) health management systems is crucial for ensuring reliability and safety. Current methods typically assume that training and testing data share the same distribution, overlooking the benefits of incorporating diverse data sources to enhance model performance. To address this limitation, we introduce a data-independent RUL prediction framework along with its domain adaptation (DA) approach, which leverages heterogeneous data sources for improved target predictions. Our approach integrates comprehensive data preprocessing, including feature extraction, denoising, and normalization, with a data-independent prediction model that combines Long Short-Term Memory (LSTM), Multihead Attention, and a Neural Ordinary Differential Equation (NODE) block, termed HybridoNet. The domain-adapted version, HybridoNet Adapt, is trained using a novel technique inspired by the Domain-Adversarial Neural Network (DANN) framework, a regression ensemble method, and Maximum Mean Discrepancy (MMD) to learn domain-invariant features from labeled cycling data in the source and target domains. Experimental results demonstrate that our approach outperforms state-of-the-art techniques, providing reliable RUL predictions for real-world applications.
☆ Investigating the Duality of Interpretability and Explainability in Machine Learning
The rapid evolution of machine learning (ML) has led to the widespread adoption of complex "black box" models, such as deep neural networks and ensemble methods. These models exhibit exceptional predictive performance, making them invaluable for critical decision-making across diverse domains within society. However, their inherently opaque nature raises concerns about transparency and interpretability, making them untrustworthy decision support systems. To alleviate such a barrier to high-stakes adoption, research community focus has been on developing methods to explain black box models as a means to address the challenges they pose. Efforts are focused on explaining these models instead of developing ones that are inherently interpretable. Designing inherently interpretable models from the outset, however, can pave the path towards responsible and beneficial applications in the field of ML. In this position paper, we clarify the chasm between explaining black boxes and adopting inherently interpretable models. We emphasize the imperative need for model interpretability and, following the purpose of attaining better (i.e., more effective or efficient w.r.t. predictive performance) and trustworthy predictors, provide an experimental evaluation of latest hybrid learning methods that integrates symbolic knowledge into neural network predictors. We demonstrate how interpretable hybrid models could potentially supplant black box ones in different domains.
☆ Using large language models to produce literature reviews: Usages and systematic biases of microphysics parametrizations in 2699 publications
Large language models afford opportunities for using computers for intensive tasks, realizing research opportunities that have not been considered before. One such opportunity could be a systematic interrogation of the scientific literature. Here, we show how a large language model can be used to construct a literature review of 2699 publications associated with microphysics parametrizations in the Weather and Research Forecasting (WRF) model, with the goal of learning how they were used and their systematic biases, when simulating precipitation. The database was constructed of publications identified from Web of Science and Scopus searches. The large language model GPT-4 Turbo was used to extract information about model configurations and performance from the text of 2699 publications. Our results reveal the landscape of how nine of the most popular microphysics parameterizations have been used around the world: Lin, Ferrier, WRF Single-Moment, Goddard Cumulus Ensemble, Morrison, Thompson, and WRF Double-Moment. More studies used one-moment parameterizations before 2020 and two-moment parameterizations after 2020. Seven out of nine parameterizations tended to overestimate precipitation. However, systematic biases of parameterizations differed in various regions. Except simulations using the Lin, Ferrier, and Goddard parameterizations that tended to underestimate precipitation over almost all locations, the remaining six parameterizations tended to overestimate, particularly over China, southeast Asia, western United States, and central Africa. This method could be used by other researchers to help understand how the increasingly massive body of scientific literature can be harnessed through the power of artificial intelligence to solve their research problems.
☆ Residual Learning Inspired Crossover Operator and Strategy Enhancements for Evolutionary Multitasking
In evolutionary multitasking, strategies such as crossover operators and skill factor assignment are critical for effective knowledge transfer. Existing improvements to crossover operators primarily focus on low-dimensional variable combinations, such as arithmetic crossover or partially mapped crossover, which are insufficient for modeling complex high-dimensional interactions.Moreover, static or semi-dynamic crossover strategies fail to adapt to the dynamic dependencies among tasks. In addition, current Multifactorial Evolutionary Algorithm frameworks often rely on fixed skill factor assignment strategies, lacking flexibility. To address these limitations, this paper proposes the Multifactorial Evolutionary Algorithm-Residual Learning (MFEA-RL) method based on residual learning. The method employs a Very Deep Super-Resolution (VDSR) model to generate high-dimensional residual representations of individuals, enhancing the modeling of complex relationships within dimensions. A ResNet-based mechanism dynamically assigns skill factors to improve task adaptability, while a random mapping mechanism efficiently performs crossover operations and mitigates the risk of negative transfer. Theoretical analysis and experimental results show that MFEA-RL outperforms state-of-the-art multitasking algorithms. It excels in both convergence and adaptability on standard evolutionary multitasking benchmarks, including CEC2017-MTSO and WCCI2020-MTSO. Additionally, its effectiveness is validated through a real-world application scenario.
comment: 9 pages, 4 figures
☆ A 71.2-$μ$W Speech Recognition Accelerator with Recurrent Spiking Neural Network
This paper introduces a 71.2-$\mu$W speech recognition accelerator designed for edge devices' real-time applications, emphasizing an ultra low power design. Achieved through algorithm and hardware co-optimizations, we propose a compact recurrent spiking neural network with two recurrent layers, one fully connected layer, and a low time step (1 or 2). The 2.79-MB model undergoes pruning and 4-bit fixed-point quantization, shrinking it by 96.42\% to 0.1 MB. On the hardware front, we take advantage of \textit{mixed-level pruning}, \textit{zero-skipping} and \textit{merged spike} techniques, reducing complexity by 90.49\% to 13.86 MMAC/S. The \textit{parallel time-step execution} addresses inter-time-step data dependencies and enables weight buffer power savings through weight sharing. Capitalizing on the sparse spike activity, an input broadcasting scheme eliminates zero computations, further saving power. Implemented on the TSMC 28-nm process, the design operates in real time at 100 kHz, consuming 71.2 $\mu$W, surpassing state-of-the-art designs. At 500 MHz, it has 28.41 TOPS/W and 1903.11 GOPS/mm$^2$ in energy and area efficiency, respectively.
☆ A Low-Power Streaming Speech Enhancement Accelerator For Edge Devices
Transformer-based speech enhancement models yield impressive results. However, their heterogeneous and complex structure restricts model compression potential, resulting in greater complexity and reduced hardware efficiency. Additionally, these models are not tailored for streaming and low-power applications. Addressing these challenges, this paper proposes a low-power streaming speech enhancement accelerator through model and hardware optimization. The proposed high performance model is optimized for hardware execution with the co-design of model compression and target application, which reduces 93.9\% of model size by the proposed domain-aware and streaming-aware pruning techniques. The required latency is further reduced with batch normalization-based transformers. Additionally, we employed softmax-free attention, complemented by an extra batch normalization, facilitating simpler hardware design. The tailored hardware accommodates these diverse computing patterns by breaking them down into element-wise multiplication and accumulation (MAC). This is achieved through a 1-D processing array, utilizing configurable SRAM addressing, thereby minimizing hardware complexities and simplifying zero skipping. Using the TSMC 40nm CMOS process, the final implementation requires merely 207.8K gates and 53.75KB SRAM. It consumes only 8.08 mW for real-time inference at a 62.5MHz frequency.
☆ ReFeed: Multi-dimensional Summarization Refinement with Reflective Reasoning on Feedback
Summarization refinement faces challenges when extending to multi-dimension. In this paper, we introduce ReFeed, a powerful summarization refinement pipeline that enhances multiple dimensions through reflective reasoning on feedback. To achieve this, we release SumFeed-CoT, a large-scale Long-CoT-based dataset optimized for training a lightweight model with reflective reasoning. Our experiments reveal how the number of dimensions, feedback exposure, and reasoning policy influence refinement performance, highlighting reflective reasoning and simultaneously addressing multiple feedback is crucial to mitigate trade-off between dimensions. Furthermore, ReFeed is robust to noisy feedback and feedback order. Lastly, our finding emphasizes that creating data with a proper goal and guideline constitutes a fundamental pillar of effective reasoning. The dataset and model will be released.
☆ HyperGraphRAG: Retrieval-Augmented Generation with Hypergraph-Structured Knowledge Representation
While standard Retrieval-Augmented Generation (RAG) based on chunks, GraphRAG structures knowledge as graphs to leverage the relations among entities. However, previous GraphRAG methods are limited by binary relations: one edge in the graph only connects two entities, which cannot well model the n-ary relations among more than two entities that widely exist in reality. To address this limitation, we propose HyperGraphRAG, a novel hypergraph-based RAG method that represents n-ary relational facts via hyperedges, modeling the complicated n-ary relations in the real world. To retrieve and generate over hypergraphs, we introduce a complete pipeline with a hypergraph construction method, a hypergraph retrieval strategy, and a hypergraph-guided generation mechanism. Experiments across medicine, agriculture, computer science, and law demonstrate that HyperGraphRAG outperforms standard RAG and GraphRAG in accuracy and generation quality.
comment: Preprint
☆ FineCIR: Explicit Parsing of Fine-Grained Modification Semantics for Composed Image Retrieval
Composed Image Retrieval (CIR) facilitates image retrieval through a multimodal query consisting of a reference image and modification text. The reference image defines the retrieval context, while the modification text specifies desired alterations. However, existing CIR datasets predominantly employ coarse-grained modification text (CoarseMT), which inadequately captures fine-grained retrieval intents. This limitation introduces two key challenges: (1) ignoring detailed differences leads to imprecise positive samples, and (2) greater ambiguity arises when retrieving visually similar images. These issues degrade retrieval accuracy, necessitating manual result filtering or repeated queries. To address these limitations, we develop a robust fine-grained CIR data annotation pipeline that minimizes imprecise positive samples and enhances CIR systems' ability to discern modification intents accurately. Using this pipeline, we refine the FashionIQ and CIRR datasets to create two fine-grained CIR datasets: Fine-FashionIQ and Fine-CIRR. Furthermore, we introduce FineCIR, the first CIR framework explicitly designed to parse the modification text. FineCIR effectively captures fine-grained modification semantics and aligns them with ambiguous visual entities, enhancing retrieval precision. Extensive experiments demonstrate that FineCIR consistently outperforms state-of-the-art CIR baselines on both fine-grained and traditional CIR benchmark datasets. Our FineCIR code and fine-grained CIR datasets are available at https://github.com/SDU-L/FineCIR.git.
☆ InternVL-X: Advancing and Accelerating InternVL Series with Efficient Visual Token Compression
Most multimodal large language models (MLLMs) treat visual tokens as "a sequence of text", integrating them with text tokens into a large language model (LLM). However, a great quantity of visual tokens significantly increases the demand for computational resources and time. In this paper, we propose InternVL-X, which outperforms the InternVL model in both performance and efficiency by incorporating three visual token compression methods. First, we propose a novel vision-language projector, PVTC. This component integrates adjacent visual embeddings to form a local query and utilizes the transformed CLS token as a global query, then performs point-to-region cross-attention through these local and global queries to more effectively convert visual features. Second, we present a layer-wise visual token compression module, LVTC, which compresses tokens in the LLM shallow layers and then expands them through upsampling and residual connections in the deeper layers. This significantly enhances the model computational efficiency. Futhermore, we propose an efficient high resolution slicing method, RVTC, which dynamically adjusts the number of visual tokens based on image area or length filtering. RVTC greatly enhances training efficiency with only a slight reduction in performance. By utilizing 20% or fewer visual tokens, InternVL-X achieves state-of-the-art performance on 7 public MLLM benchmarks, and improves the average metric by 2.34% across 12 tasks.
☆ DeBackdoor: A Deductive Framework for Detecting Backdoor Attacks on Deep Models with Limited Data
Backdoor attacks are among the most effective, practical, and stealthy attacks in deep learning. In this paper, we consider a practical scenario where a developer obtains a deep model from a third party and uses it as part of a safety-critical system. The developer wants to inspect the model for potential backdoors prior to system deployment. We find that most existing detection techniques make assumptions that are not applicable to this scenario. In this paper, we present a novel framework for detecting backdoors under realistic restrictions. We generate candidate triggers by deductively searching over the space of possible triggers. We construct and optimize a smoothed version of Attack Success Rate as our search objective. Starting from a broad class of template attacks and just using the forward pass of a deep model, we reverse engineer the backdoor attack. We conduct extensive evaluation on a wide range of attacks, models, and datasets, with our technique performing almost perfectly across these settings.
☆ Multi-Scale Invertible Neural Network for Wide-Range Variable-Rate Learned Image Compression
Autoencoder-based structures have dominated recent learned image compression methods. However, the inherent information loss associated with autoencoders limits their rate-distortion performance at high bit rates and restricts their flexibility of rate adaptation. In this paper, we present a variable-rate image compression model based on invertible transform to overcome these limitations. Specifically, we design a lightweight multi-scale invertible neural network, which bijectively maps the input image into multi-scale latent representations. To improve the compression efficiency, a multi-scale spatial-channel context model with extended gain units is devised to estimate the entropy of the latent representation from high to low levels. Experimental results demonstrate that the proposed method achieves state-of-the-art performance compared to existing variable-rate methods, and remains competitive with recent multi-model approaches. Notably, our method is the first learned image compression solution that outperforms VVC across a very wide range of bit rates using a single model, especially at high bit rates.The source code is available at \href{https://github.com/hytu99/MSINN-VRLIC}{https://github.com/hytu99/MSINN-VRLIC}.
comment: Accepted to IEEE Transactions on Multimedia 2025
☆ Reinforced Model Merging
The success of large language models has garnered widespread attention for model merging techniques, especially training-free methods which combine model capabilities within the parameter space. However, two challenges remain: (1) uniform treatment of all parameters leads to performance degradation; (2) search-based algorithms are often inefficient. In this paper, we present an innovative framework termed Reinforced Model Merging (RMM), which encompasses an environment and agent tailored for merging tasks. These components interact to execute layer-wise merging actions, aiming to search the optimal merging architecture. Notably, RMM operates without any gradient computations on the original models, rendering it feasible for edge devices. Furthermore, by utilizing data subsets during the evaluation process, we addressed the bottleneck in the reward feedback phase, thereby accelerating RMM by up to 100 times. Extensive experiments demonstrate that RMM achieves state-of-the-art performance across various vision and NLP datasets and effectively overcomes the limitations of the existing baseline methods. Our code is available at https://github.com/WuDiHJQ/Reinforced-Model-Merging.
☆ Learn by Reasoning: Analogical Weight Generation for Few-Shot Class-Incremental Learning
Few-shot class-incremental Learning (FSCIL) enables models to learn new classes from limited data while retaining performance on previously learned classes. Traditional FSCIL methods often require fine-tuning parameters with limited new class data and suffer from a separation between learning new classes and utilizing old knowledge. Inspired by the analogical learning mechanisms of the human brain, we propose a novel analogical generative method. Our approach includes the Brain-Inspired Analogical Generator (BiAG), which derives new class weights from existing classes without parameter fine-tuning during incremental stages. BiAG consists of three components: Weight Self-Attention Module (WSA), Weight & Prototype Analogical Attention Module (WPAA), and Semantic Conversion Module (SCM). SCM uses Neural Collapse theory for semantic conversion, WSA supplements new class weights, and WPAA computes analogies to generate new class weights. Experiments on miniImageNet, CUB-200, and CIFAR-100 datasets demonstrate that our method achieves higher final and average accuracy compared to SOTA methods.
☆ OminiAdapt: Learning Cross-Task Invariance for Robust and Environment-Aware Robotic Manipulation
With the rapid development of embodied intelligence, leveraging large-scale human data for high-level imitation learning on humanoid robots has become a focal point of interest in both academia and industry. However, applying humanoid robots to precision operation domains remains challenging due to the complexities they face in perception and control processes, the long-standing physical differences in morphology and actuation mechanisms between humanoid robots and humans, and the lack of task-relevant features obtained from egocentric vision. To address the issue of covariate shift in imitation learning, this paper proposes an imitation learning algorithm tailored for humanoid robots. By focusing on the primary task objectives, filtering out background information, and incorporating channel feature fusion with spatial attention mechanisms, the proposed algorithm suppresses environmental disturbances and utilizes a dynamic weight update strategy to significantly improve the success rate of humanoid robots in accomplishing target tasks. Experimental results demonstrate that the proposed method exhibits robustness and scalability across various typical task scenarios, providing new ideas and approaches for autonomous learning and control in humanoid robots. The project will be open-sourced on GitHub.
☆ Vision-to-Music Generation: A Survey
Vision-to-music Generation, including video-to-music and image-to-music tasks, is a significant branch of multimodal artificial intelligence demonstrating vast application prospects in fields such as film scoring, short video creation, and dance music synthesis. However, compared to the rapid development of modalities like text and images, research in vision-to-music is still in its preliminary stage due to its complex internal structure and the difficulty of modeling dynamic relationships with video. Existing surveys focus on general music generation without comprehensive discussion on vision-to-music. In this paper, we systematically review the research progress in the field of vision-to-music generation. We first analyze the technical characteristics and core challenges for three input types: general videos, human movement videos, and images, as well as two output types of symbolic music and audio music. We then summarize the existing methodologies on vision-to-music generation from the architecture perspective. A detailed review of common datasets and evaluation metrics is provided. Finally, we discuss current challenges and promising directions for future research. We hope our survey can inspire further innovation in vision-to-music generation and the broader field of multimodal generation in academic research and industrial applications. To follow latest works and foster further innovation in this field, we are continuously maintaining a GitHub repository at https://github.com/wzk1015/Awesome-Vision-to-Music-Generation.
☆ Dual-Splitting Conformal Prediction for Multi-Step Time Series Forecasting
Time series forecasting is crucial for applications like resource scheduling and risk management, where multi-step predictions provide a comprehensive view of future trends. Uncertainty Quantification (UQ) is a mainstream approach for addressing forecasting uncertainties, with Conformal Prediction (CP) gaining attention due to its model-agnostic nature and statistical guarantees. However, most variants of CP are designed for single-step predictions and face challenges in multi-step scenarios, such as reliance on real-time data and limited scalability. This highlights the need for CP methods specifically tailored to multi-step forecasting. We propose the Dual-Splitting Conformal Prediction (DSCP) method, a novel CP approach designed to capture inherent dependencies within time-series data for multi-step forecasting. Experimental results on real-world datasets from four different domains demonstrate that the proposed DSCP significantly outperforms existing CP variants in terms of the Winkler Score, achieving a performance improvement of up to 23.59% compared to state-of-the-art methods. Furthermore, we deployed the DSCP approach for renewable energy generation and IT load forecasting in power management of a real-world trajectory-based application, achieving an 11.25% reduction in carbon emissions through predictive optimization of data center operations and controls.
comment: 28 pages, 13 figures, 3 tables. Submitted to Applied Soft Computing. With Editor This is the first public release of the work
☆ ResearchBench: Benchmarking LLMs in Scientific Discovery via Inspiration-Based Task Decomposition
Large language models (LLMs) have demonstrated potential in assisting scientific research, yet their ability to discover high-quality research hypotheses remains unexamined due to the lack of a dedicated benchmark. To address this gap, we introduce the first large-scale benchmark for evaluating LLMs with a near-sufficient set of sub-tasks of scientific discovery: inspiration retrieval, hypothesis composition, and hypothesis ranking. We develop an automated framework that extracts critical components - research questions, background surveys, inspirations, and hypotheses - from scientific papers across 12 disciplines, with expert validation confirming its accuracy. To prevent data contamination, we focus exclusively on papers published in 2024, ensuring minimal overlap with LLM pretraining data. Our evaluation reveals that LLMs perform well in retrieving inspirations, an out-of-distribution task, suggesting their ability to surface novel knowledge associations. This positions LLMs as "research hypothesis mines", capable of facilitating automated scientific discovery by generating innovative hypotheses at scale with minimal human intervention.
☆ Improving $(α, f)$-Byzantine Resilience in Federated Learning via layerwise aggregation and cosine distance
The rapid development of artificial intelligence systems has amplified societal concerns regarding their usage, necessitating regulatory frameworks that encompass data privacy. Federated Learning (FL) is posed as potential solution to data privacy challenges in distributed machine learning by enabling collaborative model training {without data sharing}. However, FL systems remain vulnerable to Byzantine attacks, where malicious nodes contribute corrupted model updates. While Byzantine Resilient operators have emerged as a widely adopted robust aggregation algorithm to mitigate these attacks, its efficacy diminishes significantly in high-dimensional parameter spaces, sometimes leading to poor performing models. This paper introduces Layerwise Cosine Aggregation, a novel aggregation scheme designed to enhance robustness of these rules in such high-dimensional settings while preserving computational efficiency. A theoretical analysis is presented, demonstrating the superior robustness of the proposed Layerwise Cosine Aggregation compared to original robust aggregation operators. Empirical evaluation across diverse image classification datasets, under varying data distributions and Byzantine attack scenarios, consistently demonstrates the improved performance of Layerwise Cosine Aggregation, achieving up to a 16% increase in model accuracy.
comment: Submitted to Knowledge-Based Systems
☆ Feature-Enhanced Machine Learning for All-Cause Mortality Prediction in Healthcare Data
Accurate patient mortality prediction enables effective risk stratification, leading to personalized treatment plans and improved patient outcomes. However, predicting mortality in healthcare remains a significant challenge, with existing studies often focusing on specific diseases or limited predictor sets. This study evaluates machine learning models for all-cause in-hospital mortality prediction using the MIMIC-III database, employing a comprehensive feature engineering approach. Guided by clinical expertise and literature, we extracted key features such as vital signs (e.g., heart rate, blood pressure), laboratory results (e.g., creatinine, glucose), and demographic information. The Random Forest model achieved the highest performance with an AUC of 0.94, significantly outperforming other machine learning and deep learning approaches. This demonstrates Random Forest's robustness in handling high-dimensional, noisy clinical data and its potential for developing effective clinical decision support tools. Our findings highlight the importance of careful feature engineering for accurate mortality prediction. We conclude by discussing implications for clinical adoption and propose future directions, including enhancing model robustness and tailoring prediction models for specific diseases.
☆ Bias-Aware Agent: Enhancing Fairness in AI-Driven Knowledge Retrieval
Advancements in retrieving accessible information have evolved faster in the last few years compared to the decades since the internet's creation. Search engines, like Google, have been the number one way to find relevant data. They have always relied on the user's abilities to find the best information in its billions of links and sources at everybody's fingertips. The advent of large language models (LLMs) has completely transformed the field of information retrieval. The LLMs excel not only at retrieving relevant knowledge but also at summarizing it effectively, making information more accessible and consumable for users. On top of it, the rise of AI Agents has introduced another aspect to information retrieval i.e. dynamic information retrieval which enables the integration of real-time data such as weather forecasts, and financial data with the knowledge base to curate context-aware knowledge. However, despite these advancements the agents remain susceptible to issues of bias and fairness, challenges deeply rooted within the knowledge base and training of LLMs. This study introduces a novel approach to bias-aware knowledge retrieval by leveraging agentic framework and the innovative use of bias detectors as tools to identify and highlight inherent biases in the retrieved content. By empowering users with transparency and awareness, this approach aims to foster more equitable information systems and promote the development of responsible AI.
☆ Knowledge Graphs as World Models for Semantic Material-Aware Obstacle Handling in Autonomous Vehicles
The inability of autonomous vehicles (AVs) to infer the material properties of obstacles limits their decision-making capacity. While AVs rely on sensor systems such as cameras, LiDAR, and radar to detect obstacles, this study suggests combining sensors with a knowledge graph (KG)-based world model to improve AVs' comprehension of physical material qualities. Beyond sensor data, AVs can infer qualities such as malleability, density, and elasticity using a semantic KG that depicts the relationships between obstacles and their attributes. Using the CARLA autonomous driving simulator, we evaluated AV performance with and without KG integration. The findings demonstrate that the KG-based method improves obstacle management, which allows AVs to use material qualities to make better decisions about when to change lanes or apply emergency braking. For example, the KG-integrated AV changed lanes for hard impediments like traffic cones and successfully avoided collisions with flexible items such as plastic bags by passing over them. Compared to the control system, the KG framework demonstrated improved responsiveness to obstacles by resolving conflicting sensor data, causing emergency stops for 13.3% more cases. In addition, our method exhibits a 6.6% higher success rate in lane-changing maneuvers in experimental scenarios, particularly for larger, high-impact obstacles. While we focus particularly on autonomous driving, our work demonstrates the potential of KG-based world models to improve decision-making in embodied AI systems and scale to other domains, including robotics, healthcare, and environmental simulation.
☆ GenFusion: Closing the Loop between Reconstruction and Generation via Videos
Recently, 3D reconstruction and generation have demonstrated impressive novel view synthesis results, achieving high fidelity and efficiency. However, a notable conditioning gap can be observed between these two fields, e.g., scalable 3D scene reconstruction often requires densely captured views, whereas 3D generation typically relies on a single or no input view, which significantly limits their applications. We found that the source of this phenomenon lies in the misalignment between 3D constraints and generative priors. To address this problem, we propose a reconstruction-driven video diffusion model that learns to condition video frames on artifact-prone RGB-D renderings. Moreover, we propose a cyclical fusion pipeline that iteratively adds restoration frames from the generative model to the training set, enabling progressive expansion and addressing the viewpoint saturation limitations seen in previous reconstruction and generation pipelines. Our evaluation, including view synthesis from sparse view and masked input, validates the effectiveness of our approach.
☆ Integrating Large Language Models For Monte Carlo Simulation of Chemical Reaction Networks
Chemical reaction network is an important method for modeling and exploring complex biological processes, bio-chemical interactions and the behavior of different dynamics in system biology. But, formulating such reaction kinetics takes considerable time. In this paper, we leverage the efficiency of modern large language models to automate the stochastic monte carlo simulation of chemical reaction networks and enable the simulation through the reaction description provided in the form of natural languages. We also integrate this process into widely used simulation tool Copasi to further give the edge and ease to the modelers and researchers. In this work, we show the efficacy and limitations of the modern large language models to parse and create reaction kinetics for modelling complex chemical reaction processes.
comment: Accepted on MadeAI 2025 Conference
☆ Adversarial Wear and Tear: Exploiting Natural Damage for Generating Physical-World Adversarial Examples
The presence of adversarial examples in the physical world poses significant challenges to the deployment of Deep Neural Networks in safety-critical applications such as autonomous driving. Most existing methods for crafting physical-world adversarial examples are ad-hoc, relying on temporary modifications like shadows, laser beams, or stickers that are tailored to specific scenarios. In this paper, we introduce a new class of physical-world adversarial examples, AdvWT, which draws inspiration from the naturally occurring phenomenon of `wear and tear', an inherent property of physical objects. Unlike manually crafted perturbations, `wear and tear' emerges organically over time due to environmental degradation, as seen in the gradual deterioration of outdoor signboards. To achieve this, AdvWT follows a two-step approach. First, a GAN-based, unsupervised image-to-image translation network is employed to model these naturally occurring damages, particularly in the context of outdoor signboards. The translation network encodes the characteristics of damaged signs into a latent `damage style code'. In the second step, we introduce adversarial perturbations into the style code, strategically optimizing its transformation process. This manipulation subtly alters the damage style representation, guiding the network to generate adversarial images where the appearance of damages remains perceptually realistic, while simultaneously ensuring their effectiveness in misleading neural networks. Through comprehensive experiments on two traffic sign datasets, we show that AdvWT effectively misleads DNNs in both digital and physical domains. AdvWT achieves an effective attack success rate, greater robustness, and a more natural appearance compared to existing physical-world adversarial examples. Additionally, integrating AdvWT into training enhances a model's generalizability to real-world damaged signs.
comment: 11 pages, 9 figures
☆ Multi-Objective Optimization for Privacy-Utility Balance in Differentially Private Federated Learning
Federated learning (FL) enables collaborative model training across distributed clients without sharing raw data, making it a promising approach for privacy-preserving machine learning. However, ensuring differential privacy (DP) in FL presents challenges due to the trade-off between model utility and privacy protection. Clipping gradients before aggregation is a common strategy to limit privacy loss, but selecting an optimal clipping norm is non-trivial, as excessively high values compromise privacy, while overly restrictive clipping degrades model performance. In this work, we propose an adaptive clipping mechanism that dynamically adjusts the clipping norm using a multi-objective optimization framework. By integrating privacy and utility considerations into the optimization objective, our approach balances privacy preservation with model accuracy. We theoretically analyze the convergence properties of our method and demonstrate its effectiveness through extensive experiments on MNIST, Fashion-MNIST, and CIFAR-10 datasets. Our results show that adaptive clipping consistently outperforms fixed-clipping baselines, achieving improved accuracy under the same privacy constraints. This work highlights the potential of dynamic clipping strategies to enhance privacy-utility trade-offs in differentially private federated learning.
☆ Federated Learning with Differential Privacy: An Utility-Enhanced Approach
Federated learning has emerged as an attractive approach to protect data privacy by eliminating the need for sharing clients' data while reducing communication costs compared with centralized machine learning algorithms. However, recent studies have shown that federated learning alone does not guarantee privacy, as private data may still be inferred from the uploaded parameters to the central server. In order to successfully avoid data leakage, adopting differential privacy (DP) in the local optimization process or in the local update aggregation process has emerged as two feasible ways for achieving sample-level or user-level privacy guarantees respectively, in federated learning models. However, compared to their non-private equivalents, these approaches suffer from a poor utility. To improve the privacy-utility trade-off, we present a modification to these vanilla differentially private algorithms based on a Haar wavelet transformation step and a novel noise injection scheme that significantly lowers the asymptotic bound of the noise variance. We also present a holistic convergence analysis of our proposed algorithm, showing that our method yields better convergence performance than the vanilla DP algorithms. Numerical experiments on real-world datasets demonstrate that our method outperforms existing approaches in model utility while maintaining the same privacy guarantees.
☆ The Devil is in Low-Level Features for Cross-Domain Few-Shot Segmentation CVPR 2025
Cross-Domain Few-Shot Segmentation (CDFSS) is proposed to transfer the pixel-level segmentation capabilities learned from large-scale source-domain datasets to downstream target-domain datasets, with only a few annotated images per class. In this paper, we focus on a well-observed but unresolved phenomenon in CDFSS: for target domains, particularly those distant from the source domain, segmentation performance peaks at the very early epochs, and declines sharply as the source-domain training proceeds. We delve into this phenomenon for an interpretation: low-level features are vulnerable to domain shifts, leading to sharper loss landscapes during the source-domain training, which is the devil of CDFSS. Based on this phenomenon and interpretation, we further propose a method that includes two plug-and-play modules: one to flatten the loss landscapes for low-level features during source-domain training as a novel sharpness-aware minimization method, and the other to directly supplement target-domain information to the model during target-domain testing by low-level-based calibration. Extensive experiments on four target datasets validate our rationale and demonstrate that our method surpasses the state-of-the-art method in CDFSS signifcantly by 3.71% and 5.34% average MIoU in 1-shot and 5-shot scenarios, respectively.
comment: Accepted by CVPR 2025
☆ A computational theory of evaluation for parameterisable subject
Evaluation is critical to advance decision making across domains, yet existing methodologies often struggle to balance theoretical rigor and practical scalability. In order to reduce the cost of experimental evaluation, we introduce a computational theory of evaluation for parameterisable subjects. We prove upper bounds of generalized evaluation error and generalized causal effect error of evaluation metric on subject. We also prove efficiency, and consistency to estimated causal effect of subject on metric by prediction. To optimize evaluation models, we propose a meta-learner to handle heterogeneous evaluation subjects space. Comparing with other computational approaches, our (conditional) evaluation model reduced 24.1%-99.0% evaluation errors across 12 scenes, including individual medicine, scientific simulation, business activities, and quantum trade. The evaluation time is reduced 3-7 order of magnitude comparing with experiments or simulations.
☆ Optimizing Multi-DNN Inference on Mobile Devices through Heterogeneous Processor Co-Execution
Deep Neural Networks (DNNs) are increasingly deployed across diverse industries, driving demand for mobile device support. However, existing mobile inference frameworks often rely on a single processor per model, limiting hardware utilization and causing suboptimal performance and energy efficiency. Expanding DNN accessibility on mobile platforms requires adaptive, resource-efficient solutions to meet rising computational needs without compromising functionality. Parallel inference of multiple DNNs on heterogeneous processors remains challenging. Some works partition DNN operations into subgraphs for parallel execution across processors, but these often create excessive subgraphs based only on hardware compatibility, increasing scheduling complexity and memory overhead. To address this, we propose an Advanced Multi-DNN Model Scheduling (ADMS) strategy for optimizing multi-DNN inference on mobile heterogeneous processors. ADMS constructs an optimal subgraph partitioning strategy offline, balancing hardware operation support and scheduling granularity, and uses a processor-state-aware algorithm to dynamically adjust workloads based on real-time conditions. This ensures efficient workload distribution and maximizes processor utilization. Experiments show ADMS reduces multi-DNN inference latency by 4.04 times compared to vanilla frameworks.
comment: 14 pages, 12 figures, 5 tables
☆ Alleviating LLM-based Generative Retrieval Hallucination in Alipay Search
Generative retrieval (GR) has revolutionized document retrieval with the advent of large language models (LLMs), and LLM-based GR is gradually being adopted by the industry. Despite its remarkable advantages and potential, LLM-based GR suffers from hallucination and generates documents that are irrelevant to the query in some instances, severely challenging its credibility in practical applications. We thereby propose an optimized GR framework designed to alleviate retrieval hallucination, which integrates knowledge distillation reasoning in model training and incorporate decision agent to further improve retrieval precision. Specifically, we employ LLMs to assess and reason GR retrieved query-document (q-d) pairs, and then distill the reasoning data as transferred knowledge to the GR model. Moreover, we utilize a decision agent as post-processing to extend the GR retrieved documents through retrieval model and select the most relevant ones from multi perspectives as the final generative retrieval result. Extensive offline experiments on real-world datasets and online A/B tests on Fund Search and Insurance Search in Alipay demonstrate our framework's superiority and effectiveness in improving search quality and conversion gains.
comment: 4 pages
☆ Confidence Adjusted Surprise Measure for Active Resourceful Trials (CA-SMART): A Data-driven Active Learning Framework for Accelerating Material Discovery under Resource Constraints
Accelerating the discovery and manufacturing of advanced materials with specific properties is a critical yet formidable challenge due to vast search space, high costs of experiments, and time-intensive nature of material characterization. In recent years, active learning, where a surrogate machine learning (ML) model mimics the scientific discovery process of a human scientist, has emerged as a promising approach to address these challenges by guiding experimentation toward high-value outcomes with a limited budget. Among the diverse active learning philosophies, the concept of surprise (capturing the divergence between expected and observed outcomes) has demonstrated significant potential to drive experimental trials and refine predictive models. Scientific discovery often stems from surprise thereby making it a natural driver to guide the search process. Despite its promise, prior studies leveraging surprise metrics such as Shannon and Bayesian surprise lack mechanisms to account for prior confidence, leading to excessive exploration of uncertain regions that may not yield useful information. To address this, we propose the Confidence-Adjusted Surprise Measure for Active Resourceful Trials (CA-SMART), a novel Bayesian active learning framework tailored for optimizing data-driven experimentation. On a high level, CA-SMART incorporates Confidence-Adjusted Surprise (CAS) to dynamically balance exploration and exploitation by amplifying surprises in regions where the model is more certain while discounting them in highly uncertain areas. We evaluated CA-SMART on two benchmark functions (Six-Hump Camelback and Griewank) and in predicting the fatigue strength of steel. The results demonstrate superior accuracy and efficiency compared to traditional surprise metrics, standard Bayesian Optimization (BO) acquisition functions and conventional ML methods.
☆ ZJUKLAB at SemEval-2025 Task 4: Unlearning via Model Merging
This paper presents the ZJUKLAB team's submission for SemEval-2025 Task 4: Unlearning Sensitive Content from Large Language Models. This task aims to selectively erase sensitive knowledge from large language models, avoiding both over-forgetting and under-forgetting issues. We propose an unlearning system that leverages Model Merging (specifically TIES-Merging), combining two specialized models into a more balanced unlearned model. Our system achieves competitive results, ranking second among 26 teams, with an online score of 0.944 for Task Aggregate and 0.487 for overall Aggregate. In this paper, we also conduct local experiments and perform a comprehensive analysis of the unlearning process, examining performance trajectories, loss dynamics, and weight perspectives, along with several supplementary experiments, to understand the effectiveness of our method. Furthermore, we analyze the shortcomings of our method and evaluation metrics, emphasizing that MIA scores and ROUGE-based metrics alone are insufficient to fully evaluate successful unlearning. Finally, we emphasize the need for more comprehensive evaluation methodologies and rethinking of unlearning objectives in future research. Code is available at https://github.com/zjunlp/unlearn/tree/main/semeval25.
comment: Work in progress
☆ Rerouting Connection: Hybrid Computer Vision Analysis Reveals Visual Similarity Between Indus and Tibetan-Yi Corridor Writing Systems
This thesis employs a hybrid CNN-Transformer architecture, in conjunction with a detailed anthropological framework, to investigate potential historical connections between the visual morphology of the Indus Valley script and pictographic systems of the Tibetan-Yi Corridor. Through an ensemble methodology of three target scripts across 15 independently trained models, we demonstrate that Tibetan-Yi Corridor scripts exhibit approximately six-fold higher visual similarity to the Indus script (61.7%-63.5%) than to the Bronze Age Proto-Cuneiform (10.2%-10.9%) or Proto-Elamite (7.6%-8.7%) systems. Additionally and contrarily to our current understanding of the networks of the Indus Valley Civilization, the Indus script unexpectedly maps closer to Tibetan-Yi Corridor scripts, with a mean cosine similarity of 0.629, than to the aforementioned contemporaneous West Asian signaries, both of which recorded mean cosine similarities of 0.104 and 0.080 despite their close geographic proximity and evident trade relations. Across various dimensionality reduction practices and clustering methodologies, the Indus script consistently clusters closest to Tibetan-Yi Corridor scripts. Our computational results align with qualitative observations of specific pictorial parallels in numeral systems, gender markers, and key iconographic elements; this is further supported by archaeological evidence of sustained contact networks along the ancient Shu-Shendu road in tandem with the Indus Valley Civilization's decline, providing a plausible transmission pathway. While alternative explanations cannot be ruled out, the specificity and consistency of observed similarities challenge conventional narratives of isolated script development and suggest more complex ancient cultural transmission networks between South and East Asia than previously recognized.
comment: 106 pages total (main text: 42, 48 w/refs, 100 w/appendices). 21 figures, 4 tables in main; 106 figs, 8 tables total. Code and data at this URL: https://github.com/oohalakkadi/ivc2tyc. Submitted as undergrad thesis at Duke Kunshan University; accepted for presentation at the 2025 Computer Applications and Quantitative Methods in Archaeology Conference, Athens
☆ AskSport: Web Application for Sports Question-Answering
This paper introduces AskSport, a question-answering web application about sports. It allows users to ask questions using natural language and retrieve the three most relevant answers, including related information and documents. The paper describes the characteristics and functionalities of the application, including use cases demonstrating its ability to return names and numerical values. AskSport and its implementation are available for public access on HuggingFace.
comment: for accessing the application, see https://huggingface.co/spaces/leomaurodesenv/qasports-website
♻ ☆ GenoTEX: A Benchmark for Automated Gene Expression Data Analysis in Alignment with Bioinformaticians
Recent advancements in machine learning have significantly improved the identification of disease-associated genes from gene expression datasets. However, these processes often require extensive expertise and manual effort, limiting their scalability. Large Language Model (LLM)-based agents have shown promise in automating these tasks due to their increasing problem-solving abilities. To support the evaluation and development of such methods, we introduce GenoTEX, a benchmark dataset for the automated analysis of gene expression data. GenoTEX provides annotated code and results for solving a wide range of gene identification problems, encompassing dataset selection, preprocessing, and statistical analysis, in a pipeline that follows computational genomics standards. The benchmark includes expert-curated annotations from bioinformaticians to ensure accuracy and reliability. To provide baselines for these tasks, we present GenoAgent, a team of LLM-based agents that adopt a multi-step programming workflow with flexible self-correction, to collaboratively analyze gene expression datasets. Our experiments demonstrate the potential of LLM-based methods in analyzing genomic data, while error analysis highlights the challenges and areas for future improvement. We propose GenoTEX as a promising resource for benchmarking and enhancing automated methods for gene expression data analysis. The benchmark is available at https://github.com/Liu-Hy/GenoTex.
comment: 29 pages, 3 figures
♻ ☆ VIA: Unified Spatiotemporal Video Adaptation Framework for Global and Local Video Editing
Video editing serves as a fundamental pillar of digital media, spanning applications in entertainment, education, and professional communication. However, previous methods often overlook the necessity of comprehensively understanding both global and local contexts, leading to inaccurate and inconsistent edits in the spatiotemporal dimension, especially for long videos. In this paper, we introduce VIA, a unified spatiotemporal Video Adaptation framework for global and local video editing, pushing the limits of consistently editing minute-long videos. First, to ensure local consistency within individual frames, we designed test-time editing adaptation to adapt a pre-trained image editing model for improving consistency between potential editing directions and the text instruction, and adapts masked latent variables for precise local control. Furthermore, to maintain global consistency over the video sequence, we introduce spatiotemporal adaptation that recursively gather consistent attention variables in key frames and strategically applies them across the whole sequence to realize the editing effects. Extensive experiments demonstrate that, compared to baseline methods, our VIA approach produces edits that are more faithful to the source videos, more coherent in the spatiotemporal context, and more precise in local control. More importantly, we show that VIA can achieve consistent long video editing in minutes, unlocking the potential for advanced video editing tasks over long video sequences.
comment: 18 pages, 16 figures
♻ ☆ OVO-Bench: How Far is Your Video-LLMs from Real-World Online Video Understanding? CVPR 2025
Temporal Awareness, the ability to reason dynamically based on the timestamp when a question is raised, is the key distinction between offline and online video LLMs. Unlike offline models, which rely on complete videos for static, post hoc analysis, online models process video streams incrementally and dynamically adapt their responses based on the timestamp at which the question is posed. Despite its significance, temporal awareness has not been adequately evaluated in existing benchmarks. To fill this gap, we present OVO-Bench (Online-VideO-Benchmark), a novel video benchmark that emphasizes the importance of timestamps for advanced online video understanding capability benchmarking. OVO-Bench evaluates the ability of video LLMs to reason and respond to events occurring at specific timestamps under three distinct scenarios: (1) Backward tracing: trace back to past events to answer the question. (2) Real-time understanding: understand and respond to events as they unfold at the current timestamp. (3) Forward active responding: delay the response until sufficient future information becomes available to answer the question accurately. OVO-Bench comprises 12 tasks, featuring 644 unique videos and approximately human-curated 2,800 fine-grained meta-annotations with precise timestamps. We combine automated generation pipelines with human curation. With these high-quality samples, we further developed an evaluation pipeline to systematically query video LLMs along the video timeline. Evaluations of nine Video-LLMs reveal that, despite advancements on traditional benchmarks, current models struggle with online video understanding, showing a significant gap compared to human agents. We hope OVO-Bench will drive progress in video LLMs and inspire future research in online video reasoning. Our benchmark and code can be accessed at https://github.com/JoeLeelyf/OVO-Bench.
comment: CVPR 2025
♻ ☆ Multi-View and Multi-Scale Alignment for Contrastive Language-Image Pre-training in Mammography
Contrastive Language-Image Pre-training (CLIP) demonstrates strong potential in medical image analysis but requires substantial data and computational resources. Due to these restrictions, existing CLIP applications in medical imaging focus mainly on modalities like chest X-rays that have abundant image-report data available, leaving many other important modalities underexplored. Here, we propose one of the first adaptations of the full CLIP model to mammography, which presents significant challenges due to labeled data scarcity, high-resolution images with small regions of interest, and class-wise imbalance. We first develop a specialized supervision framework for mammography that leverages its multi-view nature. Furthermore, we design a symmetric local alignment module to better focus on detailed features in high-resolution images. Lastly, we incorporate a parameter-efficient fine-tuning approach for large language models pre-trained with medical knowledge to address data limitations. Our multi-view and multi-scale alignment (MaMA) method outperforms state-of-the-art baselines for three different tasks on two large real-world mammography datasets, EMBED and RSNA-Mammo, with only 52% model size compared with the largest baseline. The code is available at https://github.com/XYPB/MaMA
comment: This paper is accepted by IPMI 2025 for Oral Presentation
♻ ☆ TripoSG: High-Fidelity 3D Shape Synthesis using Large-Scale Rectified Flow Models
Recent advancements in diffusion techniques have propelled image and video generation to unprecedented levels of quality, significantly accelerating the deployment and application of generative AI. However, 3D shape generation technology has so far lagged behind, constrained by limitations in 3D data scale, complexity of 3D data processing, and insufficient exploration of advanced techniques in the 3D domain. Current approaches to 3D shape generation face substantial challenges in terms of output quality, generalization capability, and alignment with input conditions. We present TripoSG, a new streamlined shape diffusion paradigm capable of generating high-fidelity 3D meshes with precise correspondence to input images. Specifically, we propose: 1) A large-scale rectified flow transformer for 3D shape generation, achieving state-of-the-art fidelity through training on extensive, high-quality data. 2) A hybrid supervised training strategy combining SDF, normal, and eikonal losses for 3D VAE, achieving high-quality 3D reconstruction performance. 3) A data processing pipeline to generate 2 million high-quality 3D samples, highlighting the crucial rules for data quality and quantity in training 3D generative models. Through comprehensive experiments, we have validated the effectiveness of each component in our new framework. The seamless integration of these parts has enabled TripoSG to achieve state-of-the-art performance in 3D shape generation. The resulting 3D shapes exhibit enhanced detail due to high-resolution capabilities and demonstrate exceptional fidelity to input images. Moreover, TripoSG demonstrates improved versatility in generating 3D models from diverse image styles and contents, showcasing strong generalization capabilities. To foster progress and innovation in the field of 3D generation, we will make our model publicly available.
♻ ☆ Adaptive Orchestration for Large-Scale Inference on Heterogeneous Accelerator Systems Balancing Cost, Performance, and Resilience
The surge in generative AI workloads has created a need for scalable inference systems that can flexibly harness both GPUs and specialized accelerators while containing operational costs. This paper proposes a hardware-agnostic control loop that adaptively allocates requests across heterogeneous accelerators based on real-time cost and capacity signals. The approach sustains low latency and high throughput by dynamically shifting between cost-optimized and capacity-optimized modes, ensuring the most efficient use of expensive compute resources under fluctuating availability. Evaluated using the Stable Diffusion model, the framework consistently meets latency targets, automatically redirects traffic during capacity shortfalls, and capitalizes on lower-cost accelerators when possible. These results highlight how a feedback-driven deployment strategy, spanning the entire software and hardware stack, can help organizations efficiently scale generative AI workloads while maintaining resilience in the face of limited accelerator capacity.
comment: 14 pages, 7 figures
♻ ☆ TSKANMixer: Kolmogorov-Arnold Networks with MLP-Mixer Model for Time Series Forecasting AAAI 2025
Time series forecasting has long been a focus of research across diverse fields, including economics, energy, healthcare, and traffic management. Recent works have introduced innovative architectures for time series models, such as the Time-Series Mixer (TSMixer), which leverages multi-layer perceptrons (MLPs) to enhance prediction accuracy by effectively capturing both spatial and temporal dependencies within the data. In this paper, we investigate the capabilities of the Kolmogorov-Arnold Networks (KANs) for time-series forecasting by modifying TSMixer with a KAN layer (TSKANMixer). Experimental results demonstrate that TSKANMixer tends to improve prediction accuracy over the original TSMixer across multiple datasets, ranking among the top-performing models compared to other time series approaches. Our results show that the KANs are promising alternatives to improve the performance of time series forecasting by replacing or extending traditional MLPs.
comment: 8 pages, 4 figures, 7 tables and accepted at the AI4TS: AI for Time Series Analysis workshop, AAAI 2025
♻ ☆ OmniBench: Towards The Future of Universal Omni-Language Models
Recent advancements in multimodal large language models (MLLMs) have focused on integrating multiple modalities, yet their ability to simultaneously process and reason across different inputs remains underexplored. We introduce OmniBench, a novel benchmark designed to evaluate models' ability to recognize, interpret, and reason across visual, acoustic, and textual inputs simultaneously. We define language models capable of such tri-modal processing as omni-language models (OLMs). OmniBench features high-quality human annotations that require integrated understanding across all modalities. Our evaluation reveals that: i) open-source OLMs show significant limitations in instruction-following and reasoning in tri-modal contexts; and ii) most baseline models perform poorly (around 50% accuracy) even with textual alternatives to image/audio inputs. To address these limitations, we develop OmniInstruct, an 96K-sample instruction tuning dataset for training OLMs. We advocate for developing more robust tri-modal integration techniques and training strategies to enhance OLM performance. Codes and data could be found at our repo (https://github.com/multimodal-art-projection/OmniBench).
♻ ☆ CleanGen: Mitigating Backdoor Attacks for Generation Tasks in Large Language Models EMNLP 2024
The remarkable performance of large language models (LLMs) in generation tasks has enabled practitioners to leverage publicly available models to power custom applications, such as chatbots and virtual assistants. However, the data used to train or fine-tune these LLMs is often undisclosed, allowing an attacker to compromise the data and inject backdoors into the models. In this paper, we develop a novel inference time defense, named CLEANGEN, to mitigate backdoor attacks for generation tasks in LLMs. CLEANGEN is a lightweight and effective decoding strategy that is compatible with the state-of-the-art (SOTA) LLMs. Our insight behind CLEANGEN is that compared to other LLMs, backdoored LLMs assign significantly higher probabilities to tokens representing the attacker-desired contents. These discrepancies in token probabilities enable CLEANGEN to identify suspicious tokens favored by the attacker and replace them with tokens generated by another LLM that is not compromised by the same attacker, thereby avoiding generation of attacker-desired content. We evaluate CLEANGEN against five SOTA backdoor attacks. Our results show that CLEANGEN achieves lower attack success rates (ASR) compared to five SOTA baseline defenses for all five backdoor attacks. Moreover, LLMs deploying CLEANGEN maintain helpfulness in their responses when serving benign user queries with minimal added computational overhead.
comment: This paper is presented at EMNLP 2024
♻ ☆ Vision language models are blind: Failing to translate detailed visual features into words
While large language models with vision capabilities (VLMs), e.g., GPT-4o and Gemini 1.5 Pro, score high on many vision-understanding benchmarks, they are still struggling with low-level vision tasks that are easy to humans. Specifically, on BlindTest, our suite of 7 very simple tasks, including identifying (a) whether two circles overlap; (b) how many times two lines intersect; (c) which letter is being circled in a word; and (d) the number of circles in an Olympic-like logo, four state-of-the-art VLMs are only 58.07% accurate on average. Claude 3.5 Sonnet performs the best at 77.84% accuracy, far from the human expected accuracy of 100%. Across different image resolutions and line widths, VLMs including slow-thinking models consistently struggle with those tasks that require precise spatial information when geometric primitives overlap or are close. Yet, VLMs perform at near-100% accuracy when much more space is added to separate shapes and letters. Linear probing experiments show that vision encoders contain sufficient visual information to solve BlindTest and that language models fail to decode this information into correct answers. Code and data are at: https://vlmsareblind.github.io
♻ ☆ Enhancing LLM Character-Level Manipulation via Divide and Conquer
Large Language Models (LLMs) have demonstrated strong generalization capabilities across a wide range of natural language processing (NLP) tasks. However, they exhibit notable weaknesses in character-level string manipulation, struggling with fundamental operations such as character deletion, insertion, and substitution. These challenges stem primarily from tokenization constraints, despite the critical role of such operations in data preprocessing and code generation. Through systematic analysis, we derive two key insights: (1) LLMs face significant difficulties in leveraging intrinsic token knowledge for character-level reasoning, and (2) atomized word structures can substantially enhance LLMs' ability to process token-level structural information. Building on these insights, we propose Character-Level Manipulation via Divide and Conquer, a novel approach designed to bridge the gap between token-level processing and character-level manipulation. Our method decomposes complex operations into explicit character-level subtasks coupled with controlled token reconstruction phases, leading to significant improvements in accuracy. Without additional training, our method significantly improves accuracies on the $\texttt{Deletion}$, $\texttt{Insertion}$, and $\texttt{Substitution}$ tasks. To support further research, we open-source our implementation and benchmarks.
♻ ☆ Self-Contrastive Forward-Forward Algorithm
Agents that operate autonomously benefit from lifelong learning capabilities. However, compatible training algorithms must comply with the decentralized nature of these systems, which imposes constraints on both the parameter counts and the computational resources. The Forward-Forward (FF) algorithm is one of these. FF relies only on feedforward operations, the same used for inference, for optimizing layer-wise objectives. This purely forward approach eliminates the need for transpose operations required in traditional backpropagation. Despite its potential, FF has failed to reach state-of-the-art performance on most standard benchmark tasks, in part due to unreliable negative data generation methods for unsupervised learning. In this work, we propose the Self-Contrastive Forward-Forward (SCFF) algorithm, a competitive training method aimed at closing this performance gap. Inspired by standard self-supervised contrastive learning for vision tasks, SCFF generates positive and negative inputs applicable across various datasets. The method demonstrates superior performance compared to existing unsupervised local learning algorithms on several benchmark datasets, including MNIST, CIFAR-10, STL-10, and Tiny ImageNet. We extend FF's application to training recurrent neural networks, expanding its utility to sequential data tasks. These findings pave the way for high-accuracy, real-time learning on resource-constrained edge devices.
♻ ☆ Layer- and Timestep-Adaptive Differentiable Token Compression Ratios for Efficient Diffusion Transformers CVPR 2025
Diffusion Transformers (DiTs) have achieved state-of-the-art (SOTA) image generation quality but suffer from high latency and memory inefficiency, making them difficult to deploy on resource-constrained devices. One major efficiency bottleneck is that existing DiTs apply equal computation across all regions of an image. However, not all image tokens are equally important, and certain localized areas require more computation, such as objects. To address this, we propose DiffCR, a dynamic DiT inference framework with differentiable compression ratios, which automatically learns to dynamically route computation across layers and timesteps for each image token, resulting in efficient DiTs. Specifically, DiffCR integrates three features: (1) A token-level routing scheme where each DiT layer includes a router that is fine-tuned jointly with model weights to predict token importance scores. In this way, unimportant tokens bypass the entire layer's computation; (2) A layer-wise differentiable ratio mechanism where different DiT layers automatically learn varying compression ratios from a zero initialization, resulting in large compression ratios in redundant layers while others remain less compressed or even uncompressed; (3) A timestep-wise differentiable ratio mechanism where each denoising timestep learns its own compression ratio. The resulting pattern shows higher ratios for noisier timesteps and lower ratios as the image becomes clearer. Extensive experiments on text-to-image and inpainting tasks show that DiffCR effectively captures dynamism across token, layer, and timestep axes, achieving superior trade-offs between generation quality and efficiency compared to prior works. The project website is available at https://www.haoranyou.com/diffcr.
comment: Accepted by CVPR 2025
♻ ☆ Learning Multi-modal Representations by Watching Hundreds of Surgical Video Lectures
Recent advancements in surgical computer vision applications have been driven by vision-only models, which do not explicitly integrate the rich semantics of language into their design. These methods rely on manually annotated surgical videos to predict a fixed set of object categories, limiting their generalizability to unseen surgical procedures and downstream tasks. In this work, we put forward the idea that the surgical video lectures available through open surgical e-learning platforms can provide effective vision and language supervisory signals for multi-modal representation learning without relying on manual annotations. We address the surgery-specific linguistic challenges present in surgical video lectures by employing multiple complementary automatic speech recognition systems to generate text transcriptions. We then present a novel method, SurgVLP - Surgical Vision Language Pre-training, for multi-modal representation learning. Extensive experiments across diverse surgical procedures and tasks demonstrate that the multi-modal representations learned by SurgVLP exhibit strong transferability and adaptability in surgical video analysis. Furthermore, our zero-shot evaluations highlight SurgVLP's potential as a general-purpose foundation model for surgical workflow analysis, reducing the reliance on extensive manual annotations for downstream tasks, and facilitating adaptation methods such as few-shot learning to build a scalable and data-efficient solution for various downstream surgical applications. The [training code](https://github.com/CAMMA-public/SurgVLP) and [weights](https://github.com/CAMMA-public/PeskaVLP) are public.
♻ ☆ TREAD: Token Routing for Efficient Architecture-agnostic Diffusion Training
Diffusion models have emerged as the mainstream approach for visual generation. However, these models typically suffer from sample inefficiency and high training costs. Consequently, methods for efficient finetuning, inference and personalization were quickly adopted by the community. However, training these models in the first place remains very costly. While several recent approaches - including masking, distillation, and architectural modifications - have been proposed to improve training efficiency, each of these methods comes with a tradeoff: they achieve enhanced performance at the expense of increased computational cost or vice versa. In contrast, this work aims to improve training efficiency as well as generative performance at the same time through routes that act as a transport mechanism for randomly selected tokens from early layers to deeper layers of the model. Our method is not limited to the common transformer-based model - it can also be applied to state-space models and achieves this without architectural modifications or additional parameters. Finally, we show that TREAD reduces computational cost and simultaneously boosts model performance on the standard ImageNet-256 benchmark in class-conditional synthesis. Both of these benefits multiply to a convergence speedup of 14x at 400K training iterations compared to DiT and 37x compared to the best benchmark performance of DiT at 7M training iterations. Furthermore, we achieve a competitive FID of 2.09 in a guided and 3.93 in an unguided setting, which improves upon the DiT, without architectural changes.
♻ ☆ Robust Counterfactual Inference in Markov Decision Processes
This paper addresses a key limitation in existing counterfactual inference methods for Markov Decision Processes (MDPs). Current approaches assume a specific causal model to make counterfactuals identifiable. However, there are usually many causal models that align with the observational and interventional distributions of an MDP, each yielding different counterfactual distributions, so fixing a particular causal model limits the validity (and usefulness) of counterfactual inference. We propose a novel non-parametric approach that computes tight bounds on counterfactual transition probabilities across all compatible causal models. Unlike previous methods that require solving prohibitively large optimisation problems (with variables that grow exponentially in the size of the MDP), our approach provides closed-form expressions for these bounds, making computation highly efficient and scalable for non-trivial MDPs. Once such an interval counterfactual MDP is constructed, our method identifies robust counterfactual policies that optimise the worst-case reward w.r.t. the uncertain interval MDP probabilities. We evaluate our method on various case studies, demonstrating improved robustness over existing methods.
comment: Fixed typo in Equation (5)
♻ ☆ Counterfactual Influence in Markov Decision Processes
Our work addresses a fundamental problem in the context of counterfactual inference for Markov Decision Processes (MDPs). Given an MDP path $\tau$, this kind of inference allows us to derive counterfactual paths $\tau'$ describing what-if versions of $\tau$ obtained under different action sequences than those observed in $\tau$. However, as the counterfactual states and actions deviate from the observed ones over time, the observation $\tau$ may no longer influence the counterfactual world, meaning that the analysis is no longer tailored to the individual observation, resulting in interventional outcomes rather than counterfactual ones. Even though this issue specifically affects the popular Gumbel-max structural causal model used for MDP counterfactuals, it has remained overlooked until now. In this work, we introduce a formal characterisation of influence based on comparing counterfactual and interventional distributions. We devise an algorithm to construct counterfactual models that automatically satisfy influence constraints. Leveraging such models, we derive counterfactual policies that are not just optimal for a given reward structure but also remain tailored to the observed path. Even though there is an unavoidable trade-off between policy optimality and strength of influence constraints, our experiments demonstrate that it is possible to derive (near-)optimal policies while remaining under the influence of the observation.
comment: 12 pages, 6 figures
♻ ☆ Dynamics-Aware Gaussian Splatting Streaming Towards Fast On-the-Fly 4D Reconstruction
The recent development of 3D Gaussian Splatting (3DGS) has led to great interest in 4D dynamic spatial reconstruction. Existing approaches mainly rely on full-length multi-view videos, while there has been limited exploration of online reconstruction methods that enable on-the-fly training and per-timestep streaming. Current 3DGS-based streaming methods treat the Gaussian primitives uniformly and constantly renew the densified Gaussians, thereby overlooking the difference between dynamic and static features as well as neglecting the temporal continuity in the scene. To address these limitations, we propose a novel three-stage pipeline for iterative streamable 4D dynamic spatial reconstruction. Our pipeline comprises a selective inheritance stage to preserve temporal continuity, a dynamics-aware shift stage to distinguish dynamic and static primitives and optimize their movements, and an error-guided densification stage to accommodate emerging objects. Our method achieves state-of-the-art performance in online 4D reconstruction, demonstrating the fastest on-the-fly training, superior representation quality, and real-time rendering capability. Project page: https://www.liuzhening.top/DASS
comment: Project page: https://www.liuzhening.top/DASS
♻ ☆ Debiased Offline Representation Learning for Fast Online Adaptation in Non-stationary Dynamics
Developing policies that can adjust to non-stationary environments is essential for real-world reinforcement learning applications. However, learning such adaptable policies in offline settings, with only a limited set of pre-collected trajectories, presents significant challenges. A key difficulty arises because the limited offline data makes it hard for the context encoder to differentiate between changes in the environment dynamics and shifts in the behavior policy, often leading to context misassociations. To address this issue, we introduce a novel approach called Debiased Offline Representation for fast online Adaptation (DORA). DORA incorporates an information bottleneck principle that maximizes mutual information between the dynamics encoding and the environmental data, while minimizing mutual information between the dynamics encoding and the actions of the behavior policy. We present a practical implementation of DORA, leveraging tractable bounds of the information bottleneck principle. Our experimental evaluation across six benchmark MuJoCo tasks with variable parameters demonstrates that DORA not only achieves a more precise dynamics encoding but also significantly outperforms existing baselines in terms of performance.
♻ ☆ Deep Cut-informed Graph Embedding and Clustering
Graph clustering aims to divide the graph into different clusters. The recently emerging deep graph clustering approaches are largely built on graph neural networks (GNN). However, GNN is designed for general graph encoding and there is a common issue of representation collapse in existing GNN-based deep graph clustering algorithms. We attribute two main reasons for such issues: (i) the inductive bias of GNN models: GNNs tend to generate similar representations for proximal nodes. Since graphs often contain a non-negligible amount of inter-cluster links, the bias results in error message passing and leads to biased clustering; (ii) the clustering guided loss function: most traditional approaches strive to make all samples closer to pre-learned cluster centers, which causes a degenerate solution assigning all data points to a single label thus make all samples and less discriminative. To address these challenges, we investigate graph clustering from a graph cut perspective and propose an innovative and non-GNN-based Deep Cut-informed Graph embedding and Clustering framework, namely DCGC. This framework includes two modules: (i) cut-informed graph encoding; (ii) self-supervised graph clustering via optimal transport. For the encoding module, we derive a cut-informed graph embedding objective to fuse graph structure and attributes by minimizing their joint normalized cut. For the clustering module, we utilize the optimal transport theory to obtain the clustering assignments, which can balance the guidance of "proximity to the pre-learned cluster center". With the above two tailored designs, DCGC is more suitable for the graph clustering task, which can effectively alleviate the problem of representation collapse and achieve better performance. We conduct extensive experiments to demonstrate that our method is simple but effective compared with benchmarks.
♻ ☆ A Survey on Self-play Methods in Reinforcement Learning
Self-play, characterized by agents' interactions with copies or past versions of themselves, has recently gained prominence in reinforcement learning (RL). This paper first clarifies the preliminaries of self-play, including the multi-agent reinforcement learning framework and basic game theory concepts. Then, it provides a unified framework and classifies existing self-play algorithms within this framework. Moreover, the paper bridges the gap between the algorithms and their practical implications by illustrating the role of self-play in different scenarios. Finally, the survey highlights open challenges and future research directions in self-play. This paper is an essential guide map for understanding the multifaceted landscape of self-play in RL.
♻ ☆ Efficient Continual Adaptation of Pretrained Robotic Policy with Online Meta-Learned Adapters
Continual adaptation is essential for general autonomous agents. For example, a household robot pretrained with a repertoire of skills must still adapt to unseen tasks specific to each household. Motivated by this, building upon parameter-efficient fine-tuning in language models, prior works have explored lightweight adapters to adapt pretrained policies, which can preserve learned features from the pretraining phase and demonstrate good adaptation performances. However, these approaches treat task learning separately, limiting knowledge transfer between tasks. In this paper, we propose Online Meta-Learned adapters (OMLA). Instead of applying adapters directly, OMLA can facilitate knowledge transfer from previously learned tasks to current learning tasks through a novel meta-learning objective. Extensive experiments in both simulated and real-world environments demonstrate that OMLA can lead to better adaptation performances compared to the baseline methods. The project link: https://ricky-zhu.github.io/OMLA/.
comment: Project link: https://ricky-zhu.github.io/OMLA/
♻ ☆ Pretraining with random noise for uncertainty calibration
Uncertainty calibration is crucial for various machine learning applications, yet it remains challenging. Many models exhibit hallucinations - confident yet inaccurate responses - due to miscalibrated confidence. Here, we show that the common practice of random initialization in deep learning, often considered a standard technique, is an underlying cause of this miscalibration, leading to excessively high confidence in untrained networks. Our method, inspired by developmental neuroscience, addresses this issue by simply pretraining networks with random noise and labels, reducing overconfidence and bringing initial confidence levels closer to chance. This ensures optimal calibration, aligning confidence with accuracy during subsequent data training, without the need for additional pre- or post-processing. Pre-calibrated networks excel at identifying "unknown data," showing low confidence for out-of-distribution inputs, thereby resolving confidence miscalibration.
♻ ☆ Video Motion Transfer with Diffusion Transformers CVPR 2025
We propose DiTFlow, a method for transferring the motion of a reference video to a newly synthesized one, designed specifically for Diffusion Transformers (DiT). We first process the reference video with a pre-trained DiT to analyze cross-frame attention maps and extract a patch-wise motion signal called the Attention Motion Flow (AMF). We guide the latent denoising process in an optimization-based, training-free, manner by optimizing latents with our AMF loss to generate videos reproducing the motion of the reference one. We also apply our optimization strategy to transformer positional embeddings, granting us a boost in zero-shot motion transfer capabilities. We evaluate DiTFlow against recently published methods, outperforming all across multiple metrics and human evaluation.
comment: CVPR 2025 - Project page: https://ditflow.github.io/
♻ ☆ Rethinking Video Tokenization: A Conditioned Diffusion-based Approach
Existing video tokenizers typically use the traditional Variational Autoencoder (VAE) architecture for video compression and reconstruction. However, to achieve good performance, its training process often relies on complex multi-stage training tricks that go beyond basic reconstruction loss and KL regularization. Among these tricks, the most challenging is the precise tuning of adversarial training with additional Generative Adversarial Networks (GANs) in the final stage, which can hinder stable convergence. In contrast to GANs, diffusion models offer more stable training processes and can generate higher-quality results. Inspired by these advantages, we propose CDT, a novel Conditioned Diffusion-based video Tokenizer, that replaces the GAN-based decoder with a conditional causal diffusion model. The encoder compresses spatio-temporal information into compact latents, while the decoder reconstructs videos through a reverse diffusion process conditioned on these latents. During inference, we incorporate a feature cache mechanism to generate videos of arbitrary length while maintaining temporal continuity and adopt sampling acceleration technique to enhance efficiency. Trained using only a basic MSE diffusion loss for reconstruction, along with KL term and LPIPS perceptual loss from scratch, extensive experiments demonstrate that CDT achieves state-of-the-art performance in video reconstruction tasks with just a single-step sampling. Even a scaled-down version of CDT (3$\times$ inference speedup) still performs comparably with top baselines. Moreover, the latent video generation model trained with CDT also exhibits superior performance. The source code and pretrained weights are available at https://github.com/ali-vilab/CDT.
♻ ☆ Temporal-Guided Spiking Neural Networks for Event-Based Human Action Recognition
This paper explores the promising interplay between spiking neural networks (SNNs) and event-based cameras for privacy-preserving human action recognition (HAR). The unique feature of event cameras in capturing only the outlines of motion, combined with SNNs' proficiency in processing spatiotemporal data through spikes, establishes a highly synergistic compatibility for event-based HAR. Previous studies, however, have been limited by SNNs' ability to process long-term temporal information, essential for precise HAR. In this paper, we introduce two novel frameworks to address this: temporal segment-based SNN (\textit{TS-SNN}) and 3D convolutional SNN (\textit{3D-SNN}). The \textit{TS-SNN} extracts long-term temporal information by dividing actions into shorter segments, while the \textit{3D-SNN} replaces 2D spatial elements with 3D components to facilitate the transmission of temporal information. To promote further research in event-based HAR, we create a dataset, \textit{FallingDetection-CeleX}, collected using the high-resolution CeleX-V event camera $(1280 \times 800)$, comprising 7 distinct actions. Extensive experimental results show that our proposed frameworks surpass state-of-the-art SNN methods on our newly collected dataset and three other neuromorphic datasets, showcasing their effectiveness in handling long-range temporal information for event-based HAR.
♻ ☆ Online POMDP Planning with Anytime Deterministic Guarantees
Decision-making under uncertainty is a critical aspect of many practical autonomous systems due to incomplete information. Partially Observable Markov Decision Processes (POMDPs) offer a mathematically principled framework for formulating decision-making problems under such conditions. However, finding an optimal solution for a POMDP is generally intractable. In recent years, there has been a significant progress of scaling approximate solvers from small to moderately sized problems, using online tree search solvers. Often, such approximate solvers are limited to probabilistic or asymptotic guarantees towards the optimal solution. In this paper, we derive a deterministic relationship for discrete POMDPs between an approximated and the optimal solution. We show that at any time, we can derive bounds that relate between the existing solution and the optimal one. We show that our derivations provide an avenue for a new set of algorithms and can be attached to existing algorithms that have a certain structure to provide them with deterministic guarantees with marginal computational overhead. In return, not only do we certify the solution quality, but we demonstrate that making a decision based on the deterministic guarantee may result in superior performance compared to the original algorithm without the deterministic certification.
♻ ☆ FaceID-6M: A Large-Scale, Open-Source FaceID Customization Dataset
Due to the data-driven nature of current face identity (FaceID) customization methods, all state-of-the-art models rely on large-scale datasets containing millions of high-quality text-image pairs for training. However, none of these datasets are publicly available, which restricts transparency and hinders further advancements in the field. To address this issue, in this paper, we collect and release FaceID-6M, the first large-scale, open-source FaceID dataset containing 6 million high-quality text-image pairs. Filtered from LAION-5B \cite{schuhmann2022laion}, FaceID-6M undergoes a rigorous image and text filtering steps to ensure dataset quality, including resolution filtering to maintain high-quality images and faces, face filtering to remove images that lack human faces, and keyword-based strategy to retain descriptions containing human-related terms (e.g., nationality, professions and names). Through these cleaning processes, FaceID-6M provides a high-quality dataset optimized for training powerful FaceID customization models, facilitating advancements in the field by offering an open resource for research and development. We conduct extensive experiments to show the effectiveness of our FaceID-6M, demonstrating that models trained on our FaceID-6M dataset achieve performance that is comparable to, and slightly better than currently available industrial models. Additionally, to support and advance research in the FaceID customization community, we make our code, datasets, and models fully publicly available. Our codes, models, and datasets are available at: https://github.com/ShuheSH/FaceID-6M.
comment: arXiv admin note: text overlap with arXiv:2501.15407
♻ ☆ Starjob: Dataset for LLM-Driven Job Shop Scheduling
Large Language Models (LLMs) have shown remarkable capabilities across various domains, but their potential for solving combinatorial optimization problems remains largely unexplored. In this paper, we investigate the applicability of LLMs to the Job Shop Scheduling Problem (JSSP), a classic challenge in combinatorial optimization that requires efficient job allocation to machines to minimize makespan. To this end, we introduce Starjob, the first supervised dataset for JSSP, comprising 130k instances specifically designed for training LLMs. Leveraging this dataset, we fine-tune the LLaMA 8B 4-bit quantized model with the LoRA method to develop an end-to-end scheduling approach. Our evaluation on standard benchmarks demonstrates that the proposed LLM-based method not only surpasses traditional Priority Dispatching Rules (PDRs) but also achieves notable improvements over state-of-the-art neural approaches like L2D, with an average improvement of 15.36% on DMU and 7.85% on Taillard benchmarks. These results highlight the untapped potential of LLMs in tackling combinatorial optimization problems, paving the way for future advancements in this area.
comment: arXiv admin note: substantial text overlap with arXiv:2408.06993
♻ ☆ A Logic for Reasoning About Aggregate-Combine Graph Neural Networks
We propose a modal logic in which counting modalities appear in linear inequalities. We show that each formula can be transformed into an equivalent graph neural network (GNN). We also show that a broad class of GNNs can be transformed efficiently into a formula, thus significantly improving upon the literature about the logical expressiveness of GNNs. We also show that the satisfiability problem is PSPACE-complete. These results bring together the promise of using standard logical methods for reasoning about GNNs and their properties, particularly in applications such as GNN querying, equivalence checking, etc. We prove that such natural problems can be solved in polynomial space.
comment: arXiv admin note: text overlap with arXiv:2307.05150
♻ ☆ Automatically Adaptive Conformal Risk Control
Science and technology have a growing need for effective mechanisms that ensure reliable, controlled performance from black-box machine learning algorithms. These performance guarantees should ideally hold conditionally on the input-that is the performance guarantees should hold, at least approximately, no matter what the input. However, beyond stylized discrete groupings such as ethnicity and gender, the right notion of conditioning can be difficult to define. For example, in problems such as image segmentation, we want the uncertainty to reflect the intrinsic difficulty of the test sample, but this may be difficult to capture via a conditioning event. Building on the recent work of Gibbs et al. [2023], we propose a methodology for achieving approximate conditional control of statistical risks-the expected value of loss functions-by adapting to the difficulty of test samples. Our framework goes beyond traditional conditional risk control based on user-provided conditioning events to the algorithmic, data-driven determination of appropriate function classes for conditioning. We apply this framework to various regression and segmentation tasks, enabling finer-grained control over model performance and demonstrating that by continuously monitoring and adjusting these parameters, we can achieve superior precision compared to conventional risk-control methods.
♻ ☆ RGB-Th-Bench: A Dense benchmark for Visual-Thermal Understanding of Vision Language Models
We introduce RGB-Th-Bench, the first benchmark designed to evaluate the ability of Vision-Language Models (VLMs) to comprehend RGB-Thermal image pairs. While VLMs have demonstrated remarkable progress in visual reasoning and multimodal understanding, their evaluation has been predominantly limited to RGB-based benchmarks, leaving a critical gap in assessing their capabilities in infrared vision tasks. Existing visible-infrared datasets are either task-specific or lack high-quality annotations necessary for rigorous model evaluation. To address these limitations, RGB-Th-Bench provides a comprehensive evaluation framework covering 14 distinct skill dimensions, with a total of 1,600+ expert-annotated Yes/No questions. The benchmark employs two accuracy metrics: a standard question-level accuracy and a stricter skill-level accuracy, which evaluates model robustness across multiple questions within each skill dimension. This design ensures a thorough assessment of model performance, including resilience to adversarial and hallucinated responses. We conduct extensive evaluations on 19 state-of-the-art VLMs, revealing significant performance gaps in RGB-Thermal understanding. Our results show that even the strongest models struggle with thermal image comprehension, with performance heavily constrained by their RGB-based capabilities. Additionally, the lack of large-scale application-specific and expert-annotated thermal-caption-pair datasets in pre-training is an important reason of the observed performance gap. RGB-Th-Bench highlights the urgent need for further advancements in multimodal learning to bridge the gap between visible and thermal image understanding. The dataset is available through this link, and the evaluation code will also be made publicly available.
♻ ☆ Dynamic Bi-Elman Attention Networks: A Dual-Directional Context-Aware Test-Time Learning for Text Classification
Text classification, a fundamental task in natural language processing, aims to categorize textual data into predefined labels. Traditional methods struggled with complex linguistic structures and semantic dependencies. However, the advent of deep learning, particularly recurrent neural networks and Transformer-based models, has significantly advanced the field by enabling nuanced feature extraction and context-aware predictions. Despite these improvements, existing models still exhibit limitations in balancing interpretability, computational efficiency, and long-range contextual understanding. To address these challenges, this paper proposes the Dynamic Bidirectional Elman with Attention Network (DBEAN). DBEAN integrates bidirectional temporal modeling with self-attention mechanisms. It dynamically assigns weights to critical segments of input, improving contextual representation while maintaining computational efficiency.
comment: 11 pages
♻ ☆ ATM: Improving Model Merging by Alternating Tuning and Merging
Model merging has recently emerged as a cost-efficient paradigm for multi-task learning. Among current approaches, task arithmetic stands out for its simplicity and effectiveness. In this paper, we motivate the effectiveness of task vectors by linking them to multi-task gradients. We show that in a single-epoch scenario, if the optimization is performed via gradient descent, task vectors are after one step mathematically equivalent to the gradients obtained via gradient descent in a multi-task setting, and still approximate these gradients in subsequent epochs. Furthermore, we show that the effectiveness of task vectors is largely driven by the first epoch's gradient. Given this parallel between task vectors and gradients, we propose viewing model merging as a single step in an iterative process that alternates between tuning and merging (ATM). We then propose two ways to utilize ATM. The first is to replace multi-task learning with ATM in scenarios where data sharing is prohibited, such as federated learning. The second is to improve the outcome of any model merging algorithm by applying a few post-hoc iterations of ATM on a small validation dataset, which is commonly available for hyperparameter tuning. Finally, we provide both empirical and theoretical support for the effectiveness of ATM, demonstrating that it minimizes an upper bound on the loss obtained by jointly finetuning all tasks.
comment: Main paper: 9 Pages, 9 figures, 1 table
♻ ☆ ProReason: Multi-Modal Proactive Reasoning with Decoupled Eyesight and Wisdom
Large vision-language models (LVLMs) have witnessed significant progress on visual understanding tasks. However, they often prioritize language knowledge over image information on visual reasoning tasks, incurring performance degradation. To tackle this issue, we first identify the drawbacks of existing solutions (i.e., insufficient and irrelevant visual descriptions, and limited multi-modal capacities). We then decompose visual reasoning process into two stages: visual perception (i.e., eyesight) and textual reasoning (i.e., wisdom), and introduce a novel visual reasoning framework named ProReason. This framework features multi-run proactive perception and decoupled vision-reasoning capabilities. Briefly, given a multi-modal question, ProReason iterates proactive information collection and reasoning until the answer can be concluded with necessary and sufficient visual descriptions. Notably, the disassociation of capabilities allows seamless integration of existing large language models (LLMs) to compensate for the reasoning deficits of LVLMs. Our extensive experiments demonstrate that ProReason outperforms both existing multi-step reasoning frameworks and passive peer methods on a wide range of benchmarks for both open-source and closed-source models. In addition, with the assistance of LLMs, ProReason achieves a performance improvement of up to 15% on MMMU benchmark. Our insights into existing solutions and the decoupled perspective for feasible integration of LLMs illuminate future research on visual reasoning techniques, especially LLM-assisted ones.
♻ ☆ Rethinking Training for De-biasing Text-to-Image Generation: Unlocking the Potential of Stable Diffusion CVPR 2025
Recent advancements in text-to-image models, such as Stable Diffusion, show significant demographic biases. Existing de-biasing techniques rely heavily on additional training, which imposes high computational costs and risks of compromising core image generation functionality. This hinders them from being widely adopted to real-world applications. In this paper, we explore Stable Diffusion's overlooked potential to reduce bias without requiring additional training. Through our analysis, we uncover that initial noises associated with minority attributes form "minority regions" rather than scattered. We view these "minority regions" as opportunities in SD to reduce bias. To unlock the potential, we propose a novel de-biasing method called 'weak guidance,' carefully designed to guide a random noise to the minority regions without compromising semantic integrity. Through analysis and experiments on various versions of SD, we demonstrate that our proposed approach effectively reduces bias without additional training, achieving both efficiency and preservation of core image generation functionality.
comment: 19 pages; First two authors contributed equally; Accepted at CVPR 2025
♻ ☆ R2-KG: General-Purpose Dual-Agent Framework for Reliable Reasoning on Knowledge Graphs
Recent studies have combined Large Language Models (LLMs) with Knowledge Graphs (KGs) to enhance reasoning, improving inference accuracy without additional training while mitigating hallucination. However, existing frameworks are often rigid, struggling to adapt to KG or task changes. They also rely heavily on powerful LLMs for reliable (i.e., trustworthy) reasoning. To address this, We introduce R2-KG, a plug-and-play, dual-agent framework that separates reasoning into two roles: an Operator (a low-capacity LLM) that gathers evidence and a Supervisor (a high-capacity LLM) that makes final judgments. This design is cost-efficient for LLM inference while still maintaining strong reasoning accuracy. Additionally, R2-KG employs an Abstention mechanism, generating answers only when sufficient evidence is collected from KG, which significantly enhances reliability. Experiments across multiple KG-based reasoning tasks show that R2-KG consistently outperforms baselines in both accuracy and reliability, regardless of the inherent capability of LLMs used as the Operator. Further experiments reveal that the single-agent version of R2-KG, equipped with a strict self-consistency strategy, achieves significantly higher-than-baseline reliability while reducing inference cost. However, it also leads to a higher abstention rate in complex KGs. Our findings establish R2-KG as a flexible and cost-effective solution for KG-based reasoning. It reduces reliance on high-capacity LLMs while ensuring trustworthy inference. The code is available at https://github.com/ekrxjwh2009/R2-KG/.
♻ ☆ Inductive-Associative Meta-learning Pipeline with Human Cognitive Patterns for Unseen Drug-Target Interaction Prediction
Significant differences in protein structures hinder the generalization of existing drug-target interaction (DTI) models, which often rely heavily on pre-learned binding principles or detailed annotations. In contrast, BioBridge designs an Inductive-Associative pipeline inspired by the workflow of scientists who base their accumulated expertise on drawing insights into novel drug-target pairs from weakly related references. BioBridge predicts novel drug-target interactions using limited sequence data, incorporating multi-level encoders with adversarial training to accumulate transferable binding principles. On these principles basis, BioBridge employs a dynamic prototype meta-learning framework to associate insights from weakly related annotations, enabling robust predictions for previously unseen drug-target pairs. Extensive experiments demonstrate that BioBridge surpasses existing models, especially for unseen proteins. Notably, when only homologous protein binding data is available, BioBridge proves effective for virtual screening of the epidermal growth factor receptor and adenosine receptor, underscoring its potential in drug discovery.
♻ ☆ Tuning-Free Personalized Alignment via Trial-Error-Explain In-Context Learning NAACL 2025
Language models are aligned to the collective voice of many, resulting in generic outputs that do not align with specific users' styles. In this work, we present Trial-Error-Explain In-Context Learning} (ITCL), a tuning-free method that personalizes language models for text generation tasks with fewer than 10 examples per user. TICL iteratively expands an in-context learning prompt via a trial-error-explain process, adding model-generated negative samples and explanations that provide fine-grained guidance towards a specific user's style. TICL achieves favorable win rates on pairwise comparisons with LLM-as-a-judge up to 91.5% against the previous state-of-the-art and outperforms competitive tuning-free baselines for personalized alignment tasks of writing emails, essays and news articles. Both lexical and qualitative analyses show that the negative samples and explanations enable language models to learn stylistic context more effectively and overcome the bias towards structural and formal phrases observed in their zero-shot outputs. By front-loading inference compute to create a user-specific in-context learning prompt that does not require extra generation steps at test time, TICL presents a novel yet simple approach for personalized alignment.
comment: NAACL 2025 Findings
♻ ☆ Cognitive-Mental-LLM: Evaluating Reasoning in Large Language Models for Mental Health Prediction via Online Text
Large Language Models (LLMs) have demonstrated potential in predicting mental health outcomes from online text, yet traditional classification methods often lack interpretability and robustness. This study evaluates structured reasoning techniques-Chain-of-Thought (CoT), Self-Consistency (SC-CoT), and Tree-of-Thought (ToT)-to improve classification accuracy across multiple mental health datasets sourced from Reddit. We analyze reasoning-driven prompting strategies, including Zero-shot CoT and Few-shot CoT, using key performance metrics such as Balanced Accuracy, F1 score, and Sensitivity/Specificity. Our findings indicate that reasoning-enhanced techniques improve classification performance over direct prediction, particularly in complex cases. Compared to baselines such as Zero Shot non-CoT Prompting, and fine-tuned pre-trained transformers such as BERT and Mental-RoBerta, and fine-tuned Open Source LLMs such as Mental Alpaca and Mental-Flan-T5, reasoning-driven LLMs yield notable gains on datasets like Dreaddit (+0.52\% over M-LLM, +0.82\% over BERT) and SDCNL (+4.67\% over M-LLM, +2.17\% over BERT). However, performance declines in Depression Severity, and CSSRS predictions suggest dataset-specific limitations, likely due to our using a more extensive test set. Among prompting strategies, Few-shot CoT consistently outperforms others, reinforcing the effectiveness of reasoning-driven LLMs. Nonetheless, dataset variability highlights challenges in model reliability and interpretability. This study provides a comprehensive benchmark of reasoning-based LLM techniques for mental health text classification. It offers insights into their potential for scalable clinical applications while identifying key challenges for future improvements.
comment: 8 pages, 4 Figures, 3 tables
♻ ☆ Achieving >97% on GSM8K: Deeply Understanding the Problems Makes LLMs Better Solvers for Math Word Problems
Chain-of-Thought (CoT) prompting has enhanced the performance of Large Language Models (LLMs) across various reasoning tasks. However, CoT still falls short in dealing with complex math word problems, as it usually suffers from three pitfalls: semantic misunderstanding errors, calculation errors, and step-missing errors. Prior studies involve addressing the calculation errors and step-missing errors, but neglect the semantic misunderstanding errors, which is the major factor limiting the reasoning performance of LLMs. To this end, we propose a simple-yet-effective method, namely Deeply Understanding the Problems (DUP), to improve the LLMs' math problem-solving ability by addressing semantic misunderstanding errors. The core of our method is to encourage the LLMs to deeply understand the problems and extract the key problem-solving information used for better reasoning. Extensive experiments on 10 diverse reasoning benchmarks show that our DUP method consistently outperforms the other counterparts by a large margin. More encouragingly, DUP achieves a new SOTA result on the GSM8K benchmark, with an accuracy of 97.1% under the zero-shot setting.
comment: The article has been accepted by Frontiers of Computer Science (FCS), with the DOI: { 10.1007/s11704-025-41102-z }
♻ ☆ Time and Memory Trade-off of KV-Cache Compression in Tensor Transformer Decoding
The key-value (KV) cache in the tensor version of transformers presents a significant bottleneck during inference. While previous work analyzes the fundamental space complexity barriers in standard attention mechanisms [Haris and Onak, 2025], our work generalizes the space complexity barriers result to tensor attention version. Our theoretical contributions rely on a reduction from communication complexity and deduce the memory lower bound for tensor-structured attention mechanisms when $d = \Omega(\log n)$. Furthermore, we introduce two types of tensor attention cache and present a trade-off between time and memory for two scenarios. Overall, our work provides a theoretical foundation for us to understand the time-memory tradeoff of KV-Cache compression in tensor attention decoding and offers more perspectives in developing more memory-efficient tensor attention Transformer architectures.
♻ ☆ Adapting Language-Specific LLMs to a Reasoning Model in One Day via Model Merging -- An Open Recipe
This paper investigates data selection and model merging methodologies aimed at incorporating advanced reasoning capabilities such as those of DeepSeek R1 into language-specific large language models (LLMs), with a particular focus on the Thai LLM. Our goal is to enhance the reasoning capabilities of language-specific LLMs while maintaining their target language abilities. DeepSeek R1 excels in reasoning but primarily benefits high-resource languages such as English and Chinese. However, low-resource languages remain underserved due to the dominance of English-centric training data and model optimizations, which limit performance in these languages. This limitation results in unreliable code-switching and diminished effectiveness on tasks in low-resource languages. Meanwhile, local and regional LLM initiatives have attempted to bridge this gap by developing language-specific LLMs that focus on improving local linguistic fidelity. We demonstrate that, with only publicly available datasets and a computational budget of $120, it is possible to enhance the reasoning capabilities of language-specific LLMs to match the level of DeepSeek R1, without compromising their performance on target language tasks.
comment: 9 pages
♻ ☆ Typhoon T1: An Open Thai Reasoning Model
This paper introduces Typhoon T1, an open effort to develop an open Thai reasoning model. A reasoning model is a relatively new type of generative model built on top of large language models (LLMs). A reasoning model generates a long chain of thought before arriving at a final answer, an approach found to improve performance on complex tasks. However, details on developing such a model are limited, especially for reasoning models that can generate traces in a low-resource language. Typhoon T1 presents an open effort that dives into the details of developing a reasoning model in a more cost-effective way by leveraging supervised fine-tuning using open datasets, instead of reinforcement learning. This paper shares the details about synthetic data generation and training, as well as our dataset and model weights. Additionally, we provide insights gained from developing a reasoning model that generalizes across domains and is capable of generating reasoning traces in a low-resource language, using Thai as an example. We hope this open effort provides a foundation for further research in this field.
comment: 25 pages, 6 figures
♻ ☆ Hengqin-RA-v1: Advanced Large Language Model for Diagnosis and Treatment of Rheumatoid Arthritis with Dataset based Traditional Chinese Medicine AAAI-2025
Large language models (LLMs) primarily trained on English texts, often face biases and inaccuracies in Chinese contexts. Their limitations are pronounced in fields like Traditional Chinese Medicine (TCM), where cultural and clinical subtleties are vital, further hindered by a lack of domain-specific data, such as rheumatoid arthritis (RA). To address these issues, this paper introduces Hengqin-RA-v1, the first large language model specifically tailored for TCM with a focus on diagnosing and treating RA. We also present HQ-GCM-RA-C1, a comprehensive RA-specific dataset curated from ancient Chinese medical literature, classical texts, and modern clinical studies. This dataset empowers Hengqin-RA-v1 to deliver accurate and culturally informed responses, effectively bridging the gaps left by general-purpose models. Extensive experiments demonstrate that Hengqin-RA-v1 outperforms state-of-the-art models, even surpassing the diagnostic accuracy of TCM practitioners in certain cases.
comment: 8 pages, 5 figures, AAAI-2025 Workshop
♻ ☆ Group Reasoning Emission Estimation Networks
Accurate greenhouse gas (GHG) emission reporting is critical for governments, businesses, and investors. However, adoption remains limited particularly among small and medium enterprises due to high implementation costs, fragmented emission factor databases, and a lack of robust sector classification methods. To address these challenges, we introduce Group Reasoning Emission Estimation Networks (GREEN), an AI-driven carbon accounting framework that standardizes enterprise-level emission estimation, constructs a large-scale benchmark dataset, and leverages a novel reasoning approach with large language models (LLMs). Specifically, we compile textual descriptions for 20,850 companies with validated North American Industry Classification System (NAICS) labels and align these with an economic model of carbon intensity factors. By reframing sector classification as an information retrieval task, we fine-tune Sentence-BERT models using a contrastive learning loss. To overcome the limitations of single-stage models in handling thousands of hierarchical categories, we propose a Group Reasoning method that ensembles LLM classifiers based on the natural NAICS ontology, decomposing the task into multiple sub-classification steps. We theoretically prove that this approach reduces classification uncertainty and computational complexity. Experiments on 1,114 NAICS categories yield state-of-the-art performance (83.68% Top-1, 91.47% Top-10 accuracy), and case studies on 20 companies report a mean absolute percentage error (MAPE) of 45.88%. The project is available at: https://huggingface.co/datasets/Yvnminc/ExioNAICS.
♻ ☆ Flip Learning: Weakly Supervised Erase to Segment Nodules in Breast Ultrasound
Accurate segmentation of nodules in both 2D breast ultrasound (BUS) and 3D automated breast ultrasound (ABUS) is crucial for clinical diagnosis and treatment planning. Therefore, developing an automated system for nodule segmentation can enhance user independence and expedite clinical analysis. Unlike fully-supervised learning, weakly-supervised segmentation (WSS) can streamline the laborious and intricate annotation process. However, current WSS methods face challenges in achieving precise nodule segmentation, as many of them depend on inaccurate activation maps or inefficient pseudo-mask generation algorithms. In this study, we introduce a novel multi-agent reinforcement learning-based WSS framework called Flip Learning, which relies solely on 2D/3D boxes for accurate segmentation. Specifically, multiple agents are employed to erase the target from the box to facilitate classification tag flipping, with the erased region serving as the predicted segmentation mask. The key contributions of this research are as follows: (1) Adoption of a superpixel/supervoxel-based approach to encode the standardized environment, capturing boundary priors and expediting the learning process. (2) Introduction of three meticulously designed rewards, comprising a classification score reward and two intensity distribution rewards, to steer the agents' erasing process precisely, thereby avoiding both under- and over-segmentation. (3) Implementation of a progressive curriculum learning strategy to enable agents to interact with the environment in a progressively challenging manner, thereby enhancing learning efficiency. Extensively validated on the large in-house BUS and ABUS datasets, our Flip Learning method outperforms state-of-the-art WSS methods and foundation models, and achieves comparable performance as fully-supervised learning algorithms.
comment: Accepted by Medical Image Analysis. 24 pages, 13 figures, 20 tabels
♻ ☆ ReSearch: Learning to Reason with Search for LLMs via Reinforcement Learning
Large Language Models (LLMs) have shown remarkable capabilities in reasoning, exemplified by the success of OpenAI-o1 and DeepSeek-R1. However, integrating reasoning with external search processes remains challenging, especially for complex multi-hop questions requiring multiple retrieval steps. We propose ReSearch, a novel framework that trains LLMs to Reason with Search via reinforcement learning without using any supervised data on reasoning steps. Our approach treats search operations as integral components of the reasoning chain, where when and how to perform searches is guided by text-based thinking, and search results subsequently influence further reasoning. We train ReSearch on Qwen2.5-7B(-Instruct) and Qwen2.5-32B(-Instruct) models and conduct extensive experiments. Despite being trained on only one dataset, our models demonstrate strong generalizability across various benchmarks. Analysis reveals that ReSearch naturally elicits advanced reasoning capabilities such as reflection and self-correction during the reinforcement learning process.
comment: Work in progress
♻ ☆ Sparse Feature Circuits: Discovering and Editing Interpretable Causal Graphs in Language Models
We introduce methods for discovering and applying sparse feature circuits. These are causally implicated subnetworks of human-interpretable features for explaining language model behaviors. Circuits identified in prior work consist of polysemantic and difficult-to-interpret units like attention heads or neurons, rendering them unsuitable for many downstream applications. In contrast, sparse feature circuits enable detailed understanding of unanticipated mechanisms. Because they are based on fine-grained units, sparse feature circuits are useful for downstream tasks: We introduce SHIFT, where we improve the generalization of a classifier by ablating features that a human judges to be task-irrelevant. Finally, we demonstrate an entirely unsupervised and scalable interpretability pipeline by discovering thousands of sparse feature circuits for automatically discovered model behaviors.
comment: Code and data at https://github.com/saprmarks/feature-circuits. Demonstration at https://feature-circuits.xyz
♻ ☆ OODFace: Benchmarking Robustness of Face Recognition under Common Corruptions and Appearance Variations
With the rise of deep learning, facial recognition technology has seen extensive research and rapid development. Although facial recognition is considered a mature technology, we find that existing open-source models and commercial algorithms lack robustness in certain complex Out-of-Distribution (OOD) scenarios, raising concerns about the reliability of these systems. In this paper, we introduce OODFace, which explores the OOD challenges faced by facial recognition models from two perspectives: common corruptions and appearance variations. We systematically design 30 OOD scenarios across 9 major categories tailored for facial recognition. By simulating these challenges on public datasets, we establish three robustness benchmarks: LFW-C/V, CFP-FP-C/V, and YTF-C/V. We then conduct extensive experiments on 19 facial recognition models and 3 commercial APIs, along with extended physical experiments on face masks to assess their robustness. Next, we explore potential solutions from two perspectives: defense strategies and Vision-Language Models (VLMs). Based on the results, we draw several key insights, highlighting the vulnerability of facial recognition systems to OOD data and suggesting possible solutions. Additionally, we offer a unified toolkit that includes all corruption and variation types, easily extendable to other datasets. We hope that our benchmarks and findings can provide guidance for future improvements in facial recognition model robustness.
♻ ☆ DefectFill: Realistic Defect Generation with Inpainting Diffusion Model for Visual Inspection CVPR 2025
Developing effective visual inspection models remains challenging due to the scarcity of defect data. While image generation models have been used to synthesize defect images, producing highly realistic defects remains difficult. We propose DefectFill, a novel method for realistic defect generation that requires only a few reference defect images. It leverages a fine-tuned inpainting diffusion model, optimized with our custom loss functions incorporating defect, object, and attention terms. It enables precise capture of detailed, localized defect features and their seamless integration into defect-free objects. Additionally, our Low-Fidelity Selection method further enhances the defect sample quality. Experiments show that DefectFill generates high-quality defect images, enabling visual inspection models to achieve state-of-the-art performance on the MVTec AD dataset.
comment: Accepted to CVPR 2025
♻ ☆ MoReVQA: Exploring Modular Reasoning Models for Video Question Answering CVPR 2024
This paper addresses the task of video question answering (videoQA) via a decomposed multi-stage, modular reasoning framework. Previous modular methods have shown promise with a single planning stage ungrounded in visual content. However, through a simple and effective baseline, we find that such systems can lead to brittle behavior in practice for challenging videoQA settings. Thus, unlike traditional single-stage planning methods, we propose a multi-stage system consisting of an event parser, a grounding stage, and a final reasoning stage in conjunction with an external memory. All stages are training-free, and performed using few-shot prompting of large models, creating interpretable intermediate outputs at each stage. By decomposing the underlying planning and task complexity, our method, MoReVQA, improves over prior work on standard videoQA benchmarks (NExT-QA, iVQA, EgoSchema, ActivityNet-QA) with state-of-the-art results, and extensions to related tasks (grounded videoQA, paragraph captioning).
comment: CVPR 2024; updated NExT-GQA results in Appendix
♻ ☆ LaMOuR: Leveraging Language Models for Out-of-Distribution Recovery in Reinforcement Learning
Deep Reinforcement Learning (DRL) has demonstrated strong performance in robotic control but remains susceptible to out-of-distribution (OOD) states, often resulting in unreliable actions and task failure. While previous methods have focused on minimizing or preventing OOD occurrences, they largely neglect recovery once an agent encounters such states. Although the latest research has attempted to address this by guiding agents back to in-distribution states, their reliance on uncertainty estimation hinders scalability in complex environments. To overcome this limitation, we introduce Language Models for Out-of-Distribution Recovery (LaMOuR), which enables recovery learning without relying on uncertainty estimation. LaMOuR generates dense reward codes that guide the agent back to a state where it can successfully perform its original task, leveraging the capabilities of LVLMs in image description, logical reasoning, and code generation. Experimental results show that LaMOuR substantially enhances recovery efficiency across diverse locomotion tasks and even generalizes effectively to complex environments, including humanoid locomotion and mobile manipulation, where existing methods struggle. The code and supplementary materials are available at https://lamour-rl.github.io/.
comment: This paper is currently under security review and will be re-released once the review is complete
♻ ☆ iTool: Boosting Tool Use of Large Language Models via Iterative Reinforced Fine-Tuning ACL
Augmenting large language models (LLMs) with external tools is known as a promising approach to enhancing their capabilities, especially for complex tasks. Synthesizing tool-use data through real-world simulations is an effective way to achieve it. Nevertheless, our investigation reveals that (1) training gains significantly decay as synthetic data increases. The model struggles to benefit from more synthetic data due to potential data diversity issues, resulting in poor performance in complex scenarios. Moreover, we find that (2) this challenge primarily manifests as minor discrepancies between the model's output and the ground truth response (termed as deficiency), such as errors in parameter values that require complex reasoning from the context to resolve. To this end, we propose an iterative reinforced fine-tuning strategy designed to alleviate these challenges. This strategy involves: (1) enhancing the diversity of synthetic data through path exploration of Monte Carlo Tree Search. (2) iteratively identifying deficiency-related data, constructing fine-grained preference pairs to pinpoint deficiencies, and then applying preference optimization to optimize these deficiencies. Our experiments show that models trained using our method achieve about 12\% better performance than baseline models, outperforming larger open-source and closed-source models.
comment: under review ACL
♻ ☆ Towards Controllable Speech Synthesis in the Era of Large Language Models: A Survey
Text-to-speech (TTS), also known as speech synthesis, is a prominent research area that aims to generate natural-sounding human speech from text. Recently, with the increasing industrial demand, TTS technologies have evolved beyond synthesizing human-like speech to enabling controllable speech generation. This includes fine-grained control over various attributes of synthesized speech such as emotion, prosody, timbre, and duration. In addition, advancements in deep learning, such as diffusion and large language models, have significantly enhanced controllable TTS over the past several years. In this work, we conduct a comprehensive survey of controllable TTS, covering approaches ranging from basic control techniques to methods utilizing natural language prompts, aiming to provide a clear understanding of the current state of research. We examine the general controllable TTS pipeline, challenges, model architectures, and control strategies, offering a comprehensive and clear taxonomy of existing methods. Additionally, we provide a detailed summary of datasets and evaluation metrics and shed some light on the applications and future directions of controllable TTS. To the best of our knowledge, this survey paper provides the first comprehensive review of emerging controllable TTS methods, which can serve as a beneficial resource for both academic researchers and industrial practitioners.
comment: A comprehensive survey on controllable TTS, 26 pages, 7 tables, 6 figures, 317 references. Under review
♻ ☆ Reason-RFT: Reinforcement Fine-Tuning for Visual Reasoning
Visual reasoning abilities play a crucial role in understanding complex multimodal data, advancing both domain-specific applications and artificial general intelligence (AGI). Existing methods improve VLM reasoning via Chain-of-Thought (CoT) supervised fine-tuning, using meticulously annotated training data to enhance visual reasoning capabilities. However, this training paradigm may lead to overfitting and cognitive rigidity, restricting the model's ability to transfer visual reasoning skills across domains and limiting its real-world applicability. To address these limitations, we propose Reason-RFT, a novel reinforcement fine-tuning framework that significantly enhances generalization capabilities in visual reasoning tasks. Reason-RFT introduces a two-phase training framework for visual reasoning: (1) Supervised Fine-Tuning (SFT) with curated Chain-of-Thought (CoT) data activates the reasoning potential of Vision-Language Models (VLMs), followed by (2) Group Relative Policy Optimization (GRPO)-based reinforcement learning that generates multiple reasoning-response pairs, significantly enhancing generalization in visual reasoning tasks. To evaluate Reason-RFT's visual reasoning capabilities, we reconstructed a comprehensive dataset spanning visual counting, structure perception, and spatial transformation. Experimental results demonstrate Reasoning-RFT's three key advantages: (1) Performance Enhancement: achieving state-of-the-art results across multiple tasks, outperforming most mainstream open-source and proprietary models; (2) Generalization Superiority: consistently maintaining robust performance across diverse tasks and domains, outperforming alternative training paradigms; (3) Data Efficiency: excelling in few-shot learning scenarios while surpassing full-dataset SFT baselines. Project website: https://tanhuajie.github.io/ReasonRFT
comment: 35 pages, 22 figures
GR00T N1: An Open Foundation Model for Generalist Humanoid Robots
General-purpose robots need a versatile body and an intelligent mind. Recent advancements in humanoid robots have shown great promise as a hardware platform for building generalist autonomy in the human world. A robot foundation model, trained on massive and diverse data sources, is essential for enabling the robots to reason about novel situations, robustly handle real-world variability, and rapidly learn new tasks. To this end, we introduce GR00T N1, an open foundation model for humanoid robots. GR00T N1 is a Vision-Language-Action (VLA) model with a dual-system architecture. The vision-language module (System 2) interprets the environment through vision and language instructions. The subsequent diffusion transformer module (System 1) generates fluid motor actions in real time. Both modules are tightly coupled and jointly trained end-to-end. We train GR00T N1 with a heterogeneous mixture of real-robot trajectories, human videos, and synthetically generated datasets. We show that our generalist robot model GR00T N1 outperforms the state-of-the-art imitation learning baselines on standard simulation benchmarks across multiple robot embodiments. Furthermore, we deploy our model on the Fourier GR-1 humanoid robot for language-conditioned bimanual manipulation tasks, achieving strong performance with high data efficiency.
comment: Authors are listed alphabetically. Project leads are Linxi "Jim" Fan and Yuke Zhu. For more information, see https://developer.nvidia.com/isaac/gr00t
♻ ☆ Semi-Supervised Self-Learning Enhanced Music Emotion Recognition
Music emotion recognition (MER) aims to identify the emotions conveyed in a given musical piece. However, currently, in the field of MER, the available public datasets have limited sample sizes. Recently, segment-based methods for emotion-related tasks have been proposed, which train backbone networks on shorter segments instead of entire audio clips, thereby naturally augmenting training samples without requiring additional resources. Then, the predicted segment-level results are aggregated to obtain the entire song prediction. The most commonly used method is that the segment inherits the label of the clip containing it, but music emotion is not constant during the whole clip. Doing so will introduce label noise and make the training easy to overfit. To handle the noisy label issue, we propose a semi-supervised self-learning (SSSL) method, which can differentiate between samples with correct and incorrect labels in a self-learning manner, thus effectively utilizing the augmented segment-level data. Experiments on three public emotional datasets demonstrate that the proposed method can achieve better or comparable performance.
comment: 12 pages, 2 figures
♻ ☆ AnyBimanual: Transferring Unimanual Policy for General Bimanual Manipulation
Performing general language-conditioned bimanual manipulation tasks is of great importance for many applications ranging from household service to industrial assembly. However, collecting bimanual manipulation data is expensive due to the high-dimensional action space, which poses challenges for conventional methods to handle general bimanual manipulation tasks. In contrast, unimanual policy has recently demonstrated impressive generalizability across a wide range of tasks because of scaled model parameters and training data, which can provide sharable manipulation knowledge for bimanual systems. To this end, we propose a plug-and-play method named AnyBimanual, which transfers pre-trained unimanual policy to general bimanual manipulation policy with few bimanual demonstrations. Specifically, we first introduce a skill manager to dynamically schedule the skill representations discovered from pre-trained unimanual policy for bimanual manipulation tasks, which linearly combines skill primitives with task-oriented compensation to represent the bimanual manipulation instruction. To mitigate the observation discrepancy between unimanual and bimanual systems, we present a visual aligner to generate soft masks for visual embedding of the workspace, which aims to align visual input of unimanual policy model for each arm with those during pretraining stage. AnyBimanual shows superiority on 12 simulated tasks from RLBench2 with a sizable 12.67% improvement in success rate over previous methods. Experiments on 9 real-world tasks further verify its practicality with an average success rate of 84.62%.
comment: Project page: https://anybimanual.github.io/
♻ ☆ A Holistic Evaluation of Piano Sound Quality
This paper aims to develop a holistic evaluation method for piano sound quality to assist in purchasing decisions. Unlike previous studies that focused on the effect of piano performance techniques on sound quality, this study evaluates the inherent sound quality of different pianos. To derive quality evaluation systems, the study uses subjective questionnaires based on a piano sound quality dataset. The method selects the optimal piano classification models by comparing the fine-tuning results of different pre-training models of Convolutional Neural Networks (CNN). To improve the interpretability of the models, the study applies Equivalent Rectangular Bandwidth (ERB) analysis. The results reveal that musically trained individuals are better able to distinguish between the sound quality differences of different pianos. The best fine-tuned CNN pre-trained backbone achieves a high accuracy of 98.3% as the piano classifier. However, the dataset is limited, and the audio is sliced to increase its quantity, resulting in a lack of diversity and balance, so we use focal loss to reduce the impact of data imbalance. To optimize the method, the dataset will be expanded, or few-shot learning techniques will be employed in future research.
comment: 15 pages, 9 figures
♻ ☆ VERA: Explainable Video Anomaly Detection via Verbalized Learning of Vision-Language Models CVPR 2025
The rapid advancement of vision-language models (VLMs) has established a new paradigm in video anomaly detection (VAD): leveraging VLMs to simultaneously detect anomalies and provide comprehendible explanations for the decisions. Existing work in this direction often assumes the complex reasoning required for VAD exceeds the capabilities of pretrained VLMs. Consequently, these approaches either incorporate specialized reasoning modules during inference or rely on instruction tuning datasets through additional training to adapt VLMs for VAD. However, such strategies often incur substantial computational costs or data annotation overhead. To address these challenges in explainable VAD, we introduce a verbalized learning framework named VERA that enables VLMs to perform VAD without model parameter modifications. Specifically, VERA automatically decomposes the complex reasoning required for VAD into reflections on simpler, more focused guiding questions capturing distinct abnormal patterns. It treats these reflective questions as learnable parameters and optimizes them through data-driven verbal interactions between learner and optimizer VLMs, using coarsely labeled training data. During inference, VERA embeds the learned questions into model prompts to guide VLMs in generating segment-level anomaly scores, which are then refined into frame-level scores via the fusion of scene and temporal contexts. Experimental results on challenging benchmarks demonstrate that the learned questions of VERA are highly adaptable, significantly improving both detection performance and explainability of VLMs for VAD.
comment: Accepted in CVPR 2025
♻ ☆ SyncDiff: Synchronized Motion Diffusion for Multi-Body Human-Object Interaction Synthesis
Synthesizing realistic human-object interaction motions is a critical problem in VR/AR and human animation. Unlike the commonly studied scenarios involving a single human or hand interacting with one object, we address a more generic multi-body setting with arbitrary numbers of humans, hands, and objects. This complexity introduces significant challenges in synchronizing motions due to the high correlations and mutual influences among bodies. To address these challenges, we introduce SyncDiff, a novel method for multi-body interaction synthesis using a synchronized motion diffusion strategy. SyncDiff employs a single diffusion model to capture the joint distribution of multi-body motions. To enhance motion fidelity, we propose a frequency-domain motion decomposition scheme. Additionally, we introduce a new set of alignment scores to emphasize the synchronization of different body motions. SyncDiff jointly optimizes both data sample likelihood and alignment likelihood through an explicit synchronization strategy. Extensive experiments across four datasets with various multi-body configurations demonstrate the superiority of SyncDiff over existing state-of-the-art motion synthesis methods.
comment: 26 pages, 10 figures
♻ ☆ Evaluation-Driven Development of LLM Agents: A Process Model and Reference Architecture
Large Language Models (LLMs) have enabled the emergence of LLM agents: autonomous systems capable of achieving under-specified goals and adapting post-deployment, often without explicit code or model changes. Evaluating these agents is critical to ensuring their performance and safety, especially given their dynamic, probabilistic, and evolving nature. However, traditional approaches such as predefined test cases and standard redevelopment pipelines struggle to address the unique challenges of LLM agent evaluation. These challenges include capturing open-ended behaviors, handling emergent outcomes, and enabling continuous adaptation over the agent's lifecycle. To address these issues, we propose an evaluation-driven development approach, inspired by test-driven and behavior-driven development but reimagined for the unique characteristics of LLM agents. Through a multivocal literature review (MLR), we synthesize the limitations of existing LLM evaluation methods and introduce a novel process model and reference architecture tailored for evaluation-driven development of LLM agents. Our approach integrates online (runtime) and offline (redevelopment) evaluations, enabling adaptive runtime adjustments and systematic iterative refinement of pipelines, artifacts, system architecture, and LLMs themselves. By continuously incorporating evaluation results, including fine-grained feedback from human and AI evaluators, into each stage of development and operation, this framework ensures that LLM agents remain aligned with evolving goals, user needs, and governance standards.
♻ ☆ LSEAttention is All You Need for Time Series Forecasting
Transformer-based architectures have achieved remarkable success in natural language processing and computer vision. However, their performance in multivariate long-term forecasting often falls short compared to simpler linear baselines. Previous research has identified the traditional attention mechanism as a key factor limiting their effectiveness in this domain. To bridge this gap, we introduce LATST, a novel approach designed to mitigate entropy collapse and training instability common challenges in Transformer-based time series forecasting. We rigorously evaluate LATST across multiple real-world multivariate time series datasets, demonstrating its ability to outperform existing state-of-the-art Transformer models. Notably, LATST manages to achieve competitive performance with fewer parameters than some linear models on certain datasets, highlighting its efficiency and effectiveness.
comment: 8 pages with referencing, 1 figure, 5 tables
♻ ☆ SoK: How Robust is Audio Watermarking in Generative AI models?
Audio watermarking is increasingly used to verify the provenance of AI-generated content, enabling applications such as detecting AI-generated speech, protecting music IP, and defending against voice cloning. To be effective, audio watermarks must resist removal attacks that distort signals to evade detection. While many schemes claim robustness, these claims are typically tested in isolation and against a limited set of attacks. A systematic evaluation against diverse removal attacks is lacking, hindering practical deployment. In this paper, we investigate whether recent watermarking schemes that claim robustness can withstand a broad range of removal attacks. First, we introduce a taxonomy covering 22 audio watermarking schemes. Next, we summarize their underlying technologies and potential vulnerabilities. We then present a large-scale empirical study to assess their robustness. To support this, we build an evaluation framework encompassing 22 types of removal attacks (109 configurations) including signal-level, physical-level, and AI-induced distortions. We reproduce 9 watermarking schemes using open-source code, identify 8 new highly effective attacks, and highlight 11 key findings that expose the fundamental limitations of these methods across 3 public datasets. Our results reveal that none of the surveyed schemes can withstand all tested distortions. This evaluation offers a comprehensive view of how current watermarking methods perform under real-world threats. Our demo and code are available at https://sokaudiowm.github.io/.
♻ ☆ GCA-SUNet: A Gated Context-Aware Swin-UNet for Exemplar-Free Counting ICME 2025
Exemplar-Free Counting aims to count objects of interest without intensive annotations of objects or exemplars. To achieve this, we propose a Gated Context-Aware Swin-UNet (GCA-SUNet) to directly map an input image to the density map of countable objects. Specifically, a set of Swin transformers form an encoder to derive a robust feature representation, and a Gated Context-Aware Modulation block is designed to suppress irrelevant objects or background through a gate mechanism and exploit the attentive support of objects of interest through a self-similarity matrix. The gate strategy is also incorporated into the bottleneck network and the decoder of the Swin-UNet to highlight the features most relevant to objects of interest. By explicitly exploiting the attentive support among countable objects and eliminating irrelevant features through the gate mechanisms, the proposed GCA-SUNet focuses on and counts objects of interest without relying on predefined categories or exemplars. Experimental results on the real-world datasets such as FSC-147 and CARPK demonstrate that GCA-SUNet significantly and consistently outperforms state-of-the-art methods. The code is available at https://github.com/Amordia/GCA-SUNet.
comment: Accepted by ICME 2025
Computer Vision and Pattern Recognition 208
☆ Semantic Library Adaptation: LoRA Retrieval and Fusion for Open-Vocabulary Semantic Segmentation CVPR 2025
Open-vocabulary semantic segmentation models associate vision and text to label pixels from an undefined set of classes using textual queries, providing versatile performance on novel datasets. However, large shifts between training and test domains degrade their performance, requiring fine-tuning for effective real-world applications. We introduce Semantic Library Adaptation (SemLA), a novel framework for training-free, test-time domain adaptation. SemLA leverages a library of LoRA-based adapters indexed with CLIP embeddings, dynamically merging the most relevant adapters based on proximity to the target domain in the embedding space. This approach constructs an ad-hoc model tailored to each specific input without additional training. Our method scales efficiently, enhances explainability by tracking adapter contributions, and inherently protects data privacy, making it ideal for sensitive applications. Comprehensive experiments on a 20-domain benchmark built over 10 standard datasets demonstrate SemLA's superior adaptability and performance across diverse settings, establishing a new standard in domain adaptation for open-vocabulary semantic segmentation.
comment: CVPR 2025. Project page: https://thegoodailab.org/semla Code: https://github.com/rezaqorbani/SemLA
☆ VideoMage: Multi-Subject and Motion Customization of Text-to-Video Diffusion Models CVPR 2025
Customized text-to-video generation aims to produce high-quality videos that incorporate user-specified subject identities or motion patterns. However, existing methods mainly focus on personalizing a single concept, either subject identity or motion pattern, limiting their effectiveness for multiple subjects with the desired motion patterns. To tackle this challenge, we propose a unified framework VideoMage for video customization over both multiple subjects and their interactive motions. VideoMage employs subject and motion LoRAs to capture personalized content from user-provided images and videos, along with an appearance-agnostic motion learning approach to disentangle motion patterns from visual appearance. Furthermore, we develop a spatial-temporal composition scheme to guide interactions among subjects within the desired motion patterns. Extensive experiments demonstrate that VideoMage outperforms existing methods, generating coherent, user-controlled videos with consistent subject identities and interactions.
comment: CVPR 2025. Project Page: https://jasper0314-huang.github.io/videomage-customization
☆ Mobile-VideoGPT: Fast and Accurate Video Understanding Language Model
Video understanding models often struggle with high computational requirements, extensive parameter counts, and slow inference speed, making them inefficient for practical use. To tackle these challenges, we propose Mobile-VideoGPT, an efficient multimodal framework designed to operate with fewer than a billion parameters. Unlike traditional video large multimodal models (LMMs), Mobile-VideoGPT consists of lightweight dual visual encoders, efficient projectors, and a small language model (SLM), enabling real-time throughput. To further improve efficiency, we present an Attention-Based Frame Scoring mechanism to select the key-frames, along with an efficient token projector that prunes redundant visual tokens and preserves essential contextual cues. We evaluate our model across well-established six video understanding benchmarks (e.g., MVBench, EgoSchema, NextQA, and PercepTest). Our results show that Mobile-VideoGPT-0.5B can generate up to 46 tokens per second while outperforming existing state-of-the-art 0.5B-parameter models by 6 points on average with 40% fewer parameters and more than 2x higher throughput. Our code and models are publicly available at: https://github.com/Amshaker/Mobile-VideoGPT.
comment: Technical Report. Project Page: https://amshaker.github.io/Mobile-VideoGPT
☆ X$^{2}$-Gaussian: 4D Radiative Gaussian Splatting for Continuous-time Tomographic Reconstruction
Four-dimensional computed tomography (4D CT) reconstruction is crucial for capturing dynamic anatomical changes but faces inherent limitations from conventional phase-binning workflows. Current methods discretize temporal resolution into fixed phases with respiratory gating devices, introducing motion misalignment and restricting clinical practicality. In this paper, We propose X$^2$-Gaussian, a novel framework that enables continuous-time 4D-CT reconstruction by integrating dynamic radiative Gaussian splatting with self-supervised respiratory motion learning. Our approach models anatomical dynamics through a spatiotemporal encoder-decoder architecture that predicts time-varying Gaussian deformations, eliminating phase discretization. To remove dependency on external gating devices, we introduce a physiology-driven periodic consistency loss that learns patient-specific breathing cycles directly from projections via differentiable optimization. Extensive experiments demonstrate state-of-the-art performance, achieving a 9.93 dB PSNR gain over traditional methods and 2.25 dB improvement against prior Gaussian splatting techniques. By unifying continuous motion modeling with hardware-free period learning, X$^2$-Gaussian advances high-fidelity 4D CT reconstruction for dynamic clinical imaging. Project website at: https://x2-gaussian.github.io/.
comment: Project Page: https://x2-gaussian.github.io/
☆ HS-SLAM: Hybrid Representation with Structural Supervision for Improved Dense SLAM ICRA 2025
NeRF-based SLAM has recently achieved promising results in tracking and reconstruction. However, existing methods face challenges in providing sufficient scene representation, capturing structural information, and maintaining global consistency in scenes emerging significant movement or being forgotten. To this end, we present HS-SLAM to tackle these problems. To enhance scene representation capacity, we propose a hybrid encoding network that combines the complementary strengths of hash-grid, tri-planes, and one-blob, improving the completeness and smoothness of reconstruction. Additionally, we introduce structural supervision by sampling patches of non-local pixels rather than individual rays to better capture the scene structure. To ensure global consistency, we implement an active global bundle adjustment (BA) to eliminate camera drifts and mitigate accumulative errors. Experimental results demonstrate that HS-SLAM outperforms the baselines in tracking and reconstruction accuracy while maintaining the efficiency required for robotics.
comment: ICRA 2025. Project Page: https://zorangong.github.io/HS-SLAM/
☆ Test-Time Visual In-Context Tuning CVPR 2025
Visual in-context learning (VICL), as a new paradigm in computer vision, allows the model to rapidly adapt to various tasks with only a handful of prompts and examples. While effective, the existing VICL paradigm exhibits poor generalizability under distribution shifts. In this work, we propose test-time Visual In-Context Tuning (VICT), a method that can adapt VICL models on the fly with a single test sample. Specifically, we flip the role between the task prompts and the test sample and use a cycle consistency loss to reconstruct the original task prompt output. Our key insight is that a model should be aware of a new test distribution if it can successfully recover the original task prompts. Extensive experiments on six representative vision tasks ranging from high-level visual understanding to low-level image processing, with 15 common corruptions, demonstrate that our VICT can improve the generalizability of VICL to unseen new domains. In addition, we show the potential of applying VICT for unseen tasks at test time. Code: https://github.com/Jiahao000/VICT.
comment: CVPR 2025. Code: https://github.com/Jiahao000/VICT
☆ Video-R1: Reinforcing Video Reasoning in MLLMs
Inspired by DeepSeek-R1's success in eliciting reasoning abilities through rule-based reinforcement learning (RL), we introduce Video-R1 as the first attempt to systematically explore the R1 paradigm for eliciting video reasoning within multimodal large language models (MLLMs). However, directly applying RL training with the GRPO algorithm to video reasoning presents two primary challenges: (i) a lack of temporal modeling for video reasoning, and (ii) the scarcity of high-quality video-reasoning data. To address these issues, we first propose the T-GRPO algorithm, which encourages models to utilize temporal information in videos for reasoning. Additionally, instead of relying solely on video data, we incorporate high-quality image-reasoning data into the training process. We have constructed two datasets: Video-R1-COT-165k for SFT cold start and Video-R1-260k for RL training, both comprising image and video data. Experimental results demonstrate that Video-R1 achieves significant improvements on video reasoning benchmarks such as VideoMMMU and VSI-Bench, as well as on general video benchmarks including MVBench and TempCompass, etc. Notably, Video-R1-7B attains a 35.8% accuracy on video spatial reasoning benchmark VSI-bench, surpassing the commercial proprietary model GPT-4o. All codes, models, data are released.
comment: Project page: https://github.com/tulerfeng/Video-R1
☆ Optimal Stepsize for Diffusion Sampling
Diffusion models achieve remarkable generation quality but suffer from computational intensive sampling due to suboptimal step discretization. While existing works focus on optimizing denoising directions, we address the principled design of stepsize schedules. This paper proposes Optimal Stepsize Distillation, a dynamic programming framework that extracts theoretically optimal schedules by distilling knowledge from reference trajectories. By reformulating stepsize optimization as recursive error minimization, our method guarantees global discretization bounds through optimal substructure exploitation. Crucially, the distilled schedules demonstrate strong robustness across architectures, ODE solvers, and noise schedules. Experiments show 10x accelerated text-to-image generation while preserving 99.4% performance on GenEval. Our code is available at https://github.com/bebebe666/OptimalSteps.
☆ StyleMotif: Multi-Modal Motion Stylization using Style-Content Cross Fusion
We present StyleMotif, a novel Stylized Motion Latent Diffusion model, generating motion conditioned on both content and style from multiple modalities. Unlike existing approaches that either focus on generating diverse motion content or transferring style from sequences, StyleMotif seamlessly synthesizes motion across a wide range of content while incorporating stylistic cues from multi-modal inputs, including motion, text, image, video, and audio. To achieve this, we introduce a style-content cross fusion mechanism and align a style encoder with a pre-trained multi-modal model, ensuring that the generated motion accurately captures the reference style while preserving realism. Extensive experiments demonstrate that our framework surpasses existing methods in stylized motion generation and exhibits emergent capabilities for multi-modal motion stylization, enabling more nuanced motion synthesis. Source code and pre-trained models will be released upon acceptance. Project Page: https://stylemotif.github.io
comment: Project Page: https://stylemotif.github.io
☆ LOCORE: Image Re-ranking with Long-Context Sequence Modeling CVPR 2025
We introduce LOCORE, Long-Context Re-ranker, a model that takes as input local descriptors corresponding to an image query and a list of gallery images and outputs similarity scores between the query and each gallery image. This model is used for image retrieval, where typically a first ranking is performed with an efficient similarity measure, and then a shortlist of top-ranked images is re-ranked based on a more fine-grained similarity measure. Compared to existing methods that perform pair-wise similarity estimation with local descriptors or list-wise re-ranking with global descriptors, LOCORE is the first method to perform list-wise re-ranking with local descriptors. To achieve this, we leverage efficient long-context sequence models to effectively capture the dependencies between query and gallery images at the local-descriptor level. During testing, we process long shortlists with a sliding window strategy that is tailored to overcome the context size limitations of sequence models. Our approach achieves superior performance compared with other re-rankers on established image retrieval benchmarks of landmarks (ROxf and RPar), products (SOP), fashion items (In-Shop), and bird species (CUB-200) while having comparable latency to the pair-wise local descriptor re-rankers.
comment: CVPR 2025
☆ A Unified Image-Dense Annotation Generation Model for Underwater Scenes CVPR 2025
Underwater dense prediction, especially depth estimation and semantic segmentation, is crucial for gaining a comprehensive understanding of underwater scenes. Nevertheless, high-quality and large-scale underwater datasets with dense annotations remain scarce because of the complex environment and the exorbitant data collection costs. This paper proposes a unified Text-to-Image and DEnse annotation generation method (TIDE) for underwater scenes. It relies solely on text as input to simultaneously generate realistic underwater images and multiple highly consistent dense annotations. Specifically, we unify the generation of text-to-image and text-to-dense annotations within a single model. The Implicit Layout Sharing mechanism (ILS) and cross-modal interaction method called Time Adaptive Normalization (TAN) are introduced to jointly optimize the consistency between image and dense annotations. We synthesize a large-scale underwater dataset using TIDE to validate the effectiveness of our method in underwater dense prediction tasks. The results demonstrate that our method effectively improves the performance of existing underwater dense prediction models and mitigates the scarcity of underwater data with dense annotations. We hope our method can offer new perspectives on alleviating data scarcity issues in other fields. The code is available at https: //github.com/HongkLin/TIDE.
comment: Accepted by CVPR 2025. The code is available at https: //github.com/HongkLin/TIDE
☆ Visual Jenga: Discovering Object Dependencies via Counterfactual Inpainting
This paper proposes a novel scene understanding task called Visual Jenga. Drawing inspiration from the game Jenga, the proposed task involves progressively removing objects from a single image until only the background remains. Just as Jenga players must understand structural dependencies to maintain tower stability, our task reveals the intrinsic relationships between scene elements by systematically exploring which objects can be removed while preserving scene coherence in both physical and geometric sense. As a starting point for tackling the Visual Jenga task, we propose a simple, data-driven, training-free approach that is surprisingly effective on a range of real-world images. The principle behind our approach is to utilize the asymmetry in the pairwise relationships between objects within a scene and employ a large inpainting model to generate a set of counterfactuals to quantify the asymmetry.
comment: project page: https://visualjenga.github.io/
☆ Semantic Consistent Language Gaussian Splatting for Point-Level Open-vocabulary Querying
Open-vocabulary querying in 3D Gaussian Splatting aims to identify semantically relevant regions within a 3D Gaussian representation based on a given text query. Prior work, such as LangSplat, addressed this task by retrieving these regions in the form of segmentation masks on 2D renderings. More recently, OpenGaussian introduced point-level querying, which directly selects a subset of 3D Gaussians. In this work, we propose a point-level querying method that builds upon LangSplat's framework. Our approach improves the framework in two key ways: (a) we leverage masklets from the Segment Anything Model 2 (SAM2) to establish semantic consistent ground-truth for distilling the language Gaussians; (b) we introduces a novel two-step querying approach that first retrieves the distilled ground-truth and subsequently uses the ground-truth to query the individual Gaussians. Experimental evaluations on three benchmark datasets demonstrate that the proposed method achieves better performance compared to state-of-the-art approaches. For instance, our method achieves an mIoU improvement of +20.42 on the 3D-OVS dataset.
☆ Stable-SCore: A Stable Registration-based Framework for 3D Shape Correspondence CVPR 2025
Establishing character shape correspondence is a critical and fundamental task in computer vision and graphics, with diverse applications including re-topology, attribute transfer, and shape interpolation. Current dominant functional map methods, while effective in controlled scenarios, struggle in real situations with more complex challenges such as non-isometric shape discrepancies. In response, we revisit registration-for-correspondence methods and tap their potential for more stable shape correspondence estimation. To overcome their common issues including unstable deformations and the necessity for careful pre-alignment or high-quality initial 3D correspondences, we introduce Stable-SCore: A Stable Registration-based Framework for 3D Shape Correspondence. We first re-purpose a foundation model for 2D character correspondence that ensures reliable and stable 2D mappings. Crucially, we propose a novel Semantic Flow Guided Registration approach that leverages 2D correspondence to guide mesh deformations. Our framework significantly surpasses existing methods in challenging scenarios, and brings possibilities for a wide array of real applications, as demonstrated in our results.
comment: Accepted by CVPR 2025. Homepage: https://haolinliu97.github.io/Stable-Score/
☆ Exploring the Evolution of Physics Cognition in Video Generation: A Survey
Recent advancements in video generation have witnessed significant progress, especially with the rapid advancement of diffusion models. Despite this, their deficiencies in physical cognition have gradually received widespread attention - generated content often violates the fundamental laws of physics, falling into the dilemma of ''visual realism but physical absurdity". Researchers began to increasingly recognize the importance of physical fidelity in video generation and attempted to integrate heuristic physical cognition such as motion representations and physical knowledge into generative systems to simulate real-world dynamic scenarios. Considering the lack of a systematic overview in this field, this survey aims to provide a comprehensive summary of architecture designs and their applications to fill this gap. Specifically, we discuss and organize the evolutionary process of physical cognition in video generation from a cognitive science perspective, while proposing a three-tier taxonomy: 1) basic schema perception for generation, 2) passive cognition of physical knowledge for generation, and 3) active cognition for world simulation, encompassing state-of-the-art methods, classical paradigms, and benchmarks. Subsequently, we emphasize the inherent key challenges in this domain and delineate potential pathways for future research, contributing to advancing the frontiers of discussion in both academia and industry. Through structured review and interdisciplinary analysis, this survey aims to provide directional guidance for developing interpretable, controllable, and physically consistent video generation paradigms, thereby propelling generative models from the stage of ''visual mimicry'' towards a new phase of ''human-like physical comprehension''.
comment: A comprehensive list of papers studied in this survey is available at https://github.com/minnie-lin/Awesome-Physics-Cognition-based-Video-Generation
☆ Uni4D: Unifying Visual Foundation Models for 4D Modeling from a Single Video CVPR 2025
This paper presents a unified approach to understanding dynamic scenes from casual videos. Large pretrained vision foundation models, such as vision-language, video depth prediction, motion tracking, and segmentation models, offer promising capabilities. However, training a single model for comprehensive 4D understanding remains challenging. We introduce Uni4D, a multi-stage optimization framework that harnesses multiple pretrained models to advance dynamic 3D modeling, including static/dynamic reconstruction, camera pose estimation, and dense 3D motion tracking. Our results show state-of-the-art performance in dynamic 4D modeling with superior visual quality. Notably, Uni4D requires no retraining or fine-tuning, highlighting the effectiveness of repurposing visual foundation models for 4D understanding.
comment: CVPR 2025. Project page (with code): https://davidyao99.github.io/uni4d
☆ Fwd2Bot: LVLM Visual Token Compression with Double Forward Bottleneck
In this work, we aim to compress the vision tokens of a Large Vision Language Model (LVLM) into a representation that is simultaneously suitable for (a) generative and (b) discriminative tasks, (c) is nearly lossless, and (d) is storage-efficient. We propose a novel compression approach, called Fwd2Bot, that uses the LVLM itself to compress the visual information in a task-agnostic manner. At the core of Fwd2bot there exists a "double-forward pass" training strategy, whereby, during the first forward pass, the LLM (of the LVLM) creates a bottleneck by condensing the visual information into a small number of summary tokens. Then, using the same LLM, the second forward pass processes the language instruction(s) alongside the summary tokens, used as a direct replacement for the image ones. The training signal is provided by two losses: an autoregressive one applied after the second pass that provides a direct optimization objective for compression, and a contrastive loss, applied after the first pass, that further boosts the representation strength, especially for discriminative tasks. The training is further enhanced by stage-specific adapters. We accompany the proposed method by an in-depth ablation study. Overall, Fwd2Bot results in highly-informative compressed representations suitable for both generative and discriminative tasks. For generative tasks, we offer a 2x higher compression rate without compromising the generative capabilities, setting a new state-of-the-art result. For discriminative tasks, we set a new state-of-the-art on image retrieval and compositionality.
☆ Lumina-Image 2.0: A Unified and Efficient Image Generative Framework
We introduce Lumina-Image 2.0, an advanced text-to-image generation framework that achieves significant progress compared to previous work, Lumina-Next. Lumina-Image 2.0 is built upon two key principles: (1) Unification - it adopts a unified architecture (Unified Next-DiT) that treats text and image tokens as a joint sequence, enabling natural cross-modal interactions and allowing seamless task expansion. Besides, since high-quality captioners can provide semantically well-aligned text-image training pairs, we introduce a unified captioning system, Unified Captioner (UniCap), specifically designed for T2I generation tasks. UniCap excels at generating comprehensive and accurate captions, accelerating convergence and enhancing prompt adherence. (2) Efficiency - to improve the efficiency of our proposed model, we develop multi-stage progressive training strategies and introduce inference acceleration techniques without compromising image quality. Extensive evaluations on academic benchmarks and public text-to-image arenas show that Lumina-Image 2.0 delivers strong performances even with only 2.6B parameters, highlighting its scalability and design efficiency. We have released our training details, code, and models at https://github.com/Alpha-VLLM/Lumina-Image-2.0.
comment: Tech Report, 21 pages, 12 figures
☆ VBench-2.0: Advancing Video Generation Benchmark Suite for Intrinsic Faithfulness
Video generation has advanced significantly, evolving from producing unrealistic outputs to generating videos that appear visually convincing and temporally coherent. To evaluate these video generative models, benchmarks such as VBench have been developed to assess their faithfulness, measuring factors like per-frame aesthetics, temporal consistency, and basic prompt adherence. However, these aspects mainly represent superficial faithfulness, which focus on whether the video appears visually convincing rather than whether it adheres to real-world principles. While recent models perform increasingly well on these metrics, they still struggle to generate videos that are not just visually plausible but fundamentally realistic. To achieve real "world models" through video generation, the next frontier lies in intrinsic faithfulness to ensure that generated videos adhere to physical laws, commonsense reasoning, anatomical correctness, and compositional integrity. Achieving this level of realism is essential for applications such as AI-assisted filmmaking and simulated world modeling. To bridge this gap, we introduce VBench-2.0, a next-generation benchmark designed to automatically evaluate video generative models for their intrinsic faithfulness. VBench-2.0 assesses five key dimensions: Human Fidelity, Controllability, Creativity, Physics, and Commonsense, each further broken down into fine-grained capabilities. Tailored for individual dimensions, our evaluation framework integrates generalists such as state-of-the-art VLMs and LLMs, and specialists, including anomaly detection methods proposed for video generation. We conduct extensive annotations to ensure alignment with human judgment. By pushing beyond superficial faithfulness toward intrinsic faithfulness, VBench-2.0 aims to set a new standard for the next generation of video generative models in pursuit of intrinsic faithfulness.
comment: Equal contributions from first two authors. Project page: https://vchitect.github.io/VBench-2.0-project/ Code: https://github.com/Vchitect/VBench
☆ Reconstructing Humans with a Biomechanically Accurate Skeleton CVPR 2025
In this paper, we introduce a method for reconstructing 3D humans from a single image using a biomechanically accurate skeleton model. To achieve this, we train a transformer that takes an image as input and estimates the parameters of the model. Due to the lack of training data for this task, we build a pipeline to produce pseudo ground truth model parameters for single images and implement a training procedure that iteratively refines these pseudo labels. Compared to state-of-the-art methods for 3D human mesh recovery, our model achieves competitive performance on standard benchmarks, while it significantly outperforms them in settings with extreme 3D poses and viewpoints. Additionally, we show that previous reconstruction methods frequently violate joint angle limits, leading to unnatural rotations. In contrast, our approach leverages the biomechanically plausible degrees of freedom making more realistic joint rotation estimates. We validate our approach across multiple human pose estimation benchmarks. We make the code, models and data available at: https://isshikihugh.github.io/HSMR/
comment: CVPR 2025. Project Webpage: https://isshikihugh.github.io/HSMR/
☆ LeX-Art: Rethinking Text Generation via Scalable High-Quality Data Synthesis
We introduce LeX-Art, a comprehensive suite for high-quality text-image synthesis that systematically bridges the gap between prompt expressiveness and text rendering fidelity. Our approach follows a data-centric paradigm, constructing a high-quality data synthesis pipeline based on Deepseek-R1 to curate LeX-10K, a dataset of 10K high-resolution, aesthetically refined 1024$\times$1024 images. Beyond dataset construction, we develop LeX-Enhancer, a robust prompt enrichment model, and train two text-to-image models, LeX-FLUX and LeX-Lumina, achieving state-of-the-art text rendering performance. To systematically evaluate visual text generation, we introduce LeX-Bench, a benchmark that assesses fidelity, aesthetics, and alignment, complemented by Pairwise Normalized Edit Distance (PNED), a novel metric for robust text accuracy evaluation. Experiments demonstrate significant improvements, with LeX-Lumina achieving a 79.81% PNED gain on CreateBench, and LeX-FLUX outperforming baselines in color (+3.18%), positional (+4.45%), and font accuracy (+3.81%). Our codes, models, datasets, and demo are publicly available.
comment: Project page: https://zhaoshitian.github.io/lexart/
☆ CTRL-O: Language-Controllable Object-Centric Visual Representation Learning CVPR 2025
Object-centric representation learning aims to decompose visual scenes into fixed-size vectors called "slots" or "object files", where each slot captures a distinct object. Current state-of-the-art object-centric models have shown remarkable success in object discovery in diverse domains, including complex real-world scenes. However, these models suffer from a key limitation: they lack controllability. Specifically, current object-centric models learn representations based on their preconceived understanding of objects, without allowing user input to guide which objects are represented. Introducing controllability into object-centric models could unlock a range of useful capabilities, such as the ability to extract instance-specific representations from a scene. In this work, we propose a novel approach for user-directed control over slot representations by conditioning slots on language descriptions. The proposed ConTRoLlable Object-centric representation learning approach, which we term CTRL-O, achieves targeted object-language binding in complex real-world scenes without requiring mask supervision. Next, we apply these controllable slot representations on two downstream vision language tasks: text-to-image generation and visual question answering. The proposed approach enables instance-specific text-to-image generation and also achieves strong performance on visual question answering.
comment: Accepted at CVPR 2025
3DGen-Bench: Comprehensive Benchmark Suite for 3D Generative Models
3D generation is experiencing rapid advancements, while the development of 3D evaluation has not kept pace. How to keep automatic evaluation equitably aligned with human perception has become a well-recognized challenge. Recent advances in the field of language and image generation have explored human preferences and showcased respectable fitting ability. However, the 3D domain still lacks such a comprehensive preference dataset over generative models. To mitigate this absence, we develop 3DGen-Arena, an integrated platform in a battle manner. Then, we carefully design diverse text and image prompts and leverage the arena platform to gather human preferences from both public users and expert annotators, resulting in a large-scale multi-dimension human preference dataset 3DGen-Bench. Using this dataset, we further train a CLIP-based scoring model, 3DGen-Score, and a MLLM-based automatic evaluator, 3DGen-Eval. These two models innovatively unify the quality evaluation of text-to-3D and image-to-3D generation, and jointly form our automated evaluation system with their respective strengths. Extensive experiments demonstrate the efficacy of our scoring model in predicting human preferences, exhibiting a superior correlation with human ranks compared to existing metrics. We believe that our 3DGen-Bench dataset and automated evaluation system will foster a more equitable evaluation in the field of 3D generation, further promoting the development of 3D generative models and their downstream applications.
☆ SparseFlex: High-Resolution and Arbitrary-Topology 3D Shape Modeling
Creating high-fidelity 3D meshes with arbitrary topology, including open surfaces and complex interiors, remains a significant challenge. Existing implicit field methods often require costly and detail-degrading watertight conversion, while other approaches struggle with high resolutions. This paper introduces SparseFlex, a novel sparse-structured isosurface representation that enables differentiable mesh reconstruction at resolutions up to $1024^3$ directly from rendering losses. SparseFlex combines the accuracy of Flexicubes with a sparse voxel structure, focusing computation on surface-adjacent regions and efficiently handling open surfaces. Crucially, we introduce a frustum-aware sectional voxel training strategy that activates only relevant voxels during rendering, dramatically reducing memory consumption and enabling high-resolution training. This also allows, for the first time, the reconstruction of mesh interiors using only rendering supervision. Building upon this, we demonstrate a complete shape modeling pipeline by training a variational autoencoder (VAE) and a rectified flow transformer for high-quality 3D shape generation. Our experiments show state-of-the-art reconstruction accuracy, with a ~82% reduction in Chamfer Distance and a ~88% increase in F-score compared to previous methods, and demonstrate the generation of high-resolution, detailed 3D shapes with arbitrary topology. By enabling high-resolution, differentiable mesh reconstruction and generation with rendering losses, SparseFlex significantly advances the state-of-the-art in 3D shape representation and modeling.
comment: Project page: https://xianglonghe.github.io/TripoSF
☆ OccRobNet : Occlusion Robust Network for Accurate 3D Interacting Hand-Object Pose Estimation
Occlusion is one of the challenging issues when estimating 3D hand pose. This problem becomes more prominent when hand interacts with an object or two hands are involved. In the past works, much attention has not been given to these occluded regions. But these regions contain important and beneficial information that is vital for 3D hand pose estimation. Thus, in this paper, we propose an occlusion robust and accurate method for the estimation of 3D hand-object pose from the input RGB image. Our method includes first localising the hand joints using a CNN based model and then refining them by extracting contextual information. The self attention transformer then identifies the specific joints along with the hand identity. This helps the model to identify the hand belongingness of a particular joint which helps to detect the joint even in the occluded region. Further, these joints with hand identity are then used to estimate the pose using cross attention mechanism. Thus, by identifying the joints in the occluded region, the obtained network becomes robust to occlusion. Hence, this network achieves state-of-the-art results when evaluated on the InterHand2.6M, HO3D and H$_2$O3D datasets.
comment: Accepted in NATIONAL CONFERENCE ON COMMUNICATIONS (NCC) 2025
☆ Evaluating Text-to-Image Synthesis with a Conditional Fréchet Distance
Evaluating text-to-image synthesis is challenging due to misalignment between established metrics and human preferences. We propose cFreD, a metric based on the notion of Conditional Fr\'echet Distance that explicitly accounts for both visual fidelity and text-prompt alignment. Existing metrics such as Inception Score (IS), Fr\'echet Inception Distance (FID) and CLIPScore assess either image quality or image-text alignment but not both which limits their correlation with human preferences. Scoring models explicitly trained to replicate human preferences require constant updates and may not generalize to novel generation techniques or out-of-domain inputs. Through extensive experiments across multiple recently proposed text-to-image models and diverse prompt datasets, we demonstrate that cFreD exhibits a higher correlation with human judgments compared to statistical metrics, including metrics trained with human preferences. Our findings validate cFreD as a robust, future-proof metric for the systematic evaluation of text-to-image models, standardizing benchmarking in this rapidly evolving field. We release our evaluation toolkit and benchmark in the appendix.
☆ MAVERIX: Multimodal Audio-Visual Evaluation Reasoning IndeX
Frontier models have either been language-only or have primarily focused on vision and language modalities. Although recent advancements in models with vision and audio understanding capabilities have shown substantial progress, the field lacks a standardized evaluation framework for thoroughly assessing their cross-modality perception performance. We introduce MAVERIX~(Multimodal Audio-Visual Evaluation Reasoning IndeX), a novel benchmark with 700 videos and 2,556 questions explicitly designed to evaluate multimodal models through tasks that necessitate close integration of video and audio information. MAVERIX uniquely provides models with audiovisual tasks, closely mimicking the multimodal perceptual experiences available to humans during inference and decision-making processes. To our knowledge, MAVERIX is the first benchmark aimed explicitly at assessing comprehensive audiovisual integration. Experiments with state-of-the-art models, including Gemini 1.5 Pro and o1, show performance approaching human levels (around 70% accuracy), while human experts reach near-ceiling performance (95.1%). With standardized evaluation protocols, a rigorously annotated pipeline, and a public toolkit, MAVERIX establishes a challenging testbed for advancing audiovisual multimodal intelligence.
☆ Embodied-Reasoner: Synergizing Visual Search, Reasoning, and Action for Embodied Interactive Tasks
Recent advances in deep thinking models have demonstrated remarkable reasoning capabilities on mathematical and coding tasks. However, their effectiveness in embodied domains which require continuous interaction with environments through image action interleaved trajectories remains largely -unexplored. We present Embodied Reasoner, a model that extends o1 style reasoning to interactive embodied search tasks. Unlike mathematical reasoning that relies primarily on logical deduction, embodied scenarios demand spatial understanding, temporal reasoning, and ongoing self-reflection based on interaction history. To address these challenges, we synthesize 9.3k coherent Observation-Thought-Action trajectories containing 64k interactive images and 90k diverse thinking processes (analysis, spatial reasoning, reflection, planning, and verification). We develop a three-stage training pipeline that progressively enhances the model's capabilities through imitation learning, self-exploration via rejection sampling, and self-correction through reflection tuning. The evaluation shows that our model significantly outperforms those advanced visual reasoning models, e.g., it exceeds OpenAI o1, o3-mini, and Claude-3.7 by +9\%, 24\%, and +13\%. Analysis reveals our model exhibits fewer repeated searches and logical inconsistencies, with particular advantages in complex long-horizon tasks. Real-world environments also show our superiority while exhibiting fewer repeated searches and logical inconsistency cases.
comment: Code: https://github.com/zwq2018/embodied_reasoner Dataset: https://huggingface.co/datasets/zwq2018/embodied_reasoner
☆ AMA-SAM: Adversarial Multi-Domain Alignment of Segment Anything Model for High-Fidelity Histology Nuclei Segmentation
Accurate segmentation of cell nuclei in histopathology images is essential for numerous biomedical research and clinical applications. However, existing cell nucleus segmentation methods only consider a single dataset (i.e., primary domain), while neglecting to leverage supplementary data from diverse sources (i.e., auxiliary domains) to reduce overfitting and enhance the performance. Although incorporating multiple datasets could alleviate overfitting, it often exacerbates performance drops caused by domain shifts. In this work, we introduce Adversarial Multi-domain Alignment of Segment Anything Model (AMA-SAM) that extends the Segment Anything Model (SAM) to overcome these obstacles through two key innovations. First, we propose a Conditional Gradient Reversal Layer (CGRL), a multi-domain alignment module that harmonizes features from diverse domains to promote domain-invariant representation learning while preserving crucial discriminative features for the primary dataset. Second, we address SAM's inherent low-resolution output by designing a High-Resolution Decoder (HR-Decoder), which directly produces fine-grained segmentation maps in order to capture intricate nuclei boundaries in high-resolution histology images. To the best of our knowledge, this is the first attempt to adapt SAM for multi-dataset learning with application to histology nuclei segmentation. We validate our method on several publicly available datasets, demonstrating consistent and significant improvements over state-of-the-art approaches.
comment: 13 pages, 4 tables, 2 figures
☆ Progressive Rendering Distillation: Adapting Stable Diffusion for Instant Text-to-Mesh Generation without 3D Data CVPR 2025
It is highly desirable to obtain a model that can generate high-quality 3D meshes from text prompts in just seconds. While recent attempts have adapted pre-trained text-to-image diffusion models, such as Stable Diffusion (SD), into generators of 3D representations (e.g., Triplane), they often suffer from poor quality due to the lack of sufficient high-quality 3D training data. Aiming at overcoming the data shortage, we propose a novel training scheme, termed as Progressive Rendering Distillation (PRD), eliminating the need for 3D ground-truths by distilling multi-view diffusion models and adapting SD into a native 3D generator. In each iteration of training, PRD uses the U-Net to progressively denoise the latent from random noise for a few steps, and in each step it decodes the denoised latent into 3D output. Multi-view diffusion models, including MVDream and RichDreamer, are used in joint with SD to distill text-consistent textures and geometries into the 3D outputs through score distillation. Since PRD supports training without 3D ground-truths, we can easily scale up the training data and improve generation quality for challenging text prompts with creative concepts. Meanwhile, PRD can accelerate the inference speed of the generation model in just a few steps. With PRD, we train a Triplane generator, namely TriplaneTurbo, which adds only $2.5\%$ trainable parameters to adapt SD for Triplane generation. TriplaneTurbo outperforms previous text-to-3D generators in both efficiency and quality. Specifically, it can produce high-quality 3D meshes in 1.2 seconds and generalize well for challenging text input. The code is available at https://github.com/theEricMa/TriplaneTurbo.
comment: Accepted to CVPR 2025. Code:https://github.com/theEricMa/TriplaneTurbo. Demo:https://huggingface.co/spaces/ZhiyuanthePony/TriplaneTurbo
☆ RapidPoseTriangulation: Multi-view Multi-person Whole-body Human Pose Triangulation in a Millisecond
The integration of multi-view imaging and pose estimation represents a significant advance in computer vision applications, offering new possibilities for understanding human movement and interactions. This work presents a new algorithm that improves multi-view multi-person pose estimation, focusing on fast triangulation speeds and good generalization capabilities. The approach extends to whole-body pose estimation, capturing details from facial expressions to finger movements across multiple individuals and viewpoints. Adaptability to different settings is demonstrated through strong performance across unseen datasets and configurations. To support further progress in this field, all of this work is publicly accessible.
☆ CMED: A Child Micro-Expression Dataset
Micro-expressions are short bursts of emotion that are difficult to hide. Their detection in children is an important cue to assist psychotherapists in conducting better therapy. However, existing research on the detection of micro-expressions has focused on adults, whose expressions differ in their characteristics from those of children. The lack of research is a direct consequence of the lack of a child-based micro-expressions dataset as it is much more challenging to capture children's facial expressions due to the lack of predictability and controllability. This study compiles a dataset of spontaneous child micro-expression videos, the first of its kind, to the best of the authors knowledge. The dataset is captured in the wild using video conferencing software. This dataset enables us to then explore key features and differences between adult and child micro-expressions. This study also establishes a baseline for the automated spotting and recognition of micro-expressions in children using three approaches comprising of hand-created and learning-based approaches.
☆ Cognitive Science-Inspired Evaluation of Core Capabilities for Object Understanding in AI
One of the core components of our world models is 'intuitive physics' - an understanding of objects, space, and causality. This capability enables us to predict events, plan action and navigate environments, all of which rely on a composite sense of objecthood. Despite its importance, there is no single, unified account of objecthood, though multiple theoretical frameworks provide insights. In the first part of this paper, we present a comprehensive overview of the main theoretical frameworks in objecthood research - Gestalt psychology, enactive cognition, and developmental psychology - and identify the core capabilities each framework attributes to object understanding, as well as what functional roles they play in shaping world models in biological agents. Given the foundational role of objecthood in world modelling, understanding objecthood is also essential in AI. In the second part of the paper, we evaluate how current AI paradigms approach and test objecthood capabilities compared to those in cognitive science. We define an AI paradigm as a combination of how objecthood is conceptualised, the methods used for studying objecthood, the data utilised, and the evaluation techniques. We find that, whilst benchmarks can detect that AI systems model isolated aspects of objecthood, the benchmarks cannot detect when AI systems lack functional integration across these capabilities, not solving the objecthood challenge fully. Finally, we explore novel evaluation approaches that align with the integrated vision of objecthood outlined in this paper. These methods are promising candidates for advancing from isolated object capabilities toward general-purpose AI with genuine object understanding in real-world contexts.
☆ InteractionMap: Improving Online Vectorized HDMap Construction with Interaction
Vectorized high-definition (HD) maps are essential for an autonomous driving system. Recently, state-of-the-art map vectorization methods are mainly based on DETR-like framework to generate HD maps in an end-to-end manner. In this paper, we propose InteractionMap, which improves previous map vectorization methods by fully leveraging local-to-global information interaction in both time and space. Firstly, we explore enhancing DETR-like detectors by explicit position relation prior from point-level to instance-level, since map elements contain strong shape priors. Secondly, we propose a key-frame-based hierarchical temporal fusion module, which interacts temporal information from local to global. Lastly, the separate classification branch and regression branch lead to the problem of misalignment in the output distribution. We interact semantic information with geometric information by introducing a novel geometric-aware classification loss in optimization and a geometric-aware matching cost in label assignment. InteractionMap achieves state-of-the-art performance on both nuScenes and Argoverse2 benchmarks.
☆ When Astronomy Meets AI: Manazel For Crescent Visibility Prediction in Morocco
The accurate determination of the beginning of each Hijri month is essential for religious, cultural, and administrative purposes. Manazel (The code and datasets are available at https://github.com/lairgiyassir/manazel) addresses this challenge in Morocco by leveraging 13 years of crescent visibility data to refine the ODEH criterion, a widely used standard for lunar crescent visibility prediction. The study integrates two key features, the Arc of Vision (ARCV) and the total width of the crescent (W), to enhance the accuracy of lunar visibility assessments. A machine learning approach utilizing the Logistic Regression algorithm is employed to classify crescent visibility conditions, achieving a predictive accuracy of 98.83%. This data-driven methodology offers a robust and reliable framework for determining the start of the Hijri month, comparing different data classification tools, and improving the consistency of lunar calendar calculations in Morocco. The findings demonstrate the effectiveness of machine learning in astronomical applications and highlight the potential for further enhancements in the modeling of crescent visibility.
☆ The MVTec AD 2 Dataset: Advanced Scenarios for Unsupervised Anomaly Detection CVPR 2025
In recent years, performance on existing anomaly detection benchmarks like MVTec AD and VisA has started to saturate in terms of segmentation AU-PRO, with state-of-the-art models often competing in the range of less than one percentage point. This lack of discriminatory power prevents a meaningful comparison of models and thus hinders progress of the field, especially when considering the inherent stochastic nature of machine learning results. We present MVTec AD 2, a collection of eight anomaly detection scenarios with more than 8000 high-resolution images. It comprises challenging and highly relevant industrial inspection use cases that have not been considered in previous datasets, including transparent and overlapping objects, dark-field and back light illumination, objects with high variance in the normal data, and extremely small defects. We provide comprehensive evaluations of state-of-the-art methods and show that their performance remains below 60% average AU-PRO. Additionally, our dataset provides test scenarios with lighting condition changes to assess the robustness of methods under real-world distribution shifts. We host a publicly accessible evaluation server that holds the pixel-precise ground truth of the test set (https://benchmark.mvtec.com/). All image data is available at https://www.mvtec.com/company/research/datasets/mvtec-ad-2.
comment: paper under review; dataset first released for the VAND3.0 challenge @ CVPR 2025 https://sites.google.com/view/vand30cvpr2025/challenge
☆ Audio-driven Gesture Generation via Deviation Feature in the Latent Space
Gestures are essential for enhancing co-speech communication, offering visual emphasis and complementing verbal interactions. While prior work has concentrated on point-level motion or fully supervised data-driven methods, we focus on co-speech gestures, advocating for weakly supervised learning and pixel-level motion deviations. We introduce a weakly supervised framework that learns latent representation deviations, tailored for co-speech gesture video generation. Our approach employs a diffusion model to integrate latent motion features, enabling more precise and nuanced gesture representation. By leveraging weakly supervised deviations in latent space, we effectively generate hand gestures and mouth movements, crucial for realistic video production. Experiments show our method significantly improves video quality, surpassing current state-of-the-art techniques.
comment: 6 pages, 5 figures
☆ FusionSegReID: Advancing Person Re-Identification with Multimodal Retrieval and Precise Segmentation
Person re-identification (ReID) plays a critical role in applications like security surveillance and criminal investigations by matching individuals across large image galleries captured by non-overlapping cameras. Traditional ReID methods rely on unimodal inputs, typically images, but face limitations due to challenges like occlusions, lighting changes, and pose variations. While advancements in image-based and text-based ReID systems have been made, the integration of both modalities has remained under-explored. This paper presents FusionSegReID, a multimodal model that combines both image and text inputs for enhanced ReID performance. By leveraging the complementary strengths of these modalities, our model improves matching accuracy and robustness, particularly in complex, real-world scenarios where one modality may struggle. Our experiments show significant improvements in Top-1 accuracy and mean Average Precision (mAP) for ReID, as well as better segmentation results in challenging scenarios like occlusion and low-quality images. Ablation studies further confirm that multimodal fusion and segmentation modules contribute to enhanced re-identification and mask accuracy. The results show that FusionSegReID outperforms traditional unimodal models, offering a more robust and flexible solution for real-world person ReID tasks.
☆ AlignDiff: Learning Physically-Grounded Camera Alignment via Diffusion
Accurate camera calibration is a fundamental task for 3D perception, especially when dealing with real-world, in-the-wild environments where complex optical distortions are common. Existing methods often rely on pre-rectified images or calibration patterns, which limits their applicability and flexibility. In this work, we introduce a novel framework that addresses these challenges by jointly modeling camera intrinsic and extrinsic parameters using a generic ray camera model. Unlike previous approaches, AlignDiff shifts focus from semantic to geometric features, enabling more accurate modeling of local distortions. We propose AlignDiff, a diffusion model conditioned on geometric priors, enabling the simultaneous estimation of camera distortions and scene geometry. To enhance distortion prediction, we incorporate edge-aware attention, focusing the model on geometric features around image edges, rather than semantic content. Furthermore, to enhance generalizability to real-world captures, we incorporate a large database of ray-traced lenses containing over three thousand samples. This database characterizes the distortion inherent in a diverse variety of lens forms. Our experiments demonstrate that the proposed method significantly reduces the angular error of estimated ray bundles by ~8.2 degrees and overall calibration accuracy, outperforming existing approaches on challenging, real-world datasets.
☆ Bearing fault diagnosis based on multi-scale spectral images and convolutional neural network
To address the challenges of low diagnostic accuracy in traditional bearing fault diagnosis methods, this paper proposes a novel fault diagnosis approach based on multi-scale spectrum feature images and deep learning. Firstly, the vibration signal are preprocessed through mean removal and then converted to multi-length spectrum with fast Fourier transforms (FFT). Secondly, a novel feature called multi-scale spectral image (MSSI) is constructed by multi-length spectrum paving scheme. Finally, a deep learning framework, convolutional neural network (CNN), is formulated to diagnose the bearing faults. Two experimental cases are utilized to verify the effectiveness of the proposed method. Experimental results demonstrate that the proposed method significantly improves the accuracy of fault diagnosis.
comment: 12pages, 10 figures and 8 tables
☆ uLayout: Unified Room Layout Estimation for Perspective and Panoramic Images WACV-2025
We present uLayout, a unified model for estimating room layout geometries from both perspective and panoramic images, whereas traditional solutions require different model designs for each image type. The key idea of our solution is to unify both domains into the equirectangular projection, particularly, allocating perspective images into the most suitable latitude coordinate to effectively exploit both domains seamlessly. To address the Field-of-View (FoV) difference between the input domains, we design uLayout with a shared feature extractor with an extra 1D-Convolution layer to condition each domain input differently. This conditioning allows us to efficiently formulate a column-wise feature regression problem regardless of the FoV input. This simple yet effective approach achieves competitive performance with current state-of-the-art solutions and shows for the first time a single end-to-end model for both domains. Extensive experiments in the real-world datasets, LSUN, Matterport3D, PanoContext, and Stanford 2D-3D evidence the contribution of our approach. Code is available at https://github.com/JonathanLee112/uLayout.
comment: Accepted to WACV-2025
☆ SyncSDE: A Probabilistic Framework for Diffusion Synchronization CVPR2025
There have been many attempts to leverage multiple diffusion models for collaborative generation, extending beyond the original domain. A prominent approach involves synchronizing multiple diffusion trajectories by mixing the estimated scores to artificially correlate the generation processes. However, existing methods rely on naive heuristics, such as averaging, without considering task specificity. These approaches do not clarify why such methods work and often fail when a heuristic suitable for one task is blindly applied to others. In this paper, we present a probabilistic framework for analyzing why diffusion synchronization works and reveal where heuristics should be focused - modeling correlations between multiple trajectories and adapting them to each specific task. We further identify optimal correlation models per task, achieving better results than previous approaches that apply a single heuristic across all tasks without justification.
comment: Accepted to CVPR2025
☆ LOCATEdit: Graph Laplacian Optimized Cross Attention for Localized Text-Guided Image Editing
Text-guided image editing aims to modify specific regions of an image according to natural language instructions while maintaining the general structure and the background fidelity. Existing methods utilize masks derived from cross-attention maps generated from diffusion models to identify the target regions for modification. However, since cross-attention mechanisms focus on semantic relevance, they struggle to maintain the image integrity. As a result, these methods often lack spatial consistency, leading to editing artifacts and distortions. In this work, we address these limitations and introduce LOCATEdit, which enhances cross-attention maps through a graph-based approach utilizing self-attention-derived patch relationships to maintain smooth, coherent attention across image regions, ensuring that alterations are limited to the designated items while retaining the surrounding structure. \method consistently and substantially outperforms existing baselines on PIE-Bench, demonstrating its state-of-the-art performance and effectiveness on various editing tasks. Code can be found on https://github.com/LOCATEdit/LOCATEdit/
☆ ICG-MVSNet: Learning Intra-view and Cross-view Relationships for Guidance in Multi-View Stereo
Multi-view Stereo (MVS) aims to estimate depth and reconstruct 3D point clouds from a series of overlapping images. Recent learning-based MVS frameworks overlook the geometric information embedded in features and correlations, leading to weak cost matching. In this paper, we propose ICG-MVSNet, which explicitly integrates intra-view and cross-view relationships for depth estimation. Specifically, we develop an intra-view feature fusion module that leverages the feature coordinate correlations within a single image to enhance robust cost matching. Additionally, we introduce a lightweight cross-view aggregation module that efficiently utilizes the contextual information from volume correlations to guide regularization. Our method is evaluated on the DTU dataset and Tanks and Temples benchmark, consistently achieving competitive performance against state-of-the-art works, while requiring lower computational resources.
☆ Uncertainty-aware Bayesian machine learning modelling of land cover classification
Land cover classification involves the production of land cover maps, which determine the type of land through remote sensing imagery. Over recent years, such classification is being performed by machine learning classification models, which can give highly accurate predictions on land cover per pixel using large quantities of input training data. However, such models do not currently take account of input measurement uncertainty, which is vital for traceability in metrology. In this work we propose a Bayesian classification framework using generative modelling to take account of input measurement uncertainty. We take the specific case of Bayesian quadratic discriminant analysis, and apply it to land cover datasets from Copernicus Sentinel-2 in 2020 and 2021. We benchmark the performance of the model against more popular classification models used in land cover maps such as random forests and neural networks. We find that such Bayesian models are more trustworthy, in the sense that they are more interpretable, explicitly model the input measurement uncertainty, and maintain predictive performance of class probability outputs across datasets of different years and sizes, whilst also being computationally efficient.
comment: 31 pages, 10 figures
☆ Fine-Grained Evaluation of Large Vision-Language Models in Autonomous Driving
Existing benchmarks for Vision-Language Model (VLM) on autonomous driving (AD) primarily assess interpretability through open-form visual question answering (QA) within coarse-grained tasks, which remain insufficient to assess capabilities in complex driving scenarios. To this end, we introduce $\textbf{VLADBench}$, a challenging and fine-grained dataset featuring close-form QAs that progress from static foundational knowledge and elements to advanced reasoning for dynamic on-road situations. The elaborate $\textbf{VLADBench}$ spans 5 key domains: Traffic Knowledge Understanding, General Element Recognition, Traffic Graph Generation, Target Attribute Comprehension, and Ego Decision-Making and Planning. These domains are further broken down into 11 secondary aspects and 29 tertiary tasks for a granular evaluation. A thorough assessment of general and domain-specific (DS) VLMs on this benchmark reveals both their strengths and critical limitations in AD contexts. To further exploit the cognitive and reasoning interactions among the 5 domains for AD understanding, we start from a small-scale VLM and train the DS models on individual domain datasets (collected from 1.4M DS QAs across public sources). The experimental results demonstrate that the proposed benchmark provides a crucial step toward a more comprehensive assessment of VLMs in AD, paving the way for the development of more cognitively sophisticated and reasoning-capable AD systems.
☆ Keyword-Oriented Multimodal Modeling for Euphemism Identification
Euphemism identification deciphers the true meaning of euphemisms, such as linking "weed" (euphemism) to "marijuana" (target keyword) in illicit texts, aiding content moderation and combating underground markets. While existing methods are primarily text-based, the rise of social media highlights the need for multimodal analysis, incorporating text, images, and audio. However, the lack of multimodal datasets for euphemisms limits further research. To address this, we regard euphemisms and their corresponding target keywords as keywords and first introduce a keyword-oriented multimodal corpus of euphemisms (KOM-Euph), involving three datasets (Drug, Weapon, and Sexuality), including text, images, and speech. We further propose a keyword-oriented multimodal euphemism identification method (KOM-EI), which uses cross-modal feature alignment and dynamic fusion modules to explicitly utilize the visual and audio features of the keywords for efficient euphemism identification. Extensive experiments demonstrate that KOM-EI outperforms state-of-the-art models and large language models, and show the importance of our multimodal datasets.
☆ Double Blind Imaging with Generative Modeling
Blind inverse problems in imaging arise from uncertainties in the system used to collect (noisy) measurements of images. Recovering clean images from these measurements typically requires identifying the imaging system, either implicitly or explicitly. A common solution leverages generative models as priors for both the images and the imaging system parameters (e.g., a class of point spread functions). To learn these priors in a straightforward manner requires access to a dataset of clean images as well as samples of the imaging system. We propose an AmbientGAN-based generative technique to identify the distribution of parameters in unknown imaging systems, using only unpaired clean images and corrupted measurements. This learned distribution can then be used in model-based recovery algorithms to solve blind inverse problems such as blind deconvolution. We successfully demonstrate our technique for learning Gaussian blur and motion blur priors from noisy measurements and show their utility in solving blind deconvolution with diffusion posterior sampling.
☆ Shape Modeling of Longitudinal Medical Images: From Diffeomorphic Metric Mapping to Deep Learning
Living biological tissue is a complex system, constantly growing and changing in response to external and internal stimuli. These processes lead to remarkable and intricate changes in shape. Modeling and understanding both natural and pathological (or abnormal) changes in the shape of anatomical structures is highly relevant, with applications in diagnostic, prognostic, and therapeutic healthcare. Nevertheless, modeling the longitudinal shape change of biological tissue is a non-trivial task due to its inherent nonlinear nature. In this review, we highlight several existing methodologies and tools for modeling longitudinal shape change (i.e., spatiotemporal shape modeling). These methods range from diffeomorphic metric mapping to deep-learning based approaches (e.g., autoencoders, generative networks, recurrent neural networks, etc.). We discuss the synergistic combinations of existing technologies and potential directions for future research, underscoring key deficiencies in the current research landscape.
☆ Invert2Restore: Zero-Shot Degradation-Blind Image Restoration
Two of the main challenges of image restoration in real-world scenarios are the accurate characterization of an image prior and the precise modeling of the image degradation operator. Pre-trained diffusion models have been very successfully used as image priors in zero-shot image restoration methods. However, how to best handle the degradation operator is still an open problem. In real-world data, methods that rely on specific parametric assumptions about the degradation model often face limitations in their applicability. To address this, we introduce Invert2Restore, a zero-shot, training-free method that operates in both fully blind and partially blind settings -- requiring no prior knowledge of the degradation model or only partial knowledge of its parametric form without known parameters. Despite this, Invert2Restore achieves high-fidelity results and generalizes well across various types of image degradation. It leverages a pre-trained diffusion model as a deterministic mapping between normal samples and undistorted image samples. The key insight is that the input noise mapped by a diffusion model to a degraded image lies in a low-probability density region of the standard normal distribution. Thus, we can restore the degraded image by carefully guiding its input noise toward a higher-density region. We experimentally validate Invert2Restore across several image restoration tasks, demonstrating that it achieves state-of-the-art performance in scenarios where the degradation operator is either unknown or partially known.
☆ BOLT: Boost Large Vision-Language Model Without Training for Long-form Video Understanding CVPR 2025
Large video-language models (VLMs) have demonstrated promising progress in various video understanding tasks. However, their effectiveness in long-form video analysis is constrained by limited context windows. Traditional approaches, such as uniform frame sampling, often inevitably allocate resources to irrelevant content, diminishing their effectiveness in real-world scenarios. In this paper, we introduce BOLT, a method to BOost Large VLMs without additional Training through a comprehensive study of frame selection strategies. First, to enable a more realistic evaluation of VLMs in long-form video understanding, we propose a multi-source retrieval evaluation setting. Our findings reveal that uniform sampling performs poorly in noisy contexts, underscoring the importance of selecting the right frames. Second, we explore several frame selection strategies based on query-frame similarity and analyze their effectiveness at inference time. Our results show that inverse transform sampling yields the most significant performance improvement, increasing accuracy on the Video-MME benchmark from 53.8% to 56.1% and MLVU benchmark from 58.9% to 63.4%. Our code is available at https://github.com/sming256/BOLT.
comment: Accepted to CVPR 2025
☆ Fine-Grained Behavior and Lane Constraints Guided Trajectory Prediction Method
Trajectory prediction, as a critical component of autonomous driving systems, has attracted the attention of many researchers. Existing prediction algorithms focus on extracting more detailed scene features or selecting more reasonable trajectory destinations. However, in the face of dynamic and evolving future movements of the target vehicle, these algorithms cannot provide a fine-grained and continuous description of future behaviors and lane constraints, which degrades the prediction accuracy. To address this challenge, we present BLNet, a novel dualstream architecture that synergistically integrates behavioral intention recognition and lane constraint modeling through parallel attention mechanisms. The framework generates fine-grained behavior state queries (capturing spatial-temporal movement patterns) and lane queries (encoding lane topology constraints), supervised by two auxiliary losses, respectively. Subsequently, a two-stage decoder first produces trajectory proposals, then performs point-level refinement by jointly incorporating both the continuity of passed lanes and future motion features. Extensive experiments on two large datasets, nuScenes and Argoverse, show that our network exhibits significant performance gains over existing direct regression and goal-based algorithms.
comment: This work has been submitted to the IEEE TIM for possible publication
☆ Embedding Compression Distortion in Video Coding for Machines
Currently, video transmission serves not only the Human Visual System (HVS) for viewing but also machine perception for analysis. However, existing codecs are primarily optimized for pixel-domain and HVS-perception metrics rather than the needs of machine vision tasks. To address this issue, we propose a Compression Distortion Representation Embedding (CDRE) framework, which extracts machine-perception-related distortion representation and embeds it into downstream models, addressing the information lost during compression and improving task performance. Specifically, to better analyze the machine-perception-related distortion, we design a compression-sensitive extractor that identifies compression degradation in the feature domain. For efficient transmission, a lightweight distortion codec is introduced to compress the distortion information into a compact representation. Subsequently, the representation is progressively embedded into the downstream model, enabling it to be better informed about compression degradation and enhancing performance. Experiments across various codecs and downstream tasks demonstrate that our framework can effectively boost the rate-task performance of existing codecs with minimal overhead in terms of bitrate, execution time, and number of parameters. Our codes and supplementary materials are released in https://github.com/Ws-Syx/CDRE/.
☆ Retinal Fundus Multi-Disease Image Classification using Hybrid CNN-Transformer-Ensemble Architectures
Our research is motivated by the urgent global issue of a large population affected by retinal diseases, which are evenly distributed but underserved by specialized medical expertise, particularly in non-urban areas. Our primary objective is to bridge this healthcare gap by developing a comprehensive diagnostic system capable of accurately predicting retinal diseases solely from fundus images. However, we faced significant challenges due to limited, diverse datasets and imbalanced class distributions. To overcome these issues, we have devised innovative strategies. Our research introduces novel approaches, utilizing hybrid models combining deeper Convolutional Neural Networks (CNNs), Transformer encoders, and ensemble architectures sequentially and in parallel to classify retinal fundus images into 20 disease labels. Our overarching goal is to assess these advanced models' potential in practical applications, with a strong focus on enhancing retinal disease diagnosis accuracy across a broader spectrum of conditions. Importantly, our efforts have surpassed baseline model results, with the C-Tran ensemble model emerging as the leader, achieving a remarkable model score of 0.9166, surpassing the baseline score of 0.9. Additionally, experiments with the IEViT model showcased equally promising outcomes with improved computational efficiency. We've also demonstrated the effectiveness of dynamic patch extraction and the integration of domain knowledge in computer vision tasks. In summary, our research strives to contribute significantly to retinal disease diagnosis, addressing the critical need for accessible healthcare solutions in underserved regions while aiming for comprehensive and accurate disease prediction.
comment: 17 pages, 3 figures, 7 tables. Conference paper presented at the International Health Informatics Conference (IHIC 2023)
☆ RoadSocial: A Diverse VideoQA Dataset and Benchmark for Road Event Understanding from Social Video Narratives CVPR 2025
We introduce RoadSocial, a large-scale, diverse VideoQA dataset tailored for generic road event understanding from social media narratives. Unlike existing datasets limited by regional bias, viewpoint bias and expert-driven annotations, RoadSocial captures the global complexity of road events with varied geographies, camera viewpoints (CCTV, handheld, drones) and rich social discourse. Our scalable semi-automatic annotation framework leverages Text LLMs and Video LLMs to generate comprehensive question-answer pairs across 12 challenging QA tasks, pushing the boundaries of road event understanding. RoadSocial is derived from social media videos spanning 14M frames and 414K social comments, resulting in a dataset with 13.2K videos, 674 tags and 260K high-quality QA pairs. We evaluate 18 Video LLMs (open-source and proprietary, driving-specific and general-purpose) on our road event understanding benchmark. We also demonstrate RoadSocial's utility in improving road event understanding capabilities of general-purpose Video LLMs.
comment: Accepted at CVPR 2025; Project Page: https://roadsocial.github.io/
☆ FaceBench: A Multi-View Multi-Level Facial Attribute VQA Dataset for Benchmarking Face Perception MLLMs CVPR2025
Multimodal large language models (MLLMs) have demonstrated remarkable capabilities in various tasks. However, effectively evaluating these MLLMs on face perception remains largely unexplored. To address this gap, we introduce FaceBench, a dataset featuring hierarchical multi-view and multi-level attributes specifically designed to assess the comprehensive face perception abilities of MLLMs. Initially, we construct a hierarchical facial attribute structure, which encompasses five views with up to three levels of attributes, totaling over 210 attributes and 700 attribute values. Based on the structure, the proposed FaceBench consists of 49,919 visual question-answering (VQA) pairs for evaluation and 23,841 pairs for fine-tuning. Moreover, we further develop a robust face perception MLLM baseline, Face-LLaVA, by training with our proposed face VQA data. Extensive experiments on various mainstream MLLMs and Face-LLaVA are conducted to test their face perception ability, with results also compared against human performance. The results reveal that, the existing MLLMs are far from satisfactory in understanding the fine-grained facial attributes, while our Face-LLaVA significantly outperforms existing open-source models with a small amount of training data and is comparable to commercial ones like GPT-4o and Gemini. The dataset will be released at https://github.com/CVI-SZU/FaceBench.
comment: Accepted by CVPR2025
☆ Towards Generating Realistic 3D Semantic Training Data for Autonomous Driving
Semantic scene understanding is crucial for robotics and computer vision applications. In autonomous driving, 3D semantic segmentation plays an important role for enabling safe navigation. Despite significant advances in the field, the complexity of collecting and annotating 3D data is a bottleneck in this developments. To overcome that data annotation limitation, synthetic simulated data has been used to generate annotated data on demand. There is still however a domain gap between real and simulated data. More recently, diffusion models have been in the spotlight, enabling close-to-real data synthesis. Those generative models have been recently applied to the 3D data domain for generating scene-scale data with semantic annotations. Still, those methods either rely on image projection or decoupled models trained with different resolutions in a coarse-to-fine manner. Such intermediary representations impact the generated data quality due to errors added in those transformations. In this work, we propose a novel approach able to generate 3D semantic scene-scale data without relying on any projection or decoupled trained multi-resolution models, achieving more realistic semantic scene data generation compared to previous state-of-the-art methods. Besides improving 3D semantic scene-scale data synthesis, we thoroughly evaluate the use of the synthetic scene samples as labeled data to train a semantic segmentation network. In our experiments, we show that using the synthetic annotated data generated by our method as training data together with the real semantic segmentation labels, leads to an improvement in the semantic segmentation model performance. Our results show the potential of generated scene-scale point clouds to generate more training data to extend existing datasets, reducing the data annotation effort. Our code is available at https://github.com/PRBonn/3DiSS.
☆ Sparse Bayesian Learning for Label Efficiency in Cardiac Real-Time MRI
Cardiac real-time magnetic resonance imaging (MRI) is an emerging technology that images the heart at up to 50 frames per second, offering insight into the respiratory effects on the heartbeat. However, this method significantly increases the number of images that must be segmented to derive critical health indicators. Although neural networks perform well on inner slices, predictions on outer slices are often unreliable. This work proposes sparse Bayesian learning (SBL) to predict the ventricular volume on outer slices with minimal manual labeling to address this challenge. The ventricular volume over time is assumed to be dominated by sparse frequencies corresponding to the heart and respiratory rates. Moreover, SBL identifies these sparse frequencies on well-segmented inner slices by optimizing hyperparameters via type -II likelihood, automatically pruning irrelevant components. The identified sparse frequencies guide the selection of outer slice images for labeling, minimizing posterior variance. This work provides performance guarantees for the greedy algorithm. Testing on patient data demonstrates that only a few labeled images are necessary for accurate volume prediction. The labeling procedure effectively avoids selecting inefficient images. Furthermore, the Bayesian approach provides uncertainty estimates, highlighting unreliable predictions (e.g., when choosing suboptimal labels).
☆ RainyGS: Efficient Rain Synthesis with Physically-Based Gaussian Splatting
We consider the problem of adding dynamic rain effects to in-the-wild scenes in a physically-correct manner. Recent advances in scene modeling have made significant progress, with NeRF and 3DGS techniques emerging as powerful tools for reconstructing complex scenes. However, while effective for novel view synthesis, these methods typically struggle with challenging scene editing tasks, such as physics-based rain simulation. In contrast, traditional physics-based simulations can generate realistic rain effects, such as raindrops and splashes, but they often rely on skilled artists to carefully set up high-fidelity scenes. This process lacks flexibility and scalability, limiting its applicability to broader, open-world environments. In this work, we introduce RainyGS, a novel approach that leverages the strengths of both physics-based modeling and 3DGS to generate photorealistic, dynamic rain effects in open-world scenes with physical accuracy. At the core of our method is the integration of physically-based raindrop and shallow water simulation techniques within the fast 3DGS rendering framework, enabling realistic and efficient simulations of raindrop behavior, splashes, and reflections. Our method supports synthesizing rain effects at over 30 fps, offering users flexible control over rain intensity -- from light drizzles to heavy downpours. We demonstrate that RainyGS performs effectively for both real-world outdoor scenes and large-scale driving scenarios, delivering more photorealistic and physically-accurate rain effects compared to state-of-the-art methods. Project page can be found at https://pku-vcl-geometry.github.io/RainyGS/
☆ Dual-Task Learning for Dead Tree Detection and Segmentation with Hybrid Self-Attention U-Nets in Aerial Imagery
Mapping standing dead trees is critical for assessing forest health, monitoring biodiversity, and mitigating wildfire risks, for which aerial imagery has proven useful. However, dense canopy structures, spectral overlaps between living and dead vegetation, and over-segmentation errors limit the reliability of existing methods. This study introduces a hybrid postprocessing framework that refines deep learning-based tree segmentation by integrating watershed algorithms with adaptive filtering, enhancing boundary delineation, and reducing false positives in complex forest environments. Tested on high-resolution aerial imagery from boreal forests, the framework improved instance-level segmentation accuracy by 41.5% and reduced positional errors by 57%, demonstrating robust performance in densely vegetated regions. By balancing detection accuracy and over-segmentation artifacts, the method enabled the precise identification of individual dead trees, which is critical for ecological monitoring. The framework's computational efficiency supports scalable applications, such as wall-to-wall tree mortality mapping over large geographic regions using aerial or satellite imagery. These capabilities directly benefit wildfire risk assessment (identifying fuel accumulations), carbon stock estimation (tracking emissions from decaying biomass), and precision forestry (targeting salvage loggings). By bridging advanced remote sensing techniques with practical forest management needs, this work advances tools for large-scale ecological conservation and climate resilience planning.
comment: 11 pages, 4 figures, 4 tables
☆ STAMICS: Splat, Track And Map with Integrated Consistency and Semantics for Dense RGB-D SLAM
Simultaneous Localization and Mapping (SLAM) is a critical task in robotics, enabling systems to autonomously navigate and understand complex environments. Current SLAM approaches predominantly rely on geometric cues for mapping and localization, but they often fail to ensure semantic consistency, particularly in dynamic or densely populated scenes. To address this limitation, we introduce STAMICS, a novel method that integrates semantic information with 3D Gaussian representations to enhance both localization and mapping accuracy. STAMICS consists of three key components: a 3D Gaussian-based scene representation for high-fidelity reconstruction, a graph-based clustering technique that enforces temporal semantic consistency, and an open-vocabulary system that allows for the classification of unseen objects. Extensive experiments show that STAMICS significantly improves camera pose estimation and map quality, outperforming state-of-the-art methods while reducing reconstruction errors. Code will be public available.
Diffusion Image Prior
Zero-shot image restoration (IR) methods based on pretrained diffusion models have recently achieved significant success. These methods typically require at least a parametric form of the degradation model. However, in real-world scenarios, the degradation may be too complex to define explicitly. To handle this general case, we introduce the Diffusion Image Prior (DIIP). We take inspiration from the Deep Image Prior (DIP)[16], since it can be used to remove artifacts without the need for an explicit degradation model. However, in contrast to DIP, we find that pretrained diffusion models offer a much stronger prior, despite being trained without knowledge from corrupted data. We show that, the optimization process in DIIP first reconstructs a clean version of the image before eventually overfitting to the degraded input, but it does so for a broader range of degradations than DIP. In light of this result, we propose a blind image restoration (IR) method based on early stopping, which does not require prior knowledge of the degradation model. We validate DIIP on various degradation-blind IR tasks, including JPEG artifact removal, waterdrop removal, denoising and super-resolution with state-of-the-art results.
☆ VALLR: Visual ASR Language Model for Lip Reading
Lip Reading, or Visual Automatic Speech Recognition (V-ASR), is a complex task requiring the interpretation of spoken language exclusively from visual cues, primarily lip movements and facial expressions. This task is especially challenging due to the absence of auditory information and the inherent ambiguity when visually distinguishing phonemes that have overlapping visemes where different phonemes appear identical on the lips. Current methods typically attempt to predict words or characters directly from these visual cues, but this approach frequently encounters high error rates due to coarticulation effects and viseme ambiguity. We propose a novel two-stage, phoneme-centric framework for Visual Automatic Speech Recognition (V-ASR) that addresses these longstanding challenges. First, our model predicts a compact sequence of phonemes from visual inputs using a Video Transformer with a CTC head, thereby reducing the task complexity and achieving robust speaker invariance. This phoneme output then serves as the input to a fine-tuned Large Language Model (LLM), which reconstructs coherent words and sentences by leveraging broader linguistic context. Unlike existing methods that either predict words directly-often faltering on visually similar phonemes-or rely on large-scale multimodal pre-training, our approach explicitly encodes intermediate linguistic structure while remaining highly data efficient. We demonstrate state-of-the-art performance on two challenging datasets, LRS2 and LRS3, where our method achieves significant reductions in Word Error Rate (WER) achieving a SOTA WER of 18.7 on LRS3 despite using 99.4% less labelled data than the next best approach.
☆ ProHOC: Probabilistic Hierarchical Out-of-Distribution Classification via Multi-Depth Networks CVPR2025
Out-of-distribution (OOD) detection in deep learning has traditionally been framed as a binary task, where samples are either classified as belonging to the known classes or marked as OOD, with little attention given to the semantic relationships between OOD samples and the in-distribution (ID) classes. We propose a framework for detecting and classifying OOD samples in a given class hierarchy. Specifically, we aim to predict OOD data to their correct internal nodes of the class hierarchy, whereas the known ID classes should be predicted as their corresponding leaf nodes. Our approach leverages the class hierarchy to create a probabilistic model and we implement this model by using networks trained for ID classification at multiple hierarchy depths. We conduct experiments on three datasets with predefined class hierarchies and show the effectiveness of our method. Our code is available at https://github.com/walline/prohoc.
comment: CVPR2025
☆ Unsupervised Real-World Denoising: Sparsity is All You Need
Supervised training for real-world denoising presents challenges due to the difficulty of collecting large datasets of paired noisy and clean images. Recent methods have attempted to address this by utilizing unpaired datasets of clean and noisy images. Some approaches leverage such unpaired data to train denoisers in a supervised manner by generating synthetic clean-noisy pairs. However, these methods often fall short due to the distribution gap between synthetic and real noisy images. To mitigate this issue, we propose a solution based on input sparsification, specifically using random input masking. Our method, which we refer to as Mask, Inpaint and Denoise (MID), trains a denoiser to simultaneously denoise and inpaint synthetic clean-noisy pairs. On one hand, input sparsification reduces the gap between synthetic and real noisy images. On the other hand, an inpainter trained in a supervised manner can still accurately reconstruct sparse inputs by predicting missing clean pixels using the remaining unmasked pixels. Our approach begins with a synthetic Gaussian noise sampler and iteratively refines it using a noise dataset derived from the denoiser's predictions. The noise dataset is created by subtracting predicted pseudo-clean images from real noisy images at each iteration. The core intuition is that improving the denoiser results in a more accurate noise dataset and, consequently, a better noise sampler. We validate our method through extensive experiments on real-world noisy image datasets, demonstrating competitive performance compared to existing unsupervised denoising methods.
☆ Multimodal surface defect detection from wooden logs for sawing optimization
We propose a novel, good-quality, and less demanding method for detecting knots on the surface of wooden logs using multimodal data fusion. Knots are a primary factor affecting the quality of sawn timber, making their detection fundamental to any timber grading or cutting optimization system. While X-ray computed tomography provides accurate knot locations and internal structures, it is often too slow or expensive for practical use. An attractive alternative is to use fast and cost-effective log surface measurements, such as laser scanners or RGB cameras, to detect surface knots and estimate the internal structure of wood. However, due to the small size of knots and noise caused by factors, such as bark and other natural variations, detection accuracy often remains low when only one measurement modality is used. In this paper, we demonstrate that by using a data fusion pipeline consisting of separate streams for RGB and point cloud data, combined by a late fusion module, higher knot detection accuracy can be achieved compared to using either modality alone. We further propose a simple yet efficient sawing angle optimization method that utilizes surface knot detections and cross-correlation to minimize the amount of unwanted arris knots, demonstrating its benefits over randomized sawing angles.
☆ LandMarkSystem Technical Report
3D reconstruction is vital for applications in autonomous driving, virtual reality, augmented reality, and the metaverse. Recent advancements such as Neural Radiance Fields(NeRF) and 3D Gaussian Splatting (3DGS) have transformed the field, yet traditional deep learning frameworks struggle to meet the increasing demands for scene quality and scale. This paper introduces LandMarkSystem, a novel computing framework designed to enhance multi-scale scene reconstruction and rendering. By leveraging a componentized model adaptation layer, LandMarkSystem supports various NeRF and 3DGS structures while optimizing computational efficiency through distributed parallel computing and model parameter offloading. Our system addresses the limitations of existing frameworks, providing dedicated operators for complex 3D sparse computations, thus facilitating efficient training and rapid inference over extensive scenes. Key contributions include a modular architecture, a dynamic loading strategy for limited resources, and proven capabilities across multiple representative algorithms.This comprehensive solution aims to advance the efficiency and effectiveness of 3D reconstruction tasks.To facilitate further research and collaboration, the source code and documentation for the LandMarkSystem project are publicly available in an open-source repository, accessing the repository at: https://github.com/InternLandMark/LandMarkSystem.
☆ UGNA-VPR: A Novel Training Paradigm for Visual Place Recognition Based on Uncertainty-Guided NeRF Augmentation
Visual place recognition (VPR) is crucial for robots to identify previously visited locations, playing an important role in autonomous navigation in both indoor and outdoor environments. However, most existing VPR datasets are limited to single-viewpoint scenarios, leading to reduced recognition accuracy, particularly in multi-directional driving or feature-sparse scenes. Moreover, obtaining additional data to mitigate these limitations is often expensive. This paper introduces a novel training paradigm to improve the performance of existing VPR networks by enhancing multi-view diversity within current datasets through uncertainty estimation and NeRF-based data augmentation. Specifically, we initially train NeRF using the existing VPR dataset. Then, our devised self-supervised uncertainty estimation network identifies places with high uncertainty. The poses of these uncertain places are input into NeRF to generate new synthetic observations for further training of VPR networks. Additionally, we propose an improved storage method for efficient organization of augmented and original training data. We conducted extensive experiments on three datasets and tested three different VPR backbone networks. The results demonstrate that our proposed training paradigm significantly improves VPR performance by fully utilizing existing data, outperforming other training approaches. We further validated the effectiveness of our approach on self-recorded indoor and outdoor datasets, consistently demonstrating superior results. Our dataset and code have been released at \href{https://github.com/nubot-nudt/UGNA-VPR}{https://github.com/nubot-nudt/UGNA-VPR}.
comment: Accepted to IEEE Robotics and Automation Letters (RA-L)
☆ DuckSegmentation: A segmentation model based on the AnYue Hemp Duck Dataset
The modernization of smart farming is a way to improve agricultural production efficiency, and improve the agricultural production environment. Although many large models have achieved high accuracy in the task of object recognition and segmentation, they cannot really be put into use in the farming industry due to their own poor interpretability and limitations in computational volume. In this paper, we built AnYue Shelduck Dateset, which contains a total of 1951 Shelduck datasets, and performed target detection and segmentation annotation with the help of professional annotators. Based on AnYue ShelduckDateset, this paper describes DuckProcessing, an efficient and powerful module for duck identification based on real shelduckfarms. First of all, using the YOLOv8 module designed to divide the mahjong between them, Precision reached 98.10%, Recall reached 96.53% and F1 score reached 0.95 on the test set. Again using the DuckSegmentation segmentation model, DuckSegmentation reached 96.43% mIoU. Finally, the excellent DuckSegmentation was used as the teacher model, and through knowledge distillation, Deeplabv3 r50 was used as the student model, and the final student model achieved 94.49% mIoU on the test set. The method provides a new way of thinking in practical sisal duck smart farming.
☆ HORT: Monocular Hand-held Objects Reconstruction with Transformers
Reconstructing hand-held objects in 3D from monocular images remains a significant challenge in computer vision. Most existing approaches rely on implicit 3D representations, which produce overly smooth reconstructions and are time-consuming to generate explicit 3D shapes. While more recent methods directly reconstruct point clouds with diffusion models, the multi-step denoising makes high-resolution reconstruction inefficient. To address these limitations, we propose a transformer-based model to efficiently reconstruct dense 3D point clouds of hand-held objects. Our method follows a coarse-to-fine strategy, first generating a sparse point cloud from the image and progressively refining it into a dense representation using pixel-aligned image features. To enhance reconstruction accuracy, we integrate image features with 3D hand geometry to jointly predict the object point cloud and its pose relative to the hand. Our model is trained end-to-end for optimal performance. Experimental results on both synthetic and real datasets demonstrate that our method achieves state-of-the-art accuracy with much faster inference speed, while generalizing well to in-the-wild images.
comment: Project Page: https://zerchen.github.io/projects/hort.html
☆ FineCIR: Explicit Parsing of Fine-Grained Modification Semantics for Composed Image Retrieval
Composed Image Retrieval (CIR) facilitates image retrieval through a multimodal query consisting of a reference image and modification text. The reference image defines the retrieval context, while the modification text specifies desired alterations. However, existing CIR datasets predominantly employ coarse-grained modification text (CoarseMT), which inadequately captures fine-grained retrieval intents. This limitation introduces two key challenges: (1) ignoring detailed differences leads to imprecise positive samples, and (2) greater ambiguity arises when retrieving visually similar images. These issues degrade retrieval accuracy, necessitating manual result filtering or repeated queries. To address these limitations, we develop a robust fine-grained CIR data annotation pipeline that minimizes imprecise positive samples and enhances CIR systems' ability to discern modification intents accurately. Using this pipeline, we refine the FashionIQ and CIRR datasets to create two fine-grained CIR datasets: Fine-FashionIQ and Fine-CIRR. Furthermore, we introduce FineCIR, the first CIR framework explicitly designed to parse the modification text. FineCIR effectively captures fine-grained modification semantics and aligns them with ambiguous visual entities, enhancing retrieval precision. Extensive experiments demonstrate that FineCIR consistently outperforms state-of-the-art CIR baselines on both fine-grained and traditional CIR benchmark datasets. Our FineCIR code and fine-grained CIR datasets are available at https://github.com/SDU-L/FineCIR.git.
☆ InternVL-X: Advancing and Accelerating InternVL Series with Efficient Visual Token Compression
Most multimodal large language models (MLLMs) treat visual tokens as "a sequence of text", integrating them with text tokens into a large language model (LLM). However, a great quantity of visual tokens significantly increases the demand for computational resources and time. In this paper, we propose InternVL-X, which outperforms the InternVL model in both performance and efficiency by incorporating three visual token compression methods. First, we propose a novel vision-language projector, PVTC. This component integrates adjacent visual embeddings to form a local query and utilizes the transformed CLS token as a global query, then performs point-to-region cross-attention through these local and global queries to more effectively convert visual features. Second, we present a layer-wise visual token compression module, LVTC, which compresses tokens in the LLM shallow layers and then expands them through upsampling and residual connections in the deeper layers. This significantly enhances the model computational efficiency. Futhermore, we propose an efficient high resolution slicing method, RVTC, which dynamically adjusts the number of visual tokens based on image area or length filtering. RVTC greatly enhances training efficiency with only a slight reduction in performance. By utilizing 20% or fewer visual tokens, InternVL-X achieves state-of-the-art performance on 7 public MLLM benchmarks, and improves the average metric by 2.34% across 12 tasks.
☆ Multi-Scale Invertible Neural Network for Wide-Range Variable-Rate Learned Image Compression
Autoencoder-based structures have dominated recent learned image compression methods. However, the inherent information loss associated with autoencoders limits their rate-distortion performance at high bit rates and restricts their flexibility of rate adaptation. In this paper, we present a variable-rate image compression model based on invertible transform to overcome these limitations. Specifically, we design a lightweight multi-scale invertible neural network, which bijectively maps the input image into multi-scale latent representations. To improve the compression efficiency, a multi-scale spatial-channel context model with extended gain units is devised to estimate the entropy of the latent representation from high to low levels. Experimental results demonstrate that the proposed method achieves state-of-the-art performance compared to existing variable-rate methods, and remains competitive with recent multi-model approaches. Notably, our method is the first learned image compression solution that outperforms VVC across a very wide range of bit rates using a single model, especially at high bit rates.The source code is available at \href{https://github.com/hytu99/MSINN-VRLIC}{https://github.com/hytu99/MSINN-VRLIC}.
comment: Accepted to IEEE Transactions on Multimedia 2025
☆ Zero-Shot Visual Concept Blending Without Text Guidance
We propose a novel, zero-shot image generation technique called "Visual Concept Blending" that provides fine-grained control over which features from multiple reference images are transferred to a source image. If only a single reference image is available, it is difficult to isolate which specific elements should be transferred. However, using multiple reference images, the proposed approach distinguishes between common and unique features by selectively incorporating them into a generated output. By operating within a partially disentangled Contrastive Language-Image Pre-training (CLIP) embedding space (from IP-Adapter), our method enables the flexible transfer of texture, shape, motion, style, and more abstract conceptual transformations without requiring additional training or text prompts. We demonstrate its effectiveness across a diverse range of tasks, including style transfer, form metamorphosis, and conceptual transformations, showing how subtle or abstract attributes (e.g., brushstroke style, aerodynamic lines, and dynamism) can be seamlessly combined into a new image. In a user study, participants accurately recognized which features were intended to be transferred. Its simplicity, flexibility, and high-level control make Visual Concept Blending valuable for creative fields such as art, design, and content creation, where combining specific visual qualities from multiple inspirations is crucial.
☆ Delving Deep into Semantic Relation Distillation
Knowledge distillation has become a cornerstone technique in deep learning, facilitating the transfer of knowledge from complex models to lightweight counterparts. Traditional distillation approaches focus on transferring knowledge at the instance level, but fail to capture nuanced semantic relationships within the data. In response, this paper introduces a novel methodology, Semantics-based Relation Knowledge Distillation (SeRKD), which reimagines knowledge distillation through a semantics-relation lens among each sample. By leveraging semantic components, \ie, superpixels, SeRKD enables a more comprehensive and context-aware transfer of knowledge, which skillfully integrates superpixel-based semantic extraction with relation-based knowledge distillation for a sophisticated model compression and distillation. Particularly, the proposed method is naturally relevant in the domain of Vision Transformers (ViTs), where visual tokens serve as fundamental units of representation. Experimental evaluations on benchmark datasets demonstrate the superiority of SeRKD over existing methods, underscoring its efficacy in enhancing model performance and generalization capabilities.
☆ ClimbingCap: Multi-Modal Dataset and Method for Rock Climbing in World Coordinate CVPR2025
Human Motion Recovery (HMR) research mainly focuses on ground-based motions such as running. The study on capturing climbing motion, an off-ground motion, is sparse. This is partly due to the limited availability of climbing motion datasets, especially large-scale and challenging 3D labeled datasets. To address the insufficiency of climbing motion datasets, we collect AscendMotion, a large-scale well-annotated, and challenging climbing motion dataset. It consists of 412k RGB, LiDAR frames, and IMU measurements, including the challenging climbing motions of 22 skilled climbing coaches across 12 different rock walls. Capturing the climbing motions is challenging as it requires precise recovery of not only the complex pose but also the global position of climbers. Although multiple global HMR methods have been proposed, they cannot faithfully capture climbing motions. To address the limitations of HMR methods for climbing, we propose ClimbingCap, a motion recovery method that reconstructs continuous 3D human climbing motion in a global coordinate system. One key insight is to use the RGB and LiDAR modalities to separately reconstruct motions in camera coordinates and global coordinates and to optimize them jointly. We demonstrate the quality of the AscendMotion dataset and present promising results from ClimbingCap. The AscendMotion dataset and source code release publicly at \href{this link}{http://www.lidarhumanmotion.net/climbingcap/}
comment: CVPR2025, project in \href{this link}{http://www.lidarhumanmotion.net/climbingcap/}
☆ vGamba: Attentive State Space Bottleneck for efficient Long-range Dependencies in Visual Recognition
Capturing long-range dependencies efficiently is essential for visual recognition tasks, yet existing methods face limitations. Convolutional neural networks (CNNs) struggle with restricted receptive fields, while Vision Transformers (ViTs) achieve global context and long-range modeling at a high computational cost. State-space models (SSMs) offer an alternative, but their application in vision remains underexplored. This work introduces vGamba, a hybrid vision backbone that integrates SSMs with attention mechanisms to enhance efficiency and expressiveness. At its core, the Gamba bottleneck block that includes, Gamba Cell, an adaptation of Mamba for 2D spatial structures, alongside a Multi-Head Self-Attention (MHSA) mechanism and a Gated Fusion Module for effective feature representation. The interplay of these components ensures that vGamba leverages the low computational demands of SSMs while maintaining the accuracy of attention mechanisms for modeling long-range dependencies in vision tasks. Additionally, the Fusion module enables seamless interaction between these components. Extensive experiments on classification, detection, and segmentation tasks demonstrate that vGamba achieves a superior trade-off between accuracy and computational efficiency, outperforming several existing models.
☆ Reducing CT Metal Artifacts by Learning Latent Space Alignment with Gemstone Spectral Imaging Data
Metal artifacts in CT slices have long posed challenges in medical diagnostics. These artifacts degrade image quality, resulting in suboptimal visualization and complicating the accurate interpretation of tissues adjacent to metal implants. To address these issues, we introduce the Latent Gemstone Spectral Imaging (GSI) Alignment Framework, which effectively reduces metal artifacts while avoiding the introduction of noise information. Our work is based on a key finding that even artifact-affected ordinary CT sequences contain sufficient information to discern detailed structures. The challenge lies in the inability to clearly represent this information. To address this issue, we developed an Alignment Framework that adjusts the representation of ordinary CT images to match GSI CT sequences. GSI is an advanced imaging technique using multiple energy levels to mitigate artifacts caused by metal implants. By aligning the representation to GSI data, we can effectively suppress metal artifacts while clearly revealing detailed structure, without introducing extraneous information into CT sequences. To facilitate the application, we propose a new dataset, Artifacts-GSI, captured from real patients with metal implants, and establish a new benchmark based on this dataset. Experimental results show that our method significantly reduces metal artifacts and greatly enhances the readability of CT slices. All our code and data are available at: https://um-lab.github.io/GSI-MAR/
☆ Learn by Reasoning: Analogical Weight Generation for Few-Shot Class-Incremental Learning
Few-shot class-incremental Learning (FSCIL) enables models to learn new classes from limited data while retaining performance on previously learned classes. Traditional FSCIL methods often require fine-tuning parameters with limited new class data and suffer from a separation between learning new classes and utilizing old knowledge. Inspired by the analogical learning mechanisms of the human brain, we propose a novel analogical generative method. Our approach includes the Brain-Inspired Analogical Generator (BiAG), which derives new class weights from existing classes without parameter fine-tuning during incremental stages. BiAG consists of three components: Weight Self-Attention Module (WSA), Weight & Prototype Analogical Attention Module (WPAA), and Semantic Conversion Module (SCM). SCM uses Neural Collapse theory for semantic conversion, WSA supplements new class weights, and WPAA computes analogies to generate new class weights. Experiments on miniImageNet, CUB-200, and CIFAR-100 datasets demonstrate that our method achieves higher final and average accuracy compared to SOTA methods.
☆ Vision-to-Music Generation: A Survey
Vision-to-music Generation, including video-to-music and image-to-music tasks, is a significant branch of multimodal artificial intelligence demonstrating vast application prospects in fields such as film scoring, short video creation, and dance music synthesis. However, compared to the rapid development of modalities like text and images, research in vision-to-music is still in its preliminary stage due to its complex internal structure and the difficulty of modeling dynamic relationships with video. Existing surveys focus on general music generation without comprehensive discussion on vision-to-music. In this paper, we systematically review the research progress in the field of vision-to-music generation. We first analyze the technical characteristics and core challenges for three input types: general videos, human movement videos, and images, as well as two output types of symbolic music and audio music. We then summarize the existing methodologies on vision-to-music generation from the architecture perspective. A detailed review of common datasets and evaluation metrics is provided. Finally, we discuss current challenges and promising directions for future research. We hope our survey can inspire further innovation in vision-to-music generation and the broader field of multimodal generation in academic research and industrial applications. To follow latest works and foster further innovation in this field, we are continuously maintaining a GitHub repository at https://github.com/wzk1015/Awesome-Vision-to-Music-Generation.
☆ Orange Quality Grading with Deep Learning
Orange grading is a crucial step in the fruit industry, as it helps to sort oranges according to different criteria such as size, quality, ripeness, and health condition, ensuring safety for human consumption and better price allocation and client satisfaction. Automated grading enables faster processing, precision, and reduced human labor. In this paper, we implement a deep learning-based solution for orange grading via machine vision. Unlike typical grading systems that analyze fruits from a single view, we capture multiview images of each single orange in order to enable a richer representation. Afterwards, we compose the acquired images into one collage. This enables the analysis of the whole orange skin. We train a convolutional neural network (CNN) on the composed images to grade the oranges into three classes, namely good, bad, and undefined. We also evaluate the performance with two different CNNs (ResNet-18 and SqueezeNet). We show experimentally that multi-view grading is superior to single view grading.
☆ DynamiCtrl: Rethinking the Basic Structure and the Role of Text for High-quality Human Image Animation
Human image animation has recently gained significant attention due to advancements in generative models. However, existing methods still face two major challenges: (1) architectural limitations, most models rely on U-Net, which underperforms compared to the MM-DiT; and (2) the neglect of textual information, which can enhance controllability. In this work, we introduce DynamiCtrl, a novel framework that not only explores different pose-guided control structures in MM-DiT, but also reemphasizes the crucial role of text in this task. Specifically, we employ a Shared VAE encoder for both reference images and driving pose videos, eliminating the need for an additional pose encoder and simplifying the overall framework. To incorporate pose features into the full attention blocks, we propose Pose-adaptive Layer Norm (PadaLN), which utilizes adaptive layer normalization to encode sparse pose features. The encoded features are directly added to the visual input, preserving the spatiotemporal consistency of the backbone while effectively introducing pose control into MM-DiT. Furthermore, within the full attention mechanism, we align textual and visual features to enhance controllability. By leveraging text, we not only enable fine-grained control over the generated content, but also, for the first time, achieve simultaneous control over both background and motion. Experimental results verify the superiority of DynamiCtrl on benchmark datasets, demonstrating its strong identity preservation, heterogeneous character driving, background controllability, and high-quality synthesis. The project page is available at https://gulucaptain.github.io/DynamiCtrl/.
comment: 11 pages, 10 figures
☆ PLAIN: Scalable Estimation Architecture for Integrated Sensing and Communication
Integrated sensing and communication (ISAC) is envisioned be to one of the paradigms upon which next-generation mobile networks will be built, extending localization and tracking capabilities, as well as giving birth to environment-aware wireless access. A key aspect of sensing integration is parameter estimation, which involves extracting information about the surrounding environment, such as the direction, distance, and velocity of various objects within. This is typically of a high-dimensional nature, which leads to significant computational complexity, if performed jointly across multiple sensing dimensions, such as space, frequency, and time. Additionally, due to the incorporation of sensing on top of the data transmission, the time window available for sensing is likely to be short, resulting in an estimation problem where only a single snapshot is accessible. In this work, we propose PLAIN, a tensor-based estimation architecture that flexibly scales with multiple sensing dimensions and can handle high dimensionality, limited measurement time, and super-resolution requirements. It consists of three stages: a compression stage, where the high dimensional input is converted into lower dimensionality, without sacrificing resolution; a decoupled estimation stage, where the parameters across the different dimensions are estimated in parallel with low complexity; an input-based fusion stage, where the decoupled parameters are fused together to form a paired multidimensional estimate. We investigate the performance of the architecture for different configurations and compare it against practical sequential and joint estimation baselines, as well as theoretical bounds. Our results show that PLAIN, using tools from tensor algebra, subspace-based processing, and compressed sensing, can scale flexibly with dimensionality, while operating with low complexity and maintaining super-resolution.
comment: Submitted to the IEEE Transactions on Wireless Communications. Code available at GitHub: https://github.com/bashar-tahir/plain
☆ Clean Image May be Dangerous: Data Poisoning Attacks Against Deep Hashing
Large-scale image retrieval using deep hashing has become increasingly popular due to the exponential growth of image data and the remarkable feature extraction capabilities of deep neural networks (DNNs). However, deep hashing methods are vulnerable to malicious attacks, including adversarial and backdoor attacks. It is worth noting that these attacks typically involve altering the query images, which is not a practical concern in real-world scenarios. In this paper, we point out that even clean query images can be dangerous, inducing malicious target retrieval results, like undesired or illegal images. To the best of our knowledge, we are the first to study data \textbf{p}oisoning \textbf{a}ttacks against \textbf{d}eep \textbf{hash}ing \textbf{(\textit{PADHASH})}. Specifically, we first train a surrogate model to simulate the behavior of the target deep hashing model. Then, a strict gradient matching strategy is proposed to generate the poisoned images. Extensive experiments on different models, datasets, hash methods, and hash code lengths demonstrate the effectiveness and generality of our attack method.
comment: Accepted by TMM
☆ Frequency-Aware Gaussian Splatting Decomposition
3D Gaussian Splatting (3D-GS) has revolutionized novel view synthesis with its efficient, explicit representation. However, it lacks frequency interpretability, making it difficult to separate low-frequency structures from fine details. We introduce a frequency-decomposed 3D-GS framework that groups 3D Gaussians that correspond to subbands in the Laplacian Pyrmaids of the input images. Our approach enforces coherence within each subband (i.e., group of 3D Gaussians) through dedicated regularization, ensuring well-separated frequency components. We extend color values to both positive and negative ranges, allowing higher-frequency layers to add or subtract residual details. To stabilize optimization, we employ a progressive training scheme that refines details in a coarse-to-fine manner. Beyond interpretability, this frequency-aware design unlocks a range of practical benefits. Explicit frequency separation enables advanced 3D editing and stylization, allowing precise manipulation of specific frequency bands. It also supports dynamic level-of-detail control for progressive rendering, streaming, foveated rendering and fast geometry interaction. Through extensive experiments, we demonstrate that our method provides improved control and flexibility for emerging applications in scene editing and interactive rendering. Our code will be made publicly available.
☆ GenFusion: Closing the Loop between Reconstruction and Generation via Videos
Recently, 3D reconstruction and generation have demonstrated impressive novel view synthesis results, achieving high fidelity and efficiency. However, a notable conditioning gap can be observed between these two fields, e.g., scalable 3D scene reconstruction often requires densely captured views, whereas 3D generation typically relies on a single or no input view, which significantly limits their applications. We found that the source of this phenomenon lies in the misalignment between 3D constraints and generative priors. To address this problem, we propose a reconstruction-driven video diffusion model that learns to condition video frames on artifact-prone RGB-D renderings. Moreover, we propose a cyclical fusion pipeline that iteratively adds restoration frames from the generative model to the training set, enabling progressive expansion and addressing the viewpoint saturation limitations seen in previous reconstruction and generation pipelines. Our evaluation, including view synthesis from sparse view and masked input, validates the effectiveness of our approach.
☆ VoxRep: Enhancing 3D Spatial Understanding in 2D Vision-Language Models via Voxel Representation
Comprehending 3D environments is vital for intelligent systems in domains like robotics and autonomous navigation. Voxel grids offer a structured representation of 3D space, but extracting high-level semantic meaning remains challenging. This paper proposes a novel approach utilizing a Vision-Language Model (VLM) to extract "voxel semantics"-object identity, color, and location-from voxel data. Critically, instead of employing complex 3D networks, our method processes the voxel space by systematically slicing it along a primary axis (e.g., the Z-axis, analogous to CT scan slices). These 2D slices are then formatted and sequentially fed into the image encoder of a standard VLM. The model learns to aggregate information across slices and correlate spatial patterns with semantic concepts provided by the language component. This slice-based strategy aims to leverage the power of pre-trained 2D VLMs for efficient 3D semantic understanding directly from voxel representations.
☆ FakeReasoning: Towards Generalizable Forgery Detection and Reasoning
Accurate and interpretable detection of AI-generated images is essential for mitigating risks associated with AI misuse. However, the substantial domain gap among generative models makes it challenging to develop a generalizable forgery detection model. Moreover, since every pixel in an AI-generated image is synthesized, traditional saliency-based forgery explanation methods are not well suited for this task. To address these challenges, we propose modeling AI-generated image detection and explanation as a Forgery Detection and Reasoning task (FDR-Task), leveraging vision-language models (VLMs) to provide accurate detection through structured and reliable reasoning over forgery attributes. To facilitate this task, we introduce the Multi-Modal Forgery Reasoning dataset (MMFR-Dataset), a large-scale dataset containing 100K images across 10 generative models, with 10 types of forgery reasoning annotations, enabling comprehensive evaluation of FDR-Task. Additionally, we propose FakeReasoning, a forgery detection and reasoning framework with two key components. First, Forgery-Aligned Contrastive Learning enhances VLMs' understanding of forgery-related semantics through both cross-modal and intra-modal contrastive learning between images and forgery attribute reasoning. Second, a Classification Probability Mapper bridges the optimization gap between forgery detection and language modeling by mapping the output logits of VLMs to calibrated binary classification probabilities. Experiments across multiple generative models demonstrate that FakeReasoning not only achieves robust generalization but also outperforms state-of-the-art methods on both detection and reasoning tasks.
☆ An improved EfficientNetV2 for garbage classification
This paper presents an enhanced waste classification framework based on EfficientNetV2 to address challenges in data acquisition cost, generalization, and real-time performance. We propose a Channel-Efficient Attention (CE-Attention) module that mitigates feature loss during global pooling without introducing dimensional scaling, effectively enhancing critical feature extraction. Additionally, a lightweight multi-scale spatial feature extraction module (SAFM) is developed by integrating depthwise separable convolutions, significantly reducing model complexity. Comprehensive data augmentation strategies are further employed to improve generalization. Experiments on the Huawei Cloud waste classification dataset demonstrate that our method achieves a classification accuracy of 95.4\%, surpassing the baseline by 3.2\% and outperforming mainstream models. The results validate the effectiveness of our approach in balancing accuracy and efficiency for practical waste classification scenarios.
☆ WVSC: Wireless Video Semantic Communication with Multi-frame Compensation
Existing wireless video transmission schemes directly conduct video coding in pixel level, while neglecting the inner semantics contained in videos. In this paper, we propose a wireless video semantic communication framework, abbreviated as WVSC, which integrates the idea of semantic communication into wireless video transmission scenarios. WVSC first encodes original video frames as semantic frames and then conducts video coding based on such compact representations, enabling the video coding in semantic level rather than pixel level. Moreover, to further reduce the communication overhead, a reference semantic frame is introduced to substitute motion vectors of each frame in common video coding methods. At the receiver, multi-frame compensation (MFC) is proposed to produce compensated current semantic frame with a multi-frame fusion attention module. With both the reference frame transmission and MFC, the bandwidth efficiency improves with satisfying video transmission performance. Experimental results verify the performance gain of WVSC over other DL-based methods e.g. DVSC about 1 dB and traditional schemes about 2 dB in terms of PSNR.
☆ UGen: Unified Autoregressive Multimodal Model with Progressive Vocabulary Learning
We introduce UGen, a unified autoregressive multimodal model that demonstrates strong performance across text processing, image understanding, and image generation tasks simultaneously. UGen converts both texts and images into discrete token sequences and utilizes a single transformer to generate them uniformly in an autoregressive manner. To address the challenges associated with unified multimodal learning, UGen is trained using a novel mechanism, namely progressive vocabulary learning. In this process, visual token IDs are incrementally activated and integrated into the training phase, ultimately enhancing the effectiveness of unified multimodal learning. Experiments on comprehensive text and image tasks show that UGen achieves a significant overall performance improvement of 13.3% compared to the vanilla unified autoregressive method, and it also delivers competitive results across all tasks against several task-specific models.
☆ Leveraging LLMs with Iterative Loop Structure for Enhanced Social Intelligence in Video Question Answering
Social intelligence, the ability to interpret emotions, intentions, and behaviors, is essential for effective communication and adaptive responses. As robots and AI systems become more prevalent in caregiving, healthcare, and education, the demand for AI that can interact naturally with humans grows. However, creating AI that seamlessly integrates multiple modalities, such as vision and speech, remains a challenge. Current video-based methods for social intelligence rely on general video recognition or emotion recognition techniques, often overlook the unique elements inherent in human interactions. To address this, we propose the Looped Video Debating (LVD) framework, which integrates Large Language Models (LLMs) with visual information, such as facial expressions and body movements, to enhance the transparency and reliability of question-answering tasks involving human interaction videos. Our results on the Social-IQ 2.0 benchmark show that LVD achieves state-of-the-art performance without fine-tuning. Furthermore, supplementary human annotations on existing datasets provide insights into the model's accuracy, guiding future improvements in AI-driven social intelligence.
☆ DGSUnet: An Improved Unet Model with DINO-Guided SAM2 for Multi-Scale Feature Collaboration
Despite the significant advancements in general image segmentation achieved by large-scale pre-trained foundation models (such as Meta's Segment Any-thing Model (SAM) series and DINOv2), their performance in specialized fields remains limited by two critical issues: the excessive training costs due to large model parameters, and the insufficient ability to represent specific domain characteristics. This paper proposes a multi-scale feature collabora-tion framework guided by DINOv2 for SAM2, with core innovations in three aspects: (1) Establishing a feature collaboration mechanism between DINOv2 and SAM2 backbones, where high-dimensional semantic features extracted by the self-supervised model guide multi-scale feature fusion; (2) Designing lightweight adapter modules and cross-modal, cross-layer feature fusion units to inject cross-domain knowledge while freezing the base model parameters; (3) Constructing a U-shaped network structure based on U-net, which utilizes attention mechanisms to achieve adaptive aggregation decoding of multi-granularity features. This framework surpasses existing state-of-the-art meth-ods in downstream tasks such as camouflage target detection and salient ob-ject detection, without requiring costly training processes. It provides a tech-nical pathway for efficient deployment of visual image segmentation, demon-strating significant application value in a wide range of downstream tasks and specialized fields within image segmentation.Project page: https://github.com/CheneyXuYiMin/SAM2DINO-Seg
☆ Model as a Game: On Numerical and Spatial Consistency for Generative Games
Recent advances in generative models have significantly impacted game generation. However, despite producing high-quality graphics and adequately receiving player input, existing models often fail to maintain fundamental game properties such as numerical and spatial consistency. Numerical consistency ensures gameplay mechanics correctly reflect score changes and other quantitative elements, while spatial consistency prevents jarring scene transitions, providing seamless player experiences. In this paper, we revisit the paradigm of generative games to explore what truly constitutes a Model as a Game (MaaG) with a well-developed mechanism. We begin with an empirical study on ``Traveler'', a 2D game created by an LLM featuring minimalist rules yet challenging generative models in maintaining consistency. Based on the DiT architecture, we design two specialized modules: (1) a numerical module that integrates a LogicNet to determine event triggers, with calculations processed externally as conditions for image generation; and (2) a spatial module that maintains a map of explored areas, retrieving location-specific information during generation and linking new observations to ensure continuity. Experiments across three games demonstrate that our integrated modules significantly enhance performance on consistency metrics compared to baselines, while incurring minimal time overhead during inference.
comment: Technical Report
☆ VADMamba: Exploring State Space Models for Fast Video Anomaly Detection ICME 2025
Video anomaly detection (VAD) methods are mostly CNN-based or Transformer-based, achieving impressive results, but the focus on detection accuracy often comes at the expense of inference speed. The emergence of state space models in computer vision, exemplified by the Mamba model, demonstrates improved computational efficiency through selective scans and showcases the great potential for long-range modeling. Our study pioneers the application of Mamba to VAD, dubbed VADMamba, which is based on multi-task learning for frame prediction and optical flow reconstruction. Specifically, we propose the VQ-Mamba Unet (VQ-MaU) framework, which incorporates a Vector Quantization (VQ) layer and Mamba-based Non-negative Visual State Space (NVSS) block. Furthermore, two individual VQ-MaU networks separately predict frames and reconstruct corresponding optical flows, further boosting accuracy through a clip-level fusion evaluation strategy. Experimental results validate the efficacy of the proposed VADMamba across three benchmark datasets, demonstrating superior performance in inference speed compared to previous work. Code is available at https://github.com/jLooo/VADMamba.
comment: Accpeted by ICME 2025
☆ Adversarial Wear and Tear: Exploiting Natural Damage for Generating Physical-World Adversarial Examples
The presence of adversarial examples in the physical world poses significant challenges to the deployment of Deep Neural Networks in safety-critical applications such as autonomous driving. Most existing methods for crafting physical-world adversarial examples are ad-hoc, relying on temporary modifications like shadows, laser beams, or stickers that are tailored to specific scenarios. In this paper, we introduce a new class of physical-world adversarial examples, AdvWT, which draws inspiration from the naturally occurring phenomenon of `wear and tear', an inherent property of physical objects. Unlike manually crafted perturbations, `wear and tear' emerges organically over time due to environmental degradation, as seen in the gradual deterioration of outdoor signboards. To achieve this, AdvWT follows a two-step approach. First, a GAN-based, unsupervised image-to-image translation network is employed to model these naturally occurring damages, particularly in the context of outdoor signboards. The translation network encodes the characteristics of damaged signs into a latent `damage style code'. In the second step, we introduce adversarial perturbations into the style code, strategically optimizing its transformation process. This manipulation subtly alters the damage style representation, guiding the network to generate adversarial images where the appearance of damages remains perceptually realistic, while simultaneously ensuring their effectiveness in misleading neural networks. Through comprehensive experiments on two traffic sign datasets, we show that AdvWT effectively misleads DNNs in both digital and physical domains. AdvWT achieves an effective attack success rate, greater robustness, and a more natural appearance compared to existing physical-world adversarial examples. Additionally, integrating AdvWT into training enhances a model's generalizability to real-world damaged signs.
comment: 11 pages, 9 figures
☆ Integrating Travel Behavior Forecasting and Generative Modeling for Predicting Future Urban Mobility and Spatial Transformations
Transportation planning plays a critical role in shaping urban development, economic mobility, and infrastructure sustainability. However, traditional planning methods often struggle to accurately predict long-term urban growth and transportation demands. This may sometimes result in infrastructure demolition to make room for current transportation planning demands. This study integrates a Temporal Fusion Transformer to predict travel patterns from demographic data with a Generative Adversarial Network to predict future urban settings through satellite imagery. The framework achieved a 0.76 R-square score in travel behavior prediction and generated high-fidelity satellite images with a Structural Similarity Index of 0.81. The results demonstrate that integrating predictive analytics and spatial visualization can significantly improve the decision-making process, fostering more sustainable and efficient urban development. This research highlights the importance of data-driven methodologies in modern transportation planning and presents a step toward optimizing infrastructure placement, capacity, and long-term viability.
☆ The Devil is in Low-Level Features for Cross-Domain Few-Shot Segmentation CVPR 2025
Cross-Domain Few-Shot Segmentation (CDFSS) is proposed to transfer the pixel-level segmentation capabilities learned from large-scale source-domain datasets to downstream target-domain datasets, with only a few annotated images per class. In this paper, we focus on a well-observed but unresolved phenomenon in CDFSS: for target domains, particularly those distant from the source domain, segmentation performance peaks at the very early epochs, and declines sharply as the source-domain training proceeds. We delve into this phenomenon for an interpretation: low-level features are vulnerable to domain shifts, leading to sharper loss landscapes during the source-domain training, which is the devil of CDFSS. Based on this phenomenon and interpretation, we further propose a method that includes two plug-and-play modules: one to flatten the loss landscapes for low-level features during source-domain training as a novel sharpness-aware minimization method, and the other to directly supplement target-domain information to the model during target-domain testing by low-level-based calibration. Extensive experiments on four target datasets validate our rationale and demonstrate that our method surpasses the state-of-the-art method in CDFSS signifcantly by 3.71% and 5.34% average MIoU in 1-shot and 5-shot scenarios, respectively.
comment: Accepted by CVPR 2025
☆ ChatAnyone: Stylized Real-time Portrait Video Generation with Hierarchical Motion Diffusion Model
Real-time interactive video-chat portraits have been increasingly recognized as the future trend, particularly due to the remarkable progress made in text and voice chat technologies. However, existing methods primarily focus on real-time generation of head movements, but struggle to produce synchronized body motions that match these head actions. Additionally, achieving fine-grained control over the speaking style and nuances of facial expressions remains a challenge. To address these limitations, we introduce a novel framework for stylized real-time portrait video generation, enabling expressive and flexible video chat that extends from talking head to upper-body interaction. Our approach consists of the following two stages. The first stage involves efficient hierarchical motion diffusion models, that take both explicit and implicit motion representations into account based on audio inputs, which can generate a diverse range of facial expressions with stylistic control and synchronization between head and body movements. The second stage aims to generate portrait video featuring upper-body movements, including hand gestures. We inject explicit hand control signals into the generator to produce more detailed hand movements, and further perform face refinement to enhance the overall realism and expressiveness of the portrait video. Additionally, our approach supports efficient and continuous generation of upper-body portrait video in maximum 512 * 768 resolution at up to 30fps on 4090 GPU, supporting interactive video-chat in real-time. Experimental results demonstrate the capability of our approach to produce portrait videos with rich expressiveness and natural upper-body movements.
comment: Project Page: https://humanaigc.github.io/chat-anyone/
☆ Recurrent Feature Mining and Keypoint Mixup Padding for Category-Agnostic Pose Estimation
Category-agnostic pose estimation aims to locate keypoints on query images according to a few annotated support images for arbitrary novel classes. Existing methods generally extract support features via heatmap pooling, and obtain interacted features from support and query via cross-attention. Hence, these works neglect to mine fine-grained and structure-aware (FGSA) features from both support and query images, which are crucial for pixel-level keypoint localization. To this end, we propose a novel yet concise framework, which recurrently mines FGSA features from both support and query images. Specifically, we design a FGSA mining module based on deformable attention mechanism. On the one hand, we mine fine-grained features by applying deformable attention head over multi-scale feature maps. On the other hand, we mine structure-aware features by offsetting the reference points of keypoints to their linked keypoints. By means of above module, we recurrently mine FGSA features from support and query images, and thus obtain better support features and query estimations. In addition, we propose to use mixup keypoints to pad various classes to a unified keypoint number, which could provide richer supervision than the zero padding used in existing works. We conduct extensive experiments and in-depth studies on large-scale MP-100 dataset, and outperform SOTA method dramatically (+3.2\%PCK@0.05). Code is avaiable at https://github.com/chenbys/FMMP.
☆ VideoMix: Aggregating How-To Videos for Task-Oriented Learning
Tutorial videos are a valuable resource for people looking to learn new tasks. People often learn these skills by viewing multiple tutorial videos to get an overall understanding of a task by looking at different approaches to achieve the task. However, navigating through multiple videos can be time-consuming and mentally demanding as these videos are scattered and not easy to skim. We propose VideoMix, a system that helps users gain a holistic understanding of a how-to task by aggregating information from multiple videos on the task. Insights from our formative study (N=12) reveal that learners value understanding potential outcomes, required materials, alternative methods, and important details shared by different videos. Powered by a Vision-Language Model pipeline, VideoMix extracts and organizes this information, presenting concise textual summaries alongside relevant video clips, enabling users to quickly digest and navigate the content. A comparative user study (N=12) demonstrated that VideoMix enabled participants to gain a more comprehensive understanding of tasks with greater efficiency than a baseline video interface, where videos are viewed independently. Our findings highlight the potential of a task-oriented, multi-video approach where videos are organized around a shared goal, offering an enhanced alternative to conventional video-based learning.
comment: In Proceedings of the 30th International Conference on Intelligent User Interfaces (IUI '25) 2025
☆ Omni-AD: Learning to Reconstruct Global and Local Features for Multi-class Anomaly Detection
In multi-class unsupervised anomaly detection(MUAD), reconstruction-based methods learn to map input images to normal patterns to identify anomalous pixels. However, this strategy easily falls into the well-known "learning shortcut" issue when decoders fail to capture normal patterns and reconstruct both normal and abnormal samples naively. To address that, we propose to learn the input features in global and local manners, forcing the network to memorize the normal patterns more comprehensively. Specifically, we design a two-branch decoder block, named Omni-block. One branch corresponds to global feature learning, where we serialize two self-attention blocks but replace the query and (key, value) with learnable tokens, respectively, thus capturing global features of normal patterns concisely and thoroughly. The local branch comprises depth-separable convolutions, whose locality enables effective and efficient learning of local features for normal patterns. By stacking Omni-blocks, we build a framework, Omni-AD, to learn normal patterns of different granularity and reconstruct them progressively. Comprehensive experiments on public anomaly detection benchmarks show that our method outperforms state-of-the-art approaches in MUAD. Code is available at https://github.com/easyoo/Omni-AD.git.
☆ AdaMHF: Adaptive Multimodal Hierarchical Fusion for Survival Prediction ICME 2025
The integration of pathologic images and genomic data for survival analysis has gained increasing attention with advances in multimodal learning. However, current methods often ignore biological characteristics, such as heterogeneity and sparsity, both within and across modalities, ultimately limiting their adaptability to clinical practice. To address these challenges, we propose AdaMHF: Adaptive Multimodal Hierarchical Fusion, a framework designed for efficient, comprehensive, and tailored feature extraction and fusion. AdaMHF is specifically adapted to the uniqueness of medical data, enabling accurate predictions with minimal resource consumption, even under challenging scenarios with missing modalities. Initially, AdaMHF employs an experts expansion and residual structure to activate specialized experts for extracting heterogeneous and sparse features. Extracted tokens undergo refinement via selection and aggregation, reducing the weight of non-dominant features while preserving comprehensive information. Subsequently, the encoded features are hierarchically fused, allowing multi-grained interactions across modalities to be captured. Furthermore, we introduce a survival prediction benchmark designed to resolve scenarios with missing modalities, mirroring real-world clinical conditions. Extensive experiments on TCGA datasets demonstrate that AdaMHF surpasses current state-of-the-art (SOTA) methods, showcasing exceptional performance in both complete and incomplete modality settings.
comment: Accepted by ICME 2025
☆ One Snapshot is All You Need: A Generalized Method for mmWave Signal Generation
Wireless sensing systems, particularly those using mmWave technology, offer distinct advantages over traditional vision-based approaches, such as enhanced privacy and effectiveness in poor lighting conditions. These systems, leveraging FMCW signals, have shown success in human-centric applications like localization, gesture recognition, and so on. However, comprehensive mmWave datasets for diverse applications are scarce, often constrained by pre-processed signatures (e.g., point clouds or RA heatmaps) and inconsistent annotation formats. To overcome these limitations, we propose mmGen, a novel and generalized framework tailored for full-scene mmWave signal generation. By constructing physical signal transmission models, mmGen synthesizes human-reflected and environment-reflected mmWave signals from the constructed 3D meshes. Additionally, we incorporate methods to account for material properties, antenna gains, and multipath reflections, enhancing the realism of the synthesized signals. We conduct extensive experiments using a prototype system with commercial mmWave devices and Kinect sensors. The results show that the average similarity of Range-Angle and micro-Doppler signatures between the synthesized and real-captured signals across three different environments exceeds 0.91 and 0.89, respectively, demonstrating the effectiveness and practical applicability of mmGen.
comment: IEEE INFOCOM 2025
☆ StyledStreets: Multi-style Street Simulator with Spatial and Temporal Consistency
Urban scene reconstruction requires modeling both static infrastructure and dynamic elements while supporting diverse environmental conditions. We present \textbf{StyledStreets}, a multi-style street simulator that achieves instruction-driven scene editing with guaranteed spatial and temporal consistency. Building on a state-of-the-art Gaussian Splatting framework for street scenarios enhanced by our proposed pose optimization and multi-view training, our method enables photorealistic style transfers across seasons, weather conditions, and camera setups through three key innovations: First, a hybrid embedding scheme disentangles persistent scene geometry from transient style attributes, allowing realistic environmental edits while preserving structural integrity. Second, uncertainty-aware rendering mitigates supervision noise from diffusion priors, enabling robust training across extreme style variations. Third, a unified parametric model prevents geometric drift through regularized updates, maintaining multi-view consistency across seven vehicle-mounted cameras. Our framework preserves the original scene's motion patterns and geometric relationships. Qualitative results demonstrate plausible transitions between diverse conditions (snow, sandstorm, night), while quantitative evaluations show state-of-the-art geometric accuracy under style transfers. The approach establishes new capabilities for urban simulation, with applications in autonomous vehicle testing and augmented reality systems requiring reliable environmental consistency. Codes will be publicly available upon publication.
comment: 14 pages
☆ Learning Class Prototypes for Unified Sparse Supervised 3D Object Detection CVPR 2025
Both indoor and outdoor scene perceptions are essential for embodied intelligence. However, current sparse supervised 3D object detection methods focus solely on outdoor scenes without considering indoor settings. To this end, we propose a unified sparse supervised 3D object detection method for both indoor and outdoor scenes through learning class prototypes to effectively utilize unlabeled objects. Specifically, we first propose a prototype-based object mining module that converts the unlabeled object mining into a matching problem between class prototypes and unlabeled features. By using optimal transport matching results, we assign prototype labels to high-confidence features, thereby achieving the mining of unlabeled objects. We then present a multi-label cooperative refinement module to effectively recover missed detections through pseudo label quality control and prototype label cooperation. Experiments show that our method achieves state-of-the-art performance under the one object per scene sparse supervised setting across indoor and outdoor datasets. With only one labeled object per scene, our method achieves about 78%, 90%, and 96% performance compared to the fully supervised detector on ScanNet V2, SUN RGB-D, and KITTI, respectively, highlighting the scalability of our method. Code is available at https://github.com/zyrant/CPDet3D.
comment: Accepted by CVPR 2025
☆ ZJUKLAB at SemEval-2025 Task 4: Unlearning via Model Merging
This paper presents the ZJUKLAB team's submission for SemEval-2025 Task 4: Unlearning Sensitive Content from Large Language Models. This task aims to selectively erase sensitive knowledge from large language models, avoiding both over-forgetting and under-forgetting issues. We propose an unlearning system that leverages Model Merging (specifically TIES-Merging), combining two specialized models into a more balanced unlearned model. Our system achieves competitive results, ranking second among 26 teams, with an online score of 0.944 for Task Aggregate and 0.487 for overall Aggregate. In this paper, we also conduct local experiments and perform a comprehensive analysis of the unlearning process, examining performance trajectories, loss dynamics, and weight perspectives, along with several supplementary experiments, to understand the effectiveness of our method. Furthermore, we analyze the shortcomings of our method and evaluation metrics, emphasizing that MIA scores and ROUGE-based metrics alone are insufficient to fully evaluate successful unlearning. Finally, we emphasize the need for more comprehensive evaluation methodologies and rethinking of unlearning objectives in future research. Code is available at https://github.com/zjunlp/unlearn/tree/main/semeval25.
comment: Work in progress
☆ Can Video Diffusion Model Reconstruct 4D Geometry?
Reconstructing dynamic 3D scenes (i.e., 4D geometry) from monocular video is an important yet challenging problem. Conventional multiview geometry-based approaches often struggle with dynamic motion, whereas recent learning-based methods either require specialized 4D representation or sophisticated optimization. In this paper, we present Sora3R, a novel framework that taps into the rich spatiotemporal priors of large-scale video diffusion models to directly infer 4D pointmaps from casual videos. Sora3R follows a two-stage pipeline: (1) we adapt a pointmap VAE from a pretrained video VAE, ensuring compatibility between the geometry and video latent spaces; (2) we finetune a diffusion backbone in combined video and pointmap latent space to generate coherent 4D pointmaps for every frame. Sora3R operates in a fully feedforward manner, requiring no external modules (e.g., depth, optical flow, or segmentation) or iterative global alignment. Extensive experiments demonstrate that Sora3R reliably recovers both camera poses and detailed scene geometry, achieving performance on par with state-of-the-art methods for dynamic 4D reconstruction across diverse scenarios.
☆ KAC: Kolmogorov-Arnold Classifier for Continual Learning CVPR 2025
Continual learning requires models to train continuously across consecutive tasks without forgetting. Most existing methods utilize linear classifiers, which struggle to maintain a stable classification space while learning new tasks. Inspired by the success of Kolmogorov-Arnold Networks (KAN) in preserving learning stability during simple continual regression tasks, we set out to explore their potential in more complex continual learning scenarios. In this paper, we introduce the Kolmogorov-Arnold Classifier (KAC), a novel classifier developed for continual learning based on the KAN structure. We delve into the impact of KAN's spline functions and introduce Radial Basis Functions (RBF) for improved compatibility with continual learning. We replace linear classifiers with KAC in several recent approaches and conduct experiments across various continual learning benchmarks, all of which demonstrate performance improvements, highlighting the effectiveness and robustness of KAC in continual learning. The code is available at https://github.com/Ethanhuhuhu/KAC.
comment: CVPR 2025
☆ Rerouting Connection: Hybrid Computer Vision Analysis Reveals Visual Similarity Between Indus and Tibetan-Yi Corridor Writing Systems
This thesis employs a hybrid CNN-Transformer architecture, in conjunction with a detailed anthropological framework, to investigate potential historical connections between the visual morphology of the Indus Valley script and pictographic systems of the Tibetan-Yi Corridor. Through an ensemble methodology of three target scripts across 15 independently trained models, we demonstrate that Tibetan-Yi Corridor scripts exhibit approximately six-fold higher visual similarity to the Indus script (61.7%-63.5%) than to the Bronze Age Proto-Cuneiform (10.2%-10.9%) or Proto-Elamite (7.6%-8.7%) systems. Additionally and contrarily to our current understanding of the networks of the Indus Valley Civilization, the Indus script unexpectedly maps closer to Tibetan-Yi Corridor scripts, with a mean cosine similarity of 0.629, than to the aforementioned contemporaneous West Asian signaries, both of which recorded mean cosine similarities of 0.104 and 0.080 despite their close geographic proximity and evident trade relations. Across various dimensionality reduction practices and clustering methodologies, the Indus script consistently clusters closest to Tibetan-Yi Corridor scripts. Our computational results align with qualitative observations of specific pictorial parallels in numeral systems, gender markers, and key iconographic elements; this is further supported by archaeological evidence of sustained contact networks along the ancient Shu-Shendu road in tandem with the Indus Valley Civilization's decline, providing a plausible transmission pathway. While alternative explanations cannot be ruled out, the specificity and consistency of observed similarities challenge conventional narratives of isolated script development and suggest more complex ancient cultural transmission networks between South and East Asia than previously recognized.
comment: 106 pages total (main text: 42, 48 w/refs, 100 w/appendices). 21 figures, 4 tables in main; 106 figs, 8 tables total. Code and data at this URL: https://github.com/oohalakkadi/ivc2tyc. Submitted as undergrad thesis at Duke Kunshan University; accepted for presentation at the 2025 Computer Applications and Quantitative Methods in Archaeology Conference, Athens
☆ HSLiNets: Evaluating Band Ordering Strategies in Hyperspectral and LiDAR Fusion
The integration of hyperspectral imaging (HSI) and Light Detection and Ranging (LiDAR) data provides complementary spectral and spatial information for remote sensing applications. While previous studies have explored the role of band selection and grouping in HSI classification, little attention has been given to how the spectral sequence or band order affects classification outcomes when fused with LiDAR. In this work, we systematically investigate the influence of band order on HSI-LiDAR fusion performance. Through extensive experiments, we demonstrate that band order significantly impacts classification accuracy, revealing a previously overlooked factor in fusion-based models. Motivated by this observation, we propose a novel fusion architecture that not only integrates HSI and LiDAR data but also learns from multiple band order configurations. The proposed method enhances feature representation by adaptively fusing different spectral sequences, leading to improved classification accuracy. Experimental results on the Houston 2013 and Trento datasets show that our approach outperforms state-of-the-art fusion models. Data and code are available at https://github.com/Judyxyang/HSLiNets.
comment: 2 figures, 5 pages
☆ Efficient Multi-Instance Generation with Janus-Pro-Dirven Prompt Parsing
Recent advances in text-guided diffusion models have revolutionized conditional image generation, yet they struggle to synthesize complex scenes with multiple objects due to imprecise spatial grounding and limited scalability. We address these challenges through two key modules: 1) Janus-Pro-driven Prompt Parsing, a prompt-layout parsing module that bridges text understanding and layout generation via a compact 1B-parameter architecture, and 2) MIGLoRA, a parameter-efficient plug-in integrating Low-Rank Adaptation (LoRA) into UNet (SD1.5) and DiT (SD3) backbones. MIGLoRA is capable of preserving the base model's parameters and ensuring plug-and-play adaptability, minimizing architectural intrusion while enabling efficient fine-tuning. To support a comprehensive evaluation, we create DescripBox and DescripBox-1024, benchmarks that span diverse scenes and resolutions. The proposed method achieves state-of-the-art performance on COCO and LVIS benchmarks while maintaining parameter efficiency, demonstrating superior layout fidelity and scalability for open-world synthesis.
☆ Neural Architecture Search by Learning a Hierarchical Search Space
Monte-Carlo Tree Search (MCTS) is a powerful tool for many non-differentiable search related problems such as adversarial games. However, the performance of such approach highly depends on the order of the nodes that are considered at each branching of the tree. If the first branches cannot distinguish between promising and deceiving configurations for the final task, the efficiency of the search is exponentially reduced. In Neural Architecture Search (NAS), as only the final architecture matters, the visiting order of the branching can be optimized to improve learning. In this paper, we study the application of MCTS to NAS for image classification. We analyze several sampling methods and branching alternatives for MCTS and propose to learn the branching by hierarchical clustering of architectures based on their similarity. The similarity is measured by the pairwise distance of output vectors of architectures. Extensive experiments on two challenging benchmarks on CIFAR10 and ImageNet show that MCTS, if provided with a good branching hierarchy, can yield promising solutions more efficiently than other approaches for NAS problems.
☆ Online Reasoning Video Segmentation with Just-in-Time Digital Twins
Reasoning segmentation (RS) aims to identify and segment objects of interest based on implicit text queries. As such, RS is a catalyst for embodied AI agents, enabling them to interpret high-level commands without requiring explicit step-by-step guidance. However, current RS approaches rely heavily on the visual perception capabilities of multimodal large language models (LLMs), leading to several major limitations. First, they struggle with queries that require multiple steps of reasoning or those that involve complex spatial/temporal relationships. Second, they necessitate LLM fine-tuning, which may require frequent updates to maintain compatibility with contemporary LLMs and may increase risks of catastrophic forgetting during fine-tuning. Finally, being primarily designed for static images or offline video processing, they scale poorly to online video data. To address these limitations, we propose an agent framework that disentangles perception and reasoning for online video RS without LLM fine-tuning. Our innovation is the introduction of a just-in-time digital twin concept, where -- given an implicit query -- a LLM plans the construction of a low-level scene representation from high-level video using specialist vision models. We refer to this approach to creating a digital twin as "just-in-time" because the LLM planner will anticipate the need for specific information and only request this limited subset instead of always evaluating every specialist model. The LLM then performs reasoning on this digital twin representation to identify target objects. To evaluate our approach, we introduce a new comprehensive video reasoning segmentation benchmark comprising 200 videos with 895 implicit text queries. The benchmark spans three reasoning categories (semantic, spatial, and temporal) with three different reasoning chain complexity.
☆ What Changed and What Could Have Changed? State-Change Counterfactuals for Procedure-Aware Video Representation Learning
Understanding a procedural activity requires modeling both how action steps transform the scene, and how evolving scene transformations can influence the sequence of action steps, even those that are accidental or erroneous. Existing work has studied procedure-aware video representations by proposing novel approaches such as modeling the temporal order of actions and has not explicitly learned the state changes (scene transformations). In this work, we study procedure-aware video representation learning by incorporating state-change descriptions generated by Large Language Models (LLMs) as supervision signals for video encoders. Moreover, we generate state-change counterfactuals that simulate hypothesized failure outcomes, allowing models to learn by imagining the unseen ``What if'' scenarios. This counterfactual reasoning facilitates the model's ability to understand the cause and effect of each step in an activity. To verify the procedure awareness of our model, we conduct extensive experiments on procedure-aware tasks, including temporal action segmentation and error detection. Our results demonstrate the effectiveness of the proposed state-change descriptions and their counterfactuals and achieve significant improvements on multiple tasks. We will make our source code and data publicly available soon.
comment: 16 pages, 4 figures
♻ ☆ Do Multimodal Large Language Models See Like Humans?
Multimodal Large Language Models (MLLMs) have achieved impressive results on various vision tasks, leveraging recent advancements in large language models. However, a critical question remains unaddressed: do MLLMs perceive visual information similarly to humans? Current benchmarks lack the ability to evaluate MLLMs from this perspective. To address this challenge, we introduce HVSBench, a large-scale benchmark designed to assess the alignment between MLLMs and the human visual system (HVS) on fundamental vision tasks that mirror human vision. HVSBench curated over 85K multimodal samples, spanning 13 categories and 5 fields in HVS, including Prominence, Subitizing, Prioritizing, Free-Viewing, and Searching. Extensive experiments demonstrate the effectiveness of our benchmark in providing a comprehensive evaluation of MLLMs. Specifically, we evaluate 13 MLLMs, revealing that even the best models show significant room for improvement, with most achieving only moderate results. Our experiments reveal that HVSBench presents a new and significant challenge for cutting-edge MLLMs. Diverse human participants attained strong performance, significantly outperforming MLLMs, which further underscores the benchmark's high quality. We believe that HVSBench will facilitate research on human-aligned and explainable MLLMs, marking a key step in understanding how MLLMs perceive and process visual information.
comment: Project page: https://jiaying.link/HVSBench/
♻ ☆ Gaga: Group Any Gaussians via 3D-aware Memory Bank
We introduce Gaga, a framework that reconstructs and segments open-world 3D scenes by leveraging inconsistent 2D masks predicted by zero-shot class-agnostic segmentation models. Contrasted to prior 3D scene segmentation approaches that rely on video object tracking or contrastive learning methods, Gaga utilizes spatial information and effectively associates object masks across diverse camera poses through a novel 3D-aware memory bank. By eliminating the assumption of continuous view changes in training images, Gaga demonstrates robustness to variations in camera poses, particularly beneficial for sparsely sampled images, ensuring precise mask label consistency. Furthermore, Gaga accommodates 2D segmentation masks from diverse sources and demonstrates robust performance with different open-world zero-shot class-agnostic segmentation models, significantly enhancing its versatility. Extensive qualitative and quantitative evaluations demonstrate that Gaga performs favorably against state-of-the-art methods, emphasizing its potential for real-world applications such as 3D scene understanding and manipulation.
comment: Project Page: https://weijielyu.github.io/Gaga
♻ ☆ ELIP: Enhanced Visual-Language Foundation Models for Image Retrieval
The objective in this paper is to improve the performance of text-to-image retrieval. To this end, we introduce a new framework that can boost the performance of large-scale pre-trained vision-language models, so that they can be used for text-to-image re-ranking. The approach, Enhanced Language-Image Pre-training (ELIP), uses the text query, via a simple MLP mapping network, to predict a set of visual prompts to condition the ViT image encoding. ELIP can easily be applied to the commonly used CLIP, SigLIP and BLIP-2 networks. To train the architecture with limited computing resources, we develop a 'student friendly' best practice, involving global hard sample mining, and curation of a large-scale dataset. On the evaluation side, we set up two new out-of-distribution (OOD) benchmarks, Occluded COCO and ImageNet-R, to assess the zero-shot generalisation of the models to different domains. The results demonstrate that ELIP significantly boosts CLIP/SigLIP/SigLIP-2 text-to-image retrieval performance and outperforms BLIP-2 on several benchmarks, as well as providing an easy means to adapt to OOD datasets.
♻ ☆ VIA: Unified Spatiotemporal Video Adaptation Framework for Global and Local Video Editing
Video editing serves as a fundamental pillar of digital media, spanning applications in entertainment, education, and professional communication. However, previous methods often overlook the necessity of comprehensively understanding both global and local contexts, leading to inaccurate and inconsistent edits in the spatiotemporal dimension, especially for long videos. In this paper, we introduce VIA, a unified spatiotemporal Video Adaptation framework for global and local video editing, pushing the limits of consistently editing minute-long videos. First, to ensure local consistency within individual frames, we designed test-time editing adaptation to adapt a pre-trained image editing model for improving consistency between potential editing directions and the text instruction, and adapts masked latent variables for precise local control. Furthermore, to maintain global consistency over the video sequence, we introduce spatiotemporal adaptation that recursively gather consistent attention variables in key frames and strategically applies them across the whole sequence to realize the editing effects. Extensive experiments demonstrate that, compared to baseline methods, our VIA approach produces edits that are more faithful to the source videos, more coherent in the spatiotemporal context, and more precise in local control. More importantly, we show that VIA can achieve consistent long video editing in minutes, unlocking the potential for advanced video editing tasks over long video sequences.
comment: 18 pages, 16 figures
♻ ☆ A Survey on Computational Solutions for Reconstructing Complete Objects by Reassembling Their Fractured Parts
Reconstructing a complete object from its parts is a fundamental problem in many scientific domains. The purpose of this article is to provide a systematic survey on this topic. The reassembly problem requires understanding the attributes of individual pieces and establishing matches between different pieces. Many approaches also model priors of the underlying complete object. Existing approaches are tightly connected problems of shape segmentation, shape matching, and learning shape priors. We provide existing algorithms in this context and emphasize their similarities and differences to general-purpose approaches. We also survey the trends from early non-deep learning approaches to more recent deep learning approaches. In addition to algorithms, this survey will also describe existing datasets, open-source software packages, and applications. To the best of our knowledge, this is the first comprehensive survey on this topic in computer graphics.
comment: 36 pages, 22 figures
♻ ☆ OVO-Bench: How Far is Your Video-LLMs from Real-World Online Video Understanding? CVPR 2025
Temporal Awareness, the ability to reason dynamically based on the timestamp when a question is raised, is the key distinction between offline and online video LLMs. Unlike offline models, which rely on complete videos for static, post hoc analysis, online models process video streams incrementally and dynamically adapt their responses based on the timestamp at which the question is posed. Despite its significance, temporal awareness has not been adequately evaluated in existing benchmarks. To fill this gap, we present OVO-Bench (Online-VideO-Benchmark), a novel video benchmark that emphasizes the importance of timestamps for advanced online video understanding capability benchmarking. OVO-Bench evaluates the ability of video LLMs to reason and respond to events occurring at specific timestamps under three distinct scenarios: (1) Backward tracing: trace back to past events to answer the question. (2) Real-time understanding: understand and respond to events as they unfold at the current timestamp. (3) Forward active responding: delay the response until sufficient future information becomes available to answer the question accurately. OVO-Bench comprises 12 tasks, featuring 644 unique videos and approximately human-curated 2,800 fine-grained meta-annotations with precise timestamps. We combine automated generation pipelines with human curation. With these high-quality samples, we further developed an evaluation pipeline to systematically query video LLMs along the video timeline. Evaluations of nine Video-LLMs reveal that, despite advancements on traditional benchmarks, current models struggle with online video understanding, showing a significant gap compared to human agents. We hope OVO-Bench will drive progress in video LLMs and inspire future research in online video reasoning. Our benchmark and code can be accessed at https://github.com/JoeLeelyf/OVO-Bench.
comment: CVPR 2025
♻ ☆ Multi-View and Multi-Scale Alignment for Contrastive Language-Image Pre-training in Mammography
Contrastive Language-Image Pre-training (CLIP) demonstrates strong potential in medical image analysis but requires substantial data and computational resources. Due to these restrictions, existing CLIP applications in medical imaging focus mainly on modalities like chest X-rays that have abundant image-report data available, leaving many other important modalities underexplored. Here, we propose one of the first adaptations of the full CLIP model to mammography, which presents significant challenges due to labeled data scarcity, high-resolution images with small regions of interest, and class-wise imbalance. We first develop a specialized supervision framework for mammography that leverages its multi-view nature. Furthermore, we design a symmetric local alignment module to better focus on detailed features in high-resolution images. Lastly, we incorporate a parameter-efficient fine-tuning approach for large language models pre-trained with medical knowledge to address data limitations. Our multi-view and multi-scale alignment (MaMA) method outperforms state-of-the-art baselines for three different tasks on two large real-world mammography datasets, EMBED and RSNA-Mammo, with only 52% model size compared with the largest baseline. The code is available at https://github.com/XYPB/MaMA
comment: This paper is accepted by IPMI 2025 for Oral Presentation
♻ ☆ SlowFast-LLaVA-1.5: A Family of Token-Efficient Video Large Language Models for Long-Form Video Understanding
We introduce SlowFast-LLaVA-1.5 (abbreviated as SF-LLaVA-1.5), a family of video large language models (LLMs) offering a token-efficient solution for long-form video understanding. We incorporate the two-stream SlowFast mechanism into a streamlined training pipeline, and perform joint video-image training on a carefully curated data mixture of only publicly available datasets. Our primary focus is on highly efficient model scales (1B and 3B), demonstrating that even relatively small Video LLMs can achieve state-of-the-art performance on video understanding, meeting the demand for mobile-friendly models. Experimental results demonstrate that SF-LLaVA-1.5 achieves superior performance on a wide range of video and image tasks, with robust results at all model sizes (ranging from 1B to 7B). Notably, SF-LLaVA-1.5 achieves state-of-the-art results in long-form video understanding (e.g., LongVideoBench and MLVU) and excels at small scales across various video benchmarks.
comment: Technical report
♻ ☆ TripoSG: High-Fidelity 3D Shape Synthesis using Large-Scale Rectified Flow Models
Recent advancements in diffusion techniques have propelled image and video generation to unprecedented levels of quality, significantly accelerating the deployment and application of generative AI. However, 3D shape generation technology has so far lagged behind, constrained by limitations in 3D data scale, complexity of 3D data processing, and insufficient exploration of advanced techniques in the 3D domain. Current approaches to 3D shape generation face substantial challenges in terms of output quality, generalization capability, and alignment with input conditions. We present TripoSG, a new streamlined shape diffusion paradigm capable of generating high-fidelity 3D meshes with precise correspondence to input images. Specifically, we propose: 1) A large-scale rectified flow transformer for 3D shape generation, achieving state-of-the-art fidelity through training on extensive, high-quality data. 2) A hybrid supervised training strategy combining SDF, normal, and eikonal losses for 3D VAE, achieving high-quality 3D reconstruction performance. 3) A data processing pipeline to generate 2 million high-quality 3D samples, highlighting the crucial rules for data quality and quantity in training 3D generative models. Through comprehensive experiments, we have validated the effectiveness of each component in our new framework. The seamless integration of these parts has enabled TripoSG to achieve state-of-the-art performance in 3D shape generation. The resulting 3D shapes exhibit enhanced detail due to high-resolution capabilities and demonstrate exceptional fidelity to input images. Moreover, TripoSG demonstrates improved versatility in generating 3D models from diverse image styles and contents, showcasing strong generalization capabilities. To foster progress and innovation in the field of 3D generation, we will make our model publicly available.
♻ ☆ BACON: Improving Clarity of Image Captions via Bag-of-Concept Graphs
Advancements in large Vision-Language Models have brought precise, accurate image captioning, vital for advancing multi-modal image understanding and processing. Yet these captions often carry lengthy, intertwined contexts that are difficult to parse and frequently overlook essential cues, posing a great barrier for models like GroundingDINO and SDXL, which lack the strong text encoding and syntax analysis needed to fully leverage dense captions. To address this, we propose BACON, a prompting method that breaks down VLM-generated captions into disentangled, structured elements such as objects, relationships, styles, and themes. This approach not only minimizes confusion from handling complex contexts but also allows for efficient transfer into a JSON dictionary, enabling models without linguistic processing capabilities to easily access key information. We annotated 100,000 image-caption pairs using BACON with GPT-4V and trained an LLaVA captioner on this dataset, enabling it to produce BACON-style captions without relying on costly GPT-4V. Evaluations of overall quality, precision, and recall-as well as user studies-demonstrate that the resulting caption model consistently outperforms other SOTA VLM models in generating high-quality captions. Besides, we show that BACON-style captions exhibit better clarity when applied to various models, enabling them to accomplish previously unattainable tasks or surpass existing SOTA solutions without training. For example, BACON-style captions help GroundingDINO achieve 1.51x higher recall scores on open-vocabulary object detection tasks compared to leading methods.
♻ ☆ StableMamba: Distillation-free Scaling of Large SSMs for Images and Videos
State-space models (SSMs), exemplified by S4, have introduced a novel context modeling method by integrating state-space techniques into deep learning. However, they struggle with global context modeling due to their data-independent matrices. The Mamba model addressed this with data-dependent variants via the S6 selective-scan algorithm, enhancing context modeling, especially for long sequences. However, Mamba-based architectures are difficult to scale with respect to the number of parameters, which is a major limitation for vision applications. This paper addresses the scalability issue of large SSMs for image classification and action recognition without requiring additional techniques like knowledge distillation. We analyze the distinct characteristics of Mamba-based and Attention-based models, proposing a Mamba-Attention interleaved architecture that enhances scalability, robustness, and performance. We demonstrate that the stable and efficient interleaved architecture resolves the scalability issue of Mamba-based architectures for images and videos and increases robustness to common artifacts like JPEG compression. Our thorough evaluation on the ImageNet-1K, Kinetics-400 and Something-Something-v2 benchmarks demonstrates that our approach improves the accuracy of state-of-the-art Mamba-based architectures by up to $+1.7$.
♻ ☆ Frequency-Controlled Diffusion Model for Versatile Text-Guided Image-to-Image Translation AAAI
Recently, large-scale text-to-image (T2I) diffusion models have emerged as a powerful tool for image-to-image translation (I2I), allowing open-domain image translation via user-provided text prompts. This paper proposes frequency-controlled diffusion model (FCDiffusion), an end-to-end diffusion-based framework that contributes a novel solution to text-guided I2I from a frequency-domain perspective. At the heart of our framework is a feature-space frequency-domain filtering module based on Discrete Cosine Transform, which filters the latent features of the source image in the DCT domain, yielding filtered image features bearing different DCT spectral bands as different control signals to the pre-trained Latent Diffusion Model. We reveal that control signals of different DCT spectral bands bridge the source image and the T2I generated image in different correlations (e.g., style, structure, layout, contour, etc.), and thus enable versatile I2I applications emphasizing different I2I correlations, including style-guided content creation, image semantic manipulation, image scene translation, and image style translation. Different from related approaches, FCDiffusion establishes a unified text-guided I2I framework suitable for diverse image translation tasks simply by switching among different frequency control branches at inference time. The effectiveness and superiority of our method for text-guided I2I are demonstrated with extensive experiments both qualitatively and quantitatively. Our project is publicly available at: https://xianggao1102.github.io/FCDiffusion/.
comment: Proceedings of the 38th AAAI Conference on Artificial Intelligence (AAAI 2024)
♻ ☆ OmniBench: Towards The Future of Universal Omni-Language Models
Recent advancements in multimodal large language models (MLLMs) have focused on integrating multiple modalities, yet their ability to simultaneously process and reason across different inputs remains underexplored. We introduce OmniBench, a novel benchmark designed to evaluate models' ability to recognize, interpret, and reason across visual, acoustic, and textual inputs simultaneously. We define language models capable of such tri-modal processing as omni-language models (OLMs). OmniBench features high-quality human annotations that require integrated understanding across all modalities. Our evaluation reveals that: i) open-source OLMs show significant limitations in instruction-following and reasoning in tri-modal contexts; and ii) most baseline models perform poorly (around 50% accuracy) even with textual alternatives to image/audio inputs. To address these limitations, we develop OmniInstruct, an 96K-sample instruction tuning dataset for training OLMs. We advocate for developing more robust tri-modal integration techniques and training strategies to enhance OLM performance. Codes and data could be found at our repo (https://github.com/multimodal-art-projection/OmniBench).
♻ ☆ Vision language models are blind: Failing to translate detailed visual features into words
While large language models with vision capabilities (VLMs), e.g., GPT-4o and Gemini 1.5 Pro, score high on many vision-understanding benchmarks, they are still struggling with low-level vision tasks that are easy to humans. Specifically, on BlindTest, our suite of 7 very simple tasks, including identifying (a) whether two circles overlap; (b) how many times two lines intersect; (c) which letter is being circled in a word; and (d) the number of circles in an Olympic-like logo, four state-of-the-art VLMs are only 58.07% accurate on average. Claude 3.5 Sonnet performs the best at 77.84% accuracy, far from the human expected accuracy of 100%. Across different image resolutions and line widths, VLMs including slow-thinking models consistently struggle with those tasks that require precise spatial information when geometric primitives overlap or are close. Yet, VLMs perform at near-100% accuracy when much more space is added to separate shapes and letters. Linear probing experiments show that vision encoders contain sufficient visual information to solve BlindTest and that language models fail to decode this information into correct answers. Code and data are at: https://vlmsareblind.github.io
♻ ☆ On the Viability of Semi-Supervised Segmentation Methods for Statistical Shape Modeling
Statistical Shape Models (SSMs) excel at identifying population level anatomical variations, which is at the core of various clinical and biomedical applications, including morphology-based diagnostics and surgical planning. However, the effectiveness of SSM is often constrained by the necessity for expert-driven manual segmentation, a process that is both time-intensive and expensive, thereby restricting their broader application and utility. Recent deep learning approaches enable the direct estimation of Statistical Shape Models (SSMs) from unsegmented images. While these models can predict SSMs without segmentation during deployment, they do not address the challenge of acquiring the manual annotations needed for training, particularly in resource-limited settings. Semi-supervised models for anatomy segmentation can mitigate the annotation burden. Yet, despite the abundance of available approaches, there are no established guidelines to inform end-users on their effectiveness for the downstream task of constructing SSMs. In this study, we systematically evaluate the potential of semi-supervised methods as viable alternatives to manual segmentations for building SSMs. We establish a new performance benchmark by employing various semi-supervised methods for anatomy segmentation under low annotation settings, utilizing the predicted segmentations for the task of SSM. Our results indicate that some methods produce noisy segmentation, which is very unfavorable for SSM tasks, while others can capture the correct modes of variations in the population cohort with 60-80% reduction in required manual annotation
♻ ☆ Self-Contrastive Forward-Forward Algorithm
Agents that operate autonomously benefit from lifelong learning capabilities. However, compatible training algorithms must comply with the decentralized nature of these systems, which imposes constraints on both the parameter counts and the computational resources. The Forward-Forward (FF) algorithm is one of these. FF relies only on feedforward operations, the same used for inference, for optimizing layer-wise objectives. This purely forward approach eliminates the need for transpose operations required in traditional backpropagation. Despite its potential, FF has failed to reach state-of-the-art performance on most standard benchmark tasks, in part due to unreliable negative data generation methods for unsupervised learning. In this work, we propose the Self-Contrastive Forward-Forward (SCFF) algorithm, a competitive training method aimed at closing this performance gap. Inspired by standard self-supervised contrastive learning for vision tasks, SCFF generates positive and negative inputs applicable across various datasets. The method demonstrates superior performance compared to existing unsupervised local learning algorithms on several benchmark datasets, including MNIST, CIFAR-10, STL-10, and Tiny ImageNet. We extend FF's application to training recurrent neural networks, expanding its utility to sequential data tasks. These findings pave the way for high-accuracy, real-time learning on resource-constrained edge devices.
♻ ☆ Layer- and Timestep-Adaptive Differentiable Token Compression Ratios for Efficient Diffusion Transformers CVPR 2025
Diffusion Transformers (DiTs) have achieved state-of-the-art (SOTA) image generation quality but suffer from high latency and memory inefficiency, making them difficult to deploy on resource-constrained devices. One major efficiency bottleneck is that existing DiTs apply equal computation across all regions of an image. However, not all image tokens are equally important, and certain localized areas require more computation, such as objects. To address this, we propose DiffCR, a dynamic DiT inference framework with differentiable compression ratios, which automatically learns to dynamically route computation across layers and timesteps for each image token, resulting in efficient DiTs. Specifically, DiffCR integrates three features: (1) A token-level routing scheme where each DiT layer includes a router that is fine-tuned jointly with model weights to predict token importance scores. In this way, unimportant tokens bypass the entire layer's computation; (2) A layer-wise differentiable ratio mechanism where different DiT layers automatically learn varying compression ratios from a zero initialization, resulting in large compression ratios in redundant layers while others remain less compressed or even uncompressed; (3) A timestep-wise differentiable ratio mechanism where each denoising timestep learns its own compression ratio. The resulting pattern shows higher ratios for noisier timesteps and lower ratios as the image becomes clearer. Extensive experiments on text-to-image and inpainting tasks show that DiffCR effectively captures dynamism across token, layer, and timestep axes, achieving superior trade-offs between generation quality and efficiency compared to prior works. The project website is available at https://www.haoranyou.com/diffcr.
comment: Accepted by CVPR 2025
♻ ☆ Learning Multi-modal Representations by Watching Hundreds of Surgical Video Lectures
Recent advancements in surgical computer vision applications have been driven by vision-only models, which do not explicitly integrate the rich semantics of language into their design. These methods rely on manually annotated surgical videos to predict a fixed set of object categories, limiting their generalizability to unseen surgical procedures and downstream tasks. In this work, we put forward the idea that the surgical video lectures available through open surgical e-learning platforms can provide effective vision and language supervisory signals for multi-modal representation learning without relying on manual annotations. We address the surgery-specific linguistic challenges present in surgical video lectures by employing multiple complementary automatic speech recognition systems to generate text transcriptions. We then present a novel method, SurgVLP - Surgical Vision Language Pre-training, for multi-modal representation learning. Extensive experiments across diverse surgical procedures and tasks demonstrate that the multi-modal representations learned by SurgVLP exhibit strong transferability and adaptability in surgical video analysis. Furthermore, our zero-shot evaluations highlight SurgVLP's potential as a general-purpose foundation model for surgical workflow analysis, reducing the reliance on extensive manual annotations for downstream tasks, and facilitating adaptation methods such as few-shot learning to build a scalable and data-efficient solution for various downstream surgical applications. The [training code](https://github.com/CAMMA-public/SurgVLP) and [weights](https://github.com/CAMMA-public/PeskaVLP) are public.
♻ ☆ GMAI-VL & GMAI-VL-5.5M: A Large Vision-Language Model and A Comprehensive Multimodal Dataset Towards General Medical AI
Despite significant advancements in general AI, its effectiveness in the medical domain is limited by the lack of specialized medical knowledge. To address this, we formulate GMAI-VL-5.5M, a multimodal medical dataset created by converting hundreds of specialized medical datasets with various annotations into high-quality image-text pairs. This dataset offers comprehensive task coverage, diverse modalities, and rich image-text data. Building upon this dataset, we develop GMAI-VL, a general medical vision-language model, with a three-stage training strategy that enhances the integration of visual and textual information. This approach significantly improves the model's ability to process multimodal data, supporting accurate diagnoses and clinical decision-making. Experiments show that GMAI-VL achieves state-of-the-art performance across various multimodal medical tasks, including visual question answering and medical image diagnosis.
♻ ☆ Gaussian Splatting Lucas-Kanade
Gaussian Splatting and its dynamic extensions are effective for reconstructing 3D scenes from 2D images when there is significant camera movement to facilitate motion parallax and when scene objects remain relatively static. However, in many real-world scenarios, these conditions are not met. As a consequence, data-driven semantic and geometric priors have been favored as regularizers, despite their bias toward training data and their neglect of broader movement dynamics. Departing from this practice, we propose a novel analytical approach that adapts the classical Lucas-Kanade method to dynamic Gaussian splatting. By leveraging the intrinsic properties of the forward warp field network, we derive an analytical velocity field that, through time integration, facilitates accurate scene flow computation. This enables the precise enforcement of motion constraints on warp fields, thus constraining both 2D motion and 3D positions of the Gaussians. Our method excels in reconstructing highly dynamic scenes with minimal camera movement, as demonstrated through experiments on both synthetic and real-world scenes.
comment: International Conference on Learning Representations
♻ ☆ Discretized Gaussian Representation for Tomographic Reconstruction
Computed Tomography (CT) is a widely used imaging technique that provides detailed cross-sectional views of objects. Over the past decade, Deep Learning-based Reconstruction (DLR) methods have led efforts to enhance image quality and reduce noise, yet they often require large amounts of data and are computationally intensive. Inspired by recent advancements in scene reconstruction, some approaches have adapted NeRF and 3D Gaussian Splatting (3DGS) techniques for CT reconstruction. However, these methods are not ideal for direct 3D volume reconstruction. In this paper, we propose a novel Discretized Gaussian Representation (DGR) for CT reconstruction, which directly reconstructs the 3D volume using a set of discretized Gaussian functions in an end-to-end manner. To further enhance computational efficiency, we introduce a Fast Volume Reconstruction technique that aggregates the contributions of these Gaussians into a discretized volume in a highly parallelized fashion. Our extensive experiments on both real-world and synthetic datasets demonstrate that DGR achieves superior reconstruction quality and significantly improved computational efficiency compared to existing DLR and instance reconstruction methods. Our code has been provided for review purposes and will be made publicly available upon publication.
♻ ☆ Contextual AD Narration with Interleaved Multimodal Sequence
The Audio Description (AD) task aims to generate descriptions of visual elements for visually impaired individuals to help them access long-form video content, like movies. With video feature, text, character bank and context information as inputs, the generated ADs are able to correspond to the characters by name and provide reasonable, contextual descriptions to help audience understand the storyline of movie. To achieve this goal, we propose to leverage pre-trained foundation models through a simple and unified framework to generate ADs with interleaved multimodal sequence as input, termed as Uni-AD. To enhance the alignment of features across various modalities with finer granularity, we introduce a simple and lightweight module that maps video features into the textual feature space. Moreover, we also propose a character-refinement module to provide more precise information by identifying the main characters who play more significant roles in the video context. With these unique designs, we further incorporate contextual information and a contrastive loss into our architecture to generate smoother and more contextually appropriate ADs. Experiments on multiple AD datasets show that Uni-AD performs well on AD generation, which demonstrates the effectiveness of our approach. Our code is available at: https://github.com/ant-research/UniAD.
♻ ☆ GI-GS: Global Illumination Decomposition on Gaussian Splatting for Inverse Rendering
We present GI-GS, a novel inverse rendering framework that leverages 3D Gaussian Splatting (3DGS) and deferred shading to achieve photo-realistic novel view synthesis and relighting. In inverse rendering, accurately modeling the shading processes of objects is essential for achieving high-fidelity results. Therefore, it is critical to incorporate global illumination to account for indirect lighting that reaches an object after multiple bounces across the scene. Previous 3DGS-based methods have attempted to model indirect lighting by characterizing indirect illumination as learnable lighting volumes or additional attributes of each Gaussian, while using baked occlusion to represent shadow effects. These methods, however, fail to accurately model the complex physical interactions between light and objects, making it impossible to construct realistic indirect illumination during relighting. To address this limitation, we propose to calculate indirect lighting using efficient path tracing with deferred shading. In our framework, we first render a G-buffer to capture the detailed geometry and material properties of the scene. Then, we perform physically-based rendering (PBR) only for direct lighting. With the G-buffer and previous rendering results, the indirect lighting can be calculated through a lightweight path tracing. Our method effectively models indirect lighting under any given lighting conditions, thereby achieving better novel view synthesis and competitive relighting. Quantitative and qualitative results show that our GI-GS outperforms existing baselines in both rendering quality and efficiency.
comment: Camera-ready version. Project page: https://stopaimme.github.io/GI-GS-site/
♻ ☆ Imitating Radiological Scrolling: A Global-Local Attention Model for 3D Chest CT Volumes Multi-Label Anomaly Classification
The rapid increase in the number of Computed Tomography (CT) scan examinations has created an urgent need for automated tools, such as organ segmentation, anomaly classification, and report generation, to assist radiologists with their growing workload. Multi-label classification of Three-Dimensional (3D) CT scans is a challenging task due to the volumetric nature of the data and the variety of anomalies to be detected. Existing deep learning methods based on Convolutional Neural Networks (CNNs) struggle to capture long-range dependencies effectively, while Vision Transformers require extensive pre-training, posing challenges for practical use. Additionally, these existing methods do not explicitly model the radiologist's navigational behavior while scrolling through CT scan slices, which requires both global context understanding and local detail awareness. In this study, we present CT-Scroll, a novel global-local attention model specifically designed to emulate the scrolling behavior of radiologists during the analysis of 3D CT scans. Our approach is evaluated on two public datasets, demonstrating its efficacy through comprehensive experiments and an ablation study that highlights the contribution of each model component.
comment: 13 pages, 4 figures. Accepted for MIDL 2025
♻ ☆ Image segmentation of treated and untreated tumor spheroids by Fully Convolutional Networks
Multicellular tumor spheroids (MCTS) are advanced cell culture systems for assessing the impact of combinatorial radio(chemo)therapy. They exhibit therapeutically relevant in-vivo-like characteristics from 3D cell-cell and cell-matrix interactions to radial pathophysiological gradients related to proliferative activity and nutrient/oxygen supply, altering cellular radioresponse. State-of-the-art assays quantify long-term curative endpoints based on collected brightfield image time series from large treated spheroid populations per irradiation dose and treatment arm. Here, spheroid control probabilities are documented analogous to in-vivo tumor control probabilities based on Kaplan-Meier curves. This analyses require laborious spheroid segmentation of up to 100.000 images per treatment arm to extract relevant structural information from the images, e.g., diameter, area, volume and circularity. While several image analysis algorithms are available for spheroid segmentation, they all focus on compact MCTS with clearly distinguishable outer rim throughout growth. However, treated MCTS may partly be detached and destroyed and are usually obscured by dead cell debris. We successfully train two Fully Convolutional Networks, UNet and HRNet, and optimize their hyperparameters to develop an automatic segmentation for both untreated and treated MCTS. We systematically validate the automatic segmentation on larger, independent data sets of spheroids derived from two human head-and-neck cancer cell lines. We find an excellent overlap between manual and automatic segmentation for most images, quantified by Jaccard indices at around 90%. For images with smaller overlap of the segmentations, we demonstrate that this error is comparable to the variations across segmentations from different biological experts, suggesting that these images represent biologically unclear or ambiguous cases.
comment: 30 pages, 23 figures
♻ ☆ TREAD: Token Routing for Efficient Architecture-agnostic Diffusion Training
Diffusion models have emerged as the mainstream approach for visual generation. However, these models typically suffer from sample inefficiency and high training costs. Consequently, methods for efficient finetuning, inference and personalization were quickly adopted by the community. However, training these models in the first place remains very costly. While several recent approaches - including masking, distillation, and architectural modifications - have been proposed to improve training efficiency, each of these methods comes with a tradeoff: they achieve enhanced performance at the expense of increased computational cost or vice versa. In contrast, this work aims to improve training efficiency as well as generative performance at the same time through routes that act as a transport mechanism for randomly selected tokens from early layers to deeper layers of the model. Our method is not limited to the common transformer-based model - it can also be applied to state-space models and achieves this without architectural modifications or additional parameters. Finally, we show that TREAD reduces computational cost and simultaneously boosts model performance on the standard ImageNet-256 benchmark in class-conditional synthesis. Both of these benefits multiply to a convergence speedup of 14x at 400K training iterations compared to DiT and 37x compared to the best benchmark performance of DiT at 7M training iterations. Furthermore, we achieve a competitive FID of 2.09 in a guided and 3.93 in an unguided setting, which improves upon the DiT, without architectural changes.
♻ ☆ Video-Panda: Parameter-efficient Alignment for Encoder-free Video-Language Models CVPR 2025
We present an efficient encoder-free approach for video-language understanding that achieves competitive performance while significantly reducing computational overhead. Current video-language models typically rely on heavyweight image encoders (300M-1.1B parameters) or video encoders (1B-1.4B parameters), creating a substantial computational burden when processing multi-frame videos. Our method introduces a novel Spatio-Temporal Alignment Block (STAB) that directly processes video inputs without requiring pre-trained encoders while using only 45M parameters for visual processing - at least a 6.5$\times$ reduction compared to traditional approaches. The STAB architecture combines Local Spatio-Temporal Encoding for fine-grained feature extraction, efficient spatial downsampling through learned attention and separate mechanisms for modeling frame-level and video-level relationships. Our model achieves comparable or superior performance to encoder-based approaches for open-ended video question answering on standard benchmarks. The fine-grained video question-answering evaluation demonstrates our model's effectiveness, outperforming the encoder-based approaches Video-ChatGPT and Video-LLaVA in key aspects like correctness and temporal understanding. Extensive ablation studies validate our architectural choices and demonstrate the effectiveness of our spatio-temporal modeling approach while achieving 3-4$\times$ faster processing speeds than previous methods. Code is available at https://jh-yi.github.io/Video-Panda.
comment: CVPR 2025 camera-ready version
♻ ☆ Demand Estimation with Text and Image Data
We propose a demand estimation method that leverages unstructured text and image data to infer substitution patterns. Using pre-trained deep learning models, we extract embeddings from product images and textual descriptions and incorporate them into a random coefficients logit model. This approach enables researchers to estimate demand even when they lack data on product attributes or when consumers value hard-to-quantify attributes, such as visual design or functional benefits. Using data from a choice experiment, we show that our approach outperforms standard attribute-based models in counterfactual predictions of consumers' second choices. We also apply it across 40 product categories on Amazon and consistently find that text and image data help identify close substitutes within each category.
♻ ☆ Quantization-aware Matrix Factorization for Low Bit Rate Image Compression
Lossy image compression is essential for efficient transmission and storage. Traditional compression methods mainly rely on discrete cosine transform (DCT) or singular value decomposition (SVD), both of which represent image data in continuous domains and, therefore, necessitate carefully designed quantizers. Notably, these methods consider quantization as a separate step, where quantization errors cannot be incorporated into the compression process. The sensitivity of these methods, especially SVD-based ones, to quantization errors significantly degrades reconstruction quality. To address this issue, we introduce a quantization-aware matrix factorization (QMF) to develop a novel lossy image compression method. QMF provides a low-rank representation of the image data as a product of two smaller factor matrices, with elements constrained to bounded integer values, thereby effectively integrating quantization with low-rank approximation. We propose an efficient, provably convergent iterative algorithm for QMF using a block coordinate descent (BCD) scheme, with subproblems having closed-form solutions. Our experiments on the Kodak and CLIC 2024 datasets demonstrate that our QMF compression method consistently outperforms JPEG at low bit rates below 0.25 bits per pixel (bpp) and remains comparable at higher bit rates. We also assessed our method's capability to preserve visual semantics by evaluating an ImageNet pre-trained classifier on compressed images. Remarkably, our method improved top-1 accuracy by over 5 percentage points compared to JPEG at bit rates under 0.25 bpp. The project is available at https://github.com/pashtari/lrf .
comment: 22 pages, 6 figures, 1 table, 1 algorithm
♻ ☆ UAV-DETR: Efficient End-to-End Object Detection for Unmanned Aerial Vehicle Imagery
Unmanned aerial vehicle object detection (UAV-OD) has been widely used in various scenarios. However, most existing UAV-OD algorithms rely on manually designed components, which require extensive tuning. End-to-end models that do not depend on such manually designed components are mainly designed for natural images, which are less effective for UAV imagery. To address such challenges, this paper proposes an efficient detection transformer (DETR) framework tailored for UAV imagery, i.e., UAV-DETR. The framework includes a multi-scale feature fusion with frequency enhancement module, which captures both spatial and frequency information at different scales. In addition, a frequency-focused down-sampling module is presented to retain critical spatial details during down-sampling. A semantic alignment and calibration module is developed to align and fuse features from different fusion paths. Experimental results demonstrate the effectiveness and generalization of our approach across various UAV imagery datasets. On the VisDrone dataset, our method improves AP by 3.1\% and $\text{AP}_{50}$ by 4.2\% over the baseline. Similar enhancements are observed on the UAVVaste dataset. The project page: https://github.com/ValiantDiligent/UAV-DETR
♻ ☆ SegMAN: Omni-scale Context Modeling with State Space Models and Local Attention for Semantic Segmentation CVPR 2025
High-quality semantic segmentation relies on three key capabilities: global context modeling, local detail encoding, and multi-scale feature extraction. However, recent methods struggle to possess all these capabilities simultaneously. Hence, we aim to empower segmentation networks to simultaneously carry out efficient global context modeling, high-quality local detail encoding, and rich multi-scale feature representation for varying input resolutions. In this paper, we introduce SegMAN, a novel linear-time model comprising a hybrid feature encoder dubbed SegMAN Encoder, and a decoder based on state space models. Specifically, the SegMAN Encoder synergistically integrates sliding local attention with dynamic state space models, enabling highly efficient global context modeling while preserving fine-grained local details. Meanwhile, the MMSCopE module in our decoder enhances multi-scale context feature extraction and adaptively scales with the input resolution. Our SegMAN-B Encoder achieves 85.1% ImageNet-1k accuracy (+1.5% over VMamba-S with fewer parameters). When paired with our decoder, the full SegMAN-B model achieves 52.6% mIoU on ADE20K (+1.6% over SegNeXt-L with 15% fewer GFLOPs), 83.8% mIoU on Cityscapes (+2.1% over SegFormer-B3 with half the GFLOPs), and 1.6% higher mIoU than VWFormer-B3 on COCO-Stuff with lower GFLOPs. Our code is available at https://github.com/yunxiangfu2001/SegMAN.
comment: CVPR 2025
♻ ☆ How NeRFs and 3D Gaussian Splatting are Reshaping SLAM: a Survey
Over the past two decades, research in the field of Simultaneous Localization and Mapping (SLAM) has undergone a significant evolution, highlighting its critical role in enabling autonomous exploration of unknown environments. This evolution ranges from hand-crafted methods, through the era of deep learning, to more recent developments focused on Neural Radiance Fields (NeRFs) and 3D Gaussian Splatting (3DGS) representations. Recognizing the growing body of research and the absence of a comprehensive survey on the topic, this paper aims to provide the first comprehensive overview of SLAM progress through the lens of the latest advancements in radiance fields. It sheds light on the background, evolutionary path, inherent strengths and limitations, and serves as a fundamental reference to highlight the dynamic progress and specific challenges.
comment: Updated to November 2024
♻ ☆ Consistency Trajectory Matching for One-Step Generative Super-Resolution
Current diffusion-based super-resolution (SR) approaches achieve commendable performance at the cost of high inference overhead. Therefore, distillation techniques are utilized to accelerate the multi-step teacher model into one-step student model. Nevertheless, these methods significantly raise training costs and constrain the performance of the student model by the teacher model. To overcome these tough challenges, we propose Consistency Trajectory Matching for Super-Resolution (CTMSR), a distillation-free strategy that is able to generate photo-realistic SR results in one step. Concretely, we first formulate a Probability Flow Ordinary Differential Equation (PF-ODE) trajectory to establish a deterministic mapping from low-resolution (LR) images with noise to high-resolution (HR) images. Then we apply the Consistency Training (CT) strategy to directly learn the mapping in one step, eliminating the necessity of pre-trained diffusion model. To further enhance the performance and better leverage the ground-truth during the training process, we aim to align the distribution of SR results more closely with that of the natural images. To this end, we propose to minimize the discrepancy between their respective PF-ODE trajectories from the LR image distribution by our meticulously designed Distribution Trajectory Matching (DTM) loss, resulting in improved realism of our recovered HR images. Comprehensive experimental results demonstrate that the proposed methods can attain comparable or even superior capabilities on both synthetic and real datasets while maintaining minimal inference latency.
♻ ☆ Improving Object Detection by Modifying Synthetic Data with Explainable AI
Limited real-world data severely impacts model performance in many computer vision domains, particularly for samples that are underrepresented in training. Synthetically generated images are a promising solution, but 1) it remains unclear how to design synthetic training data to optimally improve model performance (e.g, whether and where to introduce more realism or more abstraction) and 2) the domain expertise, time and effort required from human operators for this design and optimisation process represents a major practical challenge. Here we propose a novel conceptual approach to improve the efficiency of designing synthetic images, by using robust Explainable AI (XAI) techniques to guide a human-in-the-loop process of modifying 3D mesh models used to generate these images. Importantly, this framework allows both modifications that increase and decrease realism in synthetic data, which can both improve model performance. We illustrate this concept using a real-world example where data are sparse; detection of vehicles in infrared imagery. We fine-tune an initial YOLOv8 model on the ATR DSIAC infrared dataset and synthetic images generated from 3D mesh models in the Unity gaming engine, and then use XAI saliency maps to guide modification of our Unity models. We show that synthetic data can improve detection of vehicles in orientations unseen in training by 4.6% (to mAP50 = 94.6%). We further improve performance by an additional 1.5% (to 96.1%) through our new XAI-guided approach, which reduces misclassifications through both increasing and decreasing the realism of different parts of the synthetic data. Our proof-of-concept results pave the way for fine, XAI-controlled curation of synthetic datasets tailored to improve object detection performance, whilst simultaneously reducing the burden on human operators in designing and optimising these datasets.
♻ ☆ Dynamics-Aware Gaussian Splatting Streaming Towards Fast On-the-Fly 4D Reconstruction
The recent development of 3D Gaussian Splatting (3DGS) has led to great interest in 4D dynamic spatial reconstruction. Existing approaches mainly rely on full-length multi-view videos, while there has been limited exploration of online reconstruction methods that enable on-the-fly training and per-timestep streaming. Current 3DGS-based streaming methods treat the Gaussian primitives uniformly and constantly renew the densified Gaussians, thereby overlooking the difference between dynamic and static features as well as neglecting the temporal continuity in the scene. To address these limitations, we propose a novel three-stage pipeline for iterative streamable 4D dynamic spatial reconstruction. Our pipeline comprises a selective inheritance stage to preserve temporal continuity, a dynamics-aware shift stage to distinguish dynamic and static primitives and optimize their movements, and an error-guided densification stage to accommodate emerging objects. Our method achieves state-of-the-art performance in online 4D reconstruction, demonstrating the fastest on-the-fly training, superior representation quality, and real-time rendering capability. Project page: https://www.liuzhening.top/DASS
comment: Project page: https://www.liuzhening.top/DASS
♻ ☆ EventMamba: Enhancing Spatio-Temporal Locality with State Space Models for Event-Based Video Reconstruction
Leveraging its robust linear global modeling capability, Mamba has notably excelled in computer vision. Despite its success, existing Mamba-based vision models have overlooked the nuances of event-driven tasks, especially in video reconstruction. Event-based video reconstruction (EBVR) demands spatial translation invariance and close attention to local event relationships in the spatio-temporal domain. Unfortunately, conventional Mamba algorithms apply static window partitions and standard reshape scanning methods, leading to significant losses in local connectivity. To overcome these limitations, we introduce EventMamba--a specialized model designed for EBVR tasks. EventMamba innovates by incorporating random window offset (RWO) in the spatial domain, moving away from the restrictive fixed partitioning. Additionally, it features a new consistent traversal serialization approach in the spatio-temporal domain, which maintains the proximity of adjacent events both spatially and temporally. These enhancements enable EventMamba to retain Mamba's robust modeling capabilities while significantly preserving the spatio-temporal locality of event data. Comprehensive testing on multiple datasets shows that EventMamba markedly enhances video reconstruction, drastically improving computation speed while delivering superior visual quality compared to Transformer-based methods.
♻ ☆ AnomalyNCD: Towards Novel Anomaly Class Discovery in Industrial Scenarios CVPR2025
Recently, multi-class anomaly classification has garnered increasing attention. Previous methods directly cluster anomalies but often struggle due to the lack of anomaly-prior knowledge. Acquiring this knowledge faces two issues: the non-prominent and weak-semantics anomalies. In this paper, we propose AnomalyNCD, a multi-class anomaly classification network compatible with different anomaly detection methods. To address the non-prominence of anomalies, we design main element binarization (MEBin) to obtain anomaly-centered images, ensuring anomalies are learned while avoiding the impact of incorrect detections. Next, to learn anomalies with weak semantics, we design mask-guided representation learning, which focuses on isolated anomalies guided by masks and reduces confusion from erroneous inputs through corrected pseudo labels. Finally, to enable flexible classification at both region and image levels, we develop a region merging strategy that determines the overall image category based on the classified anomaly regions. Our method outperforms the state-of-the-art works on the MVTec AD and MTD datasets. Compared with the current methods, AnomalyNCD combined with zero-shot anomaly detection method achieves a 10.8% $F_1$ gain, 8.8% NMI gain, and 9.5% ARI gain on MVTec AD, and 12.8% $F_1$ gain, 5.7% NMI gain, and 10.8% ARI gain on MTD. Code is available at https://github.com/HUST-SLOW/AnomalyNCD.
comment: Accepted at CVPR2025
♻ ☆ Training-free Diffusion Acceleration with Bottleneck Sampling
Diffusion models have demonstrated remarkable capabilities in visual content generation but remain challenging to deploy due to their high computational cost during inference. This computational burden primarily arises from the quadratic complexity of self-attention with respect to image or video resolution. While existing acceleration methods often compromise output quality or necessitate costly retraining, we observe that most diffusion models are pre-trained at lower resolutions, presenting an opportunity to exploit these low-resolution priors for more efficient inference without degrading performance. In this work, we introduce Bottleneck Sampling, a training-free framework that leverages low-resolution priors to reduce computational overhead while preserving output fidelity. Bottleneck Sampling follows a high-low-high denoising workflow: it performs high-resolution denoising in the initial and final stages while operating at lower resolutions in intermediate steps. To mitigate aliasing and blurring artifacts, we further refine the resolution transition points and adaptively shift the denoising timesteps at each stage. We evaluate Bottleneck Sampling on both image and video generation tasks, where extensive experiments demonstrate that it accelerates inference by up to 3$\times$ for image generation and 2.5$\times$ for video generation, all while maintaining output quality comparable to the standard full-resolution sampling process across multiple evaluation metrics.
comment: Project Page: https://tyfeld.github.io/BottleneckSampling.github.io/
♻ ☆ MAR-3D: Progressive Masked Auto-regressor for High-Resolution 3D Generation CVPR 2025
Recent advances in auto-regressive transformers have revolutionized generative modeling across different domains, from language processing to visual generation, demonstrating remarkable capabilities. However, applying these advances to 3D generation presents three key challenges: the unordered nature of 3D data conflicts with sequential next-token prediction paradigm, conventional vector quantization approaches incur substantial compression loss when applied to 3D meshes, and the lack of efficient scaling strategies for higher resolution latent prediction. To address these challenges, we introduce MAR-3D, which integrates a pyramid variational autoencoder with a cascaded masked auto-regressive transformer (Cascaded MAR) for progressive latent upscaling in the continuous space. Our architecture employs random masking during training and auto-regressive denoising in random order during inference, naturally accommodating the unordered property of 3D latent tokens. Additionally, we propose a cascaded training strategy with condition augmentation that enables efficiently up-scale the latent token resolution with fast convergence. Extensive experiments demonstrate that MAR-3D not only achieves superior performance and generalization capabilities compared to existing methods but also exhibits enhanced scaling capabilities compared to joint distribution modeling approaches (e.g., diffusion transformers).
comment: Accepted to CVPR 2025
♻ ☆ Volumetric Surfaces: Representing Fuzzy Geometries with Layered Meshes
High-quality view synthesis relies on volume rendering, splatting, or surface rendering. While surface rendering is typically the fastest, it struggles to accurately model fuzzy geometry like hair. In turn, alpha-blending techniques excel at representing fuzzy materials but require an unbounded number of samples per ray (P1). Further overheads are induced by empty space skipping in volume rendering (P2) and sorting input primitives in splatting (P3). We present a novel representation for real-time view synthesis where the (P1) number of sampling locations is small and bounded, (P2) sampling locations are efficiently found via rasterization, and (P3) rendering is sorting-free. We achieve this by representing objects as semi-transparent multi-layer meshes rendered in a fixed order. First, we model surface layers as signed distance function (SDF) shells with optimal spacing learned during training. Then, we bake them as meshes and fit UV textures. Unlike single-surface methods, our multi-layer representation effectively models fuzzy objects. In contrast to volume and splatting-based methods, our approach enables real-time rendering on low-power laptops and smartphones.
♻ ☆ Video Motion Transfer with Diffusion Transformers CVPR 2025
We propose DiTFlow, a method for transferring the motion of a reference video to a newly synthesized one, designed specifically for Diffusion Transformers (DiT). We first process the reference video with a pre-trained DiT to analyze cross-frame attention maps and extract a patch-wise motion signal called the Attention Motion Flow (AMF). We guide the latent denoising process in an optimization-based, training-free, manner by optimizing latents with our AMF loss to generate videos reproducing the motion of the reference one. We also apply our optimization strategy to transformer positional embeddings, granting us a boost in zero-shot motion transfer capabilities. We evaluate DiTFlow against recently published methods, outperforming all across multiple metrics and human evaluation.
comment: CVPR 2025 - Project page: https://ditflow.github.io/
♻ ☆ LANTERN++: Enhancing Relaxed Speculative Decoding with Static Tree Drafting for Visual Auto-regressive Models ICLR 2025
Speculative decoding has been widely used to accelerate auto-regressive (AR) text generation. However, its effectiveness for visual AR models remains limited due to token selection ambiguity, where multiple tokens share similarly low probabilities and thus reduce acceptance rates. Recently, relaxed speculative decoding with dynamic tree drafting was proposed to mitigate this ambiguity, demonstrating promising results in accelerating visual AR models. However, we observe that token selection ambiguity still negatively affects dynamic tree drafting, resulting in shallow draft trees and limited acceleration. To overcome this issue, we introduce LANTERN++, a refined framework that integrates static tree drafting with a tailored relaxed acceptance condition, allowing drafts to be selected independently of low-confidence predictions. This enables the acceptance of deeper sequences, improving decoding efficiency while preserving image quality. Extensive experiments on state-of-the-art visual AR models demonstrate that LANTERN++ significantly accelerates inference, achieving up to $\mathbf{\times 2.56}$ speedup over standard AR decoding while maintaining high image quality. The code is publicly available at https://github.com/jadohu/LANTERN.
comment: ICLR 2025 Workshop at SCOPE (Oral), 16 pages, 5 figures, short paper (6 pages exclude reference and appendix)
♻ ☆ Rethinking Video Tokenization: A Conditioned Diffusion-based Approach
Existing video tokenizers typically use the traditional Variational Autoencoder (VAE) architecture for video compression and reconstruction. However, to achieve good performance, its training process often relies on complex multi-stage training tricks that go beyond basic reconstruction loss and KL regularization. Among these tricks, the most challenging is the precise tuning of adversarial training with additional Generative Adversarial Networks (GANs) in the final stage, which can hinder stable convergence. In contrast to GANs, diffusion models offer more stable training processes and can generate higher-quality results. Inspired by these advantages, we propose CDT, a novel Conditioned Diffusion-based video Tokenizer, that replaces the GAN-based decoder with a conditional causal diffusion model. The encoder compresses spatio-temporal information into compact latents, while the decoder reconstructs videos through a reverse diffusion process conditioned on these latents. During inference, we incorporate a feature cache mechanism to generate videos of arbitrary length while maintaining temporal continuity and adopt sampling acceleration technique to enhance efficiency. Trained using only a basic MSE diffusion loss for reconstruction, along with KL term and LPIPS perceptual loss from scratch, extensive experiments demonstrate that CDT achieves state-of-the-art performance in video reconstruction tasks with just a single-step sampling. Even a scaled-down version of CDT (3$\times$ inference speedup) still performs comparably with top baselines. Moreover, the latent video generation model trained with CDT also exhibits superior performance. The source code and pretrained weights are available at https://github.com/ali-vilab/CDT.
♻ ☆ Multimodal Feature-Driven Deep Learning for the Prediction of Duck Body Dimensions and Weight
Accurate body dimension and weight measurements are critical for optimizing poultry management, health assessment, and economic efficiency. This study introduces an innovative deep learning-based model leveraging multimodal data-2D RGB images from different views, depth images, and 3D point clouds-for the non-invasive estimation of duck body dimensions and weight. A dataset of 1,023 Linwu ducks, comprising over 5,000 samples with diverse postures and conditions, was collected to support model training. The proposed method innovatively employs PointNet++ to extract key feature points from point clouds, extracts and computes corresponding 3D geometric features, and fuses them with multi-view convolutional 2D features. A Transformer encoder is then utilized to capture long-range dependencies and refine feature interactions, thereby enhancing prediction robustness. The model achieved a mean absolute percentage error (MAPE) of 6.33% and an R2 of 0.953 across eight morphometric parameters, demonstrating strong predictive capability. Unlike conventional manual measurements, the proposed model enables high-precision estimation while eliminating the necessity for physical handling, thereby reducing animal stress and broadening its application scope. This study marks the first application of deep learning techniques to poultry body dimension and weight estimation, providing a valuable reference for the intelligent and precise management of the livestock industry with far-reaching practical significance.
♻ ☆ Temporal-Guided Spiking Neural Networks for Event-Based Human Action Recognition
This paper explores the promising interplay between spiking neural networks (SNNs) and event-based cameras for privacy-preserving human action recognition (HAR). The unique feature of event cameras in capturing only the outlines of motion, combined with SNNs' proficiency in processing spatiotemporal data through spikes, establishes a highly synergistic compatibility for event-based HAR. Previous studies, however, have been limited by SNNs' ability to process long-term temporal information, essential for precise HAR. In this paper, we introduce two novel frameworks to address this: temporal segment-based SNN (\textit{TS-SNN}) and 3D convolutional SNN (\textit{3D-SNN}). The \textit{TS-SNN} extracts long-term temporal information by dividing actions into shorter segments, while the \textit{3D-SNN} replaces 2D spatial elements with 3D components to facilitate the transmission of temporal information. To promote further research in event-based HAR, we create a dataset, \textit{FallingDetection-CeleX}, collected using the high-resolution CeleX-V event camera $(1280 \times 800)$, comprising 7 distinct actions. Extensive experimental results show that our proposed frameworks surpass state-of-the-art SNN methods on our newly collected dataset and three other neuromorphic datasets, showcasing their effectiveness in handling long-range temporal information for event-based HAR.
♻ ☆ FaceID-6M: A Large-Scale, Open-Source FaceID Customization Dataset
Due to the data-driven nature of current face identity (FaceID) customization methods, all state-of-the-art models rely on large-scale datasets containing millions of high-quality text-image pairs for training. However, none of these datasets are publicly available, which restricts transparency and hinders further advancements in the field. To address this issue, in this paper, we collect and release FaceID-6M, the first large-scale, open-source FaceID dataset containing 6 million high-quality text-image pairs. Filtered from LAION-5B \cite{schuhmann2022laion}, FaceID-6M undergoes a rigorous image and text filtering steps to ensure dataset quality, including resolution filtering to maintain high-quality images and faces, face filtering to remove images that lack human faces, and keyword-based strategy to retain descriptions containing human-related terms (e.g., nationality, professions and names). Through these cleaning processes, FaceID-6M provides a high-quality dataset optimized for training powerful FaceID customization models, facilitating advancements in the field by offering an open resource for research and development. We conduct extensive experiments to show the effectiveness of our FaceID-6M, demonstrating that models trained on our FaceID-6M dataset achieve performance that is comparable to, and slightly better than currently available industrial models. Additionally, to support and advance research in the FaceID customization community, we make our code, datasets, and models fully publicly available. Our codes, models, and datasets are available at: https://github.com/ShuheSH/FaceID-6M.
comment: arXiv admin note: text overlap with arXiv:2501.15407
♻ ☆ EnvGS: Modeling View-Dependent Appearance with Environment Gaussian
Reconstructing complex reflections in real-world scenes from 2D images is essential for achieving photorealistic novel view synthesis. Existing methods that utilize environment maps to model reflections from distant lighting often struggle with high-frequency reflection details and fail to account for near-field reflections. In this work, we introduce EnvGS, a novel approach that employs a set of Gaussian primitives as an explicit 3D representation for capturing reflections of environments. These environment Gaussian primitives are incorporated with base Gaussian primitives to model the appearance of the whole scene. To efficiently render these environment Gaussian primitives, we developed a ray-tracing-based renderer that leverages the GPU's RT core for fast rendering. This allows us to jointly optimize our model for high-quality reconstruction while maintaining real-time rendering speeds. Results from multiple real-world and synthetic datasets demonstrate that our method produces significantly more detailed reflections, achieving the best rendering quality in real-time novel view synthesis. The code is available at https://zju3dv.github.io/envgs.
comment: Project page: https://zju3dv.github.io/envgs
♻ ☆ Video-3D LLM: Learning Position-Aware Video Representation for 3D Scene Understanding CVPR 2025
The rapid advancement of Multimodal Large Language Models (MLLMs) has significantly impacted various multimodal tasks. However, these models face challenges in tasks that require spatial understanding within 3D environments. Efforts to enhance MLLMs, such as incorporating point cloud features, have been made, yet a considerable gap remains between the models' learned representations and the inherent complexity of 3D scenes. This discrepancy largely stems from the training of MLLMs on predominantly 2D data, which restricts their effectiveness in comprehending 3D spaces. To address this issue, in this paper, we propose a novel generalist model, i.e., Video-3D LLM, for 3D scene understanding. By treating 3D scenes as dynamic videos and incorporating 3D position encoding into these representations, our Video-3D LLM aligns video representations with real-world spatial contexts more accurately. In addition, we have implemented a maximum coverage sampling technique to optimize the trade-off between computational cost and performance. Extensive experiments demonstrate that our model achieves state-of-the-art performance on several 3D scene understanding benchmarks, including ScanRefer, Multi3DRefer, Scan2Cap, ScanQA, and SQA3D.
comment: Accepted by CVPR 2025
♻ ☆ VIRES: Video Instance Repainting via Sketch and Text Guided Generation
We introduce VIRES, a video instance repainting method with sketch and text guidance, enabling video instance repainting, replacement, generation, and removal. Existing approaches struggle with temporal consistency and accurate alignment with the provided sketch sequence. VIRES leverages the generative priors of text-to-video models to maintain temporal consistency and produce visually pleasing results. We propose the Sequential ControlNet with the standardized self-scaling, which effectively extracts structure layouts and adaptively captures high-contrast sketch details. We further augment the diffusion transformer backbone with the sketch attention to interpret and inject fine-grained sketch semantics. A sketch-aware encoder ensures that repainted results are aligned with the provided sketch sequence. Additionally, we contribute the VireSet, a dataset with detailed annotations tailored for training and evaluating video instance editing methods. Experimental results demonstrate the effectiveness of VIRES, which outperforms state-of-the-art methods in visual quality, temporal consistency, condition alignment, and human ratings. Project page: https://hjzheng.net/projects/VIRES/
♻ ☆ RGB-Th-Bench: A Dense benchmark for Visual-Thermal Understanding of Vision Language Models
We introduce RGB-Th-Bench, the first benchmark designed to evaluate the ability of Vision-Language Models (VLMs) to comprehend RGB-Thermal image pairs. While VLMs have demonstrated remarkable progress in visual reasoning and multimodal understanding, their evaluation has been predominantly limited to RGB-based benchmarks, leaving a critical gap in assessing their capabilities in infrared vision tasks. Existing visible-infrared datasets are either task-specific or lack high-quality annotations necessary for rigorous model evaluation. To address these limitations, RGB-Th-Bench provides a comprehensive evaluation framework covering 14 distinct skill dimensions, with a total of 1,600+ expert-annotated Yes/No questions. The benchmark employs two accuracy metrics: a standard question-level accuracy and a stricter skill-level accuracy, which evaluates model robustness across multiple questions within each skill dimension. This design ensures a thorough assessment of model performance, including resilience to adversarial and hallucinated responses. We conduct extensive evaluations on 19 state-of-the-art VLMs, revealing significant performance gaps in RGB-Thermal understanding. Our results show that even the strongest models struggle with thermal image comprehension, with performance heavily constrained by their RGB-based capabilities. Additionally, the lack of large-scale application-specific and expert-annotated thermal-caption-pair datasets in pre-training is an important reason of the observed performance gap. RGB-Th-Bench highlights the urgent need for further advancements in multimodal learning to bridge the gap between visible and thermal image understanding. The dataset is available through this link, and the evaluation code will also be made publicly available.
♻ ☆ Beyond [cls]: Exploring the true potential of Masked Image Modeling representations
Masked Image Modeling (MIM) has emerged as a promising approach for Self-Supervised Learning (SSL) of visual representations. However, the out-of-the-box performance of MIMs is typically inferior to competing approaches. Most users cannot afford fine-tuning due to the need for large amounts of data, high GPU consumption, and specialized user knowledge. Therefore, the practical use of MIM representations is limited. In this paper we ask what is the reason for the poor out-of-the-box performance of MIMs. Is it due to weaker features produced by MIM models, or is it due to suboptimal usage? Through detailed analysis, we show that attention in MIMs is spread almost uniformly over many patches, leading to ineffective aggregation by the [cls] token. Based on this insight, we propose Selective Aggregation to better capture the rich semantic information retained in patch tokens, which significantly improves the out-of-the-box performance of MIM.
♻ ☆ Structure Modeling Activation Free Fourier Network for Spacecraft Image Denoising
Spacecraft image denoising is a crucial fundamental technology closely related to aerospace research. However, the existing deep learning-based image denoising methods are primarily designed for natural image and fail to adequately consider the characteristics of spacecraft image(e.g. low-light conditions, repetitive periodic structures), resulting in suboptimal performance in the spacecraft image denoising task. To address the aforementioned problems, we propose a Structure modeling Activation Free Fourier Network (SAFFN), which is an efficient spacecraft image denoising method including Structure Modeling Block (SMB) and Activation Free Fourier Block (AFFB). We present SMB to effectively extract edge information and model the structure for better identification of spacecraft components from dark regions in spacecraft noise image. We present AFFB and utilize an improved Fast Fourier block to extract repetitive periodic features and long-range information in noisy spacecraft image. Extensive experimental results demonstrate that our SAFFN performs competitively compared to the state-of-the-art methods on spacecraft noise image datasets. The codes are available at: https://github.com/shenduke/SAFFN.
♻ ☆ ReCap: Better Gaussian Relighting with Cross-Environment Captures
Accurate 3D objects relighting in diverse unseen environments is crucial for realistic virtual object placement. Due to the albedo-lighting ambiguity, existing methods often fall short in producing faithful relights. Without proper constraints, observed training views can be explained by numerous combinations of lighting and material attributes, lacking physical correspondence with the actual environment maps used for relighting. In this work, we present ReCap, treating cross-environment captures as multi-task target to provide the missing supervision that cuts through the entanglement. Specifically, ReCap jointly optimizes multiple lighting representations that share a common set of material attributes. This naturally harmonizes a coherent set of lighting representations around the mutual material attributes, exploiting commonalities and differences across varied object appearances. Such coherence enables physically sound lighting reconstruction and robust material estimation - both essential for accurate relighting. Together with a streamlined shading function and effective post-processing, ReCap outperforms all leading competitors on an expanded relighting benchmark.
♻ ☆ Mapping fMRI Signal and Image Stimuli in an Artificial Neural Network Latent Space: Bringing Artificial and Natural Minds Together
The goal of this study is to investigate whether latent space representations of visual stimuli and fMRI data share common information. Decoding and reconstructing stimuli from fMRI data remains a challenge in AI and neuroscience, with significant implications for understanding neural representations and improving the interpretability of Artificial Neural Networks (ANNs). In this preliminary study, we investigate the feasibility of such reconstruction by examining the similarity between the latent spaces of one autoencoder (AE) and one vision transformer (ViT) trained on fMRI and image data, respectively. Using representational similarity analysis (RSA), we found that the latent spaces of the two domains appear different. However, these initial findings are inconclusive, and further research is needed to explore this relationship more thoroughly.
comment: 4 pages, 3 figures
♻ ☆ What Do You See? Enhancing Zero-Shot Image Classification with Multimodal Large Language Models
Large language models (LLMs) have been effectively used for many computer vision tasks, including image classification. In this paper, we present a simple yet effective approach for zero-shot image classification using multimodal LLMs. Using multimodal LLMs, we generate comprehensive textual representations from input images. These textual representations are then utilized to generate fixed-dimensional features in a cross-modal embedding space. Subsequently, these features are fused together to perform zero-shot classification using a linear classifier. Our method does not require prompt engineering for each dataset; instead, we use a single, straightforward set of prompts across all datasets. We evaluated our method on several datasets and our results demonstrate its remarkable effectiveness, surpassing benchmark accuracy on multiple datasets. On average, for ten benchmarks, our method achieved an accuracy gain of 6.2 percentage points, with an increase of 6.8 percentage points on the ImageNet dataset, compared to prior methods re-evaluated with the same setup. Our findings highlight the potential of multimodal LLMs to enhance computer vision tasks such as zero-shot image classification, offering a significant improvement over traditional methods.
♻ ☆ LongViTU: Instruction Tuning for Long-Form Video Understanding
This paper introduces LongViTU, a large-scale (~121k QA pairs, ~900h videos), automatically generated dataset for long-form video understanding. We propose a systematic approach that organizes videos into a hierarchical tree structure for QA generation and incorporates self-revision mechanisms to ensure high-quality QA pairs. Each QA pair in LongViTU features: 1) long-term context (average certificate length of 4.6 minutes); 2) rich knowledge and condensed reasoning (commonsense, causality, planning, etc.)). We also offer explicit timestamp annotations of relevant events for each QA pair. We have conducted extensive human studies on LongViTU, and the results prove the quality of our dataset. To better evaluate the challenges posed by LongViTU's emphasis on long-term context and condensed reasoning, we manually curate a subset of LongViTU into a benchmark. Evaluations using a state-of-the-art open-source model (LongVU), a proprietary model (Gemini-1.5-Pro), and human annotators yield GPT-4 scores of 49.9, 52.3, and 81.0, respectively, underscoring the substantial difficulty presented by LongViTU questions. Performing supervised fine-tuning (SFT) of LongVU and LLaVA-Video on LongViTU data results in average performance gains of 2.5% and 3.7%, respectively, across a suite of long video understanding benchmarks (EgoSchema, VideoMME-Long, MLVU, LVBench).
♻ ☆ Generalizable Prompt Learning of CLIP: A Brief Overview
Existing vision-language models (VLMs) such as CLIP have showcased an impressive capability to generalize well across various downstream tasks. These models leverage the synergy between visual and textual information, enabling them to understand and reason about the content present in images and text in a unified manner. This article provides a brief overview of CLIP based on few-shot prompt learning, including experimental data and technical characteristics of some methods. The purpose of this review is to provide a reference for researchers who have just started their research in generalizable prompting of CLIP through few-shot training for classification across 15 datasets and also to facilitate the integration of this field by researchers in other downstream tasks.
♻ ☆ Leveraging Textual Anatomical Knowledge for Class-Imbalanced Semi-Supervised Multi-Organ Segmentation
Annotating 3D medical images demands substantial time and expertise, driving the adoption of semi-supervised learning (SSL) for segmentation tasks. However, the complex anatomical structures of organs often lead to significant class imbalances, posing major challenges for deploying SSL in real-world scenarios. Despite the availability of valuable prior information, such as inter-organ relative positions and organ shape priors, existing SSL methods have yet to fully leverage these insights. To address this gap, we propose a novel approach that integrates textual anatomical knowledge (TAK) into the segmentation model. Specifically, we use GPT-4o to generate textual descriptions of anatomical priors, which are then encoded using a CLIP-based model. These encoded priors are injected into the segmentation model as parameters of the segmentation head. Additionally, contrastive learning is employed to enhance the alignment between textual priors and visual features. Extensive experiments demonstrate the superior performance of our method, significantly surpassing state-of-the-art approaches. The source code will be available at: https://github.com/Lunn88/TAK-Semi.
comment: 11 pages
♻ ☆ OmniSplat: Taming Feed-Forward 3D Gaussian Splatting for Omnidirectional Images with Editable Capabilities
Feed-forward 3D Gaussian splatting (3DGS) models have gained significant popularity due to their ability to generate scenes immediately without needing per-scene optimization. Although omnidirectional images are becoming more popular since they reduce the computation required for image stitching to composite a holistic scene, existing feed-forward models are only designed for perspective images. The unique optical properties of omnidirectional images make it difficult for feature encoders to correctly understand the context of the image and make the Gaussian non-uniform in space, which hinders the image quality synthesized from novel views. We propose OmniSplat, a training-free fast feed-forward 3DGS generation framework for omnidirectional images. We adopt a Yin-Yang grid and decompose images based on it to reduce the domain gap between omnidirectional and perspective images. The Yin-Yang grid can use the existing CNN structure as it is, but its quasi-uniform characteristic allows the decomposed image to be similar to a perspective image, so it can exploit the strong prior knowledge of the learned feed-forward network. OmniSplat demonstrates higher reconstruction accuracy than existing feed-forward networks trained on perspective images. Our project page is available on: https://robot0321.github.io/omnisplat/index.html.
♻ ☆ MAP-based Problem-Agnostic diffusion model for Inverse Problems
Diffusion models have indeed shown great promise in solving inverse problems in image processing. In this paper, we propose a novel, problem-agnostic diffusion model called the maximum a posteriori (MAP)-based guided term estimation method for inverse problems. To leverage unconditionally pretrained diffusion models to address conditional generation tasks, we divide the conditional score function into two terms according to Bayes' rule: an unconditional score function (approximated by a pretrained score network) and a guided term, which is estimated using a novel MAP-based method that incorporates a Gaussian-type prior of natural images. This innovation allows us to better capture the intrinsic properties of the data, leading to improved performance. Numerical results demonstrate that our method preserves contents more effectively compared to state-of-the-art methods--for example, maintaining the structure of glasses in super-resolution tasks and producing more coherent results in the neighborhood of masked regions during inpainting.
comment: 17 pages, 10 figures
♻ ☆ Self-Expansion of Pre-trained Models with Mixture of Adapters for Continual Learning
Continual learning (CL) aims to continually accumulate knowledge from a non-stationary data stream without catastrophic forgetting of learned knowledge, requiring a balance between stability and adaptability. Relying on the generalizable representation in pre-trained models (PTMs), PTM-based CL methods perform effective continual adaptation on downstream tasks by adding learnable adapters or prompts upon the frozen PTMs. However, many existing PTM-based CL methods use restricted adaptation on a fixed set of these modules to avoid forgetting, suffering from limited CL ability. Periodically adding task-specific modules results in linear model growth rate and impaired knowledge reuse. We propose Self-Expansion of pre-trained models with Modularized Adaptation (SEMA), a novel approach to enhance the control of stability-plasticity balance in PTM-based CL. SEMA automatically decides to reuse or add adapter modules on demand in CL, depending on whether significant distribution shift that cannot be handled is detected at different representation levels. We design modular adapter consisting of a functional adapter and a representation descriptor. The representation descriptors are trained as a distribution shift indicator and used to trigger self-expansion signals. For better composing the adapters, an expandable weighting router is learned jointly for mixture of adapter outputs. SEMA enables better knowledge reuse and sub-linear expansion rate. Extensive experiments demonstrate the effectiveness of the proposed self-expansion method, achieving state-of-the-art performance compared to PTM-based CL methods without memory rehearsal. Code is available at https://github.com/huiyiwang01/SEMA-CL.
comment: Code available at https: https://github.com/huiyiwang01/SEMA-CL
♻ ☆ SimROD: A Simple Baseline for Raw Object Detection with Global and Local Enhancements
Most visual models are designed for sRGB images, yet RAW data offers significant advantages for object detection by preserving sensor information before ISP processing. This enables improved detection accuracy and more efficient hardware designs by bypassing the ISP. However, RAW object detection is challenging due to limited training data, unbalanced pixel distributions, and sensor noise. To address this, we propose SimROD, a lightweight and effective approach for RAW object detection. We introduce a Global Gamma Enhancement (GGE) module, which applies a learnable global gamma transformation with only four parameters, improving feature representation while keeping the model efficient. Additionally, we leverage the green channel's richer signal to enhance local details, aligning with the human eye's sensitivity and Bayer filter design. Extensive experiments on multiple RAW object detection datasets and detectors demonstrate that SimROD outperforms state-of-the-art methods like RAW-Adapter and DIAP while maintaining efficiency. Our work highlights the potential of RAW data for real-world object detection. Code is available at https://ocean146.github.io/SimROD2025/.
comment: Code is available at https://ocean146.github.io/SimROD2025/
♻ ☆ ATM: Improving Model Merging by Alternating Tuning and Merging
Model merging has recently emerged as a cost-efficient paradigm for multi-task learning. Among current approaches, task arithmetic stands out for its simplicity and effectiveness. In this paper, we motivate the effectiveness of task vectors by linking them to multi-task gradients. We show that in a single-epoch scenario, if the optimization is performed via gradient descent, task vectors are after one step mathematically equivalent to the gradients obtained via gradient descent in a multi-task setting, and still approximate these gradients in subsequent epochs. Furthermore, we show that the effectiveness of task vectors is largely driven by the first epoch's gradient. Given this parallel between task vectors and gradients, we propose viewing model merging as a single step in an iterative process that alternates between tuning and merging (ATM). We then propose two ways to utilize ATM. The first is to replace multi-task learning with ATM in scenarios where data sharing is prohibited, such as federated learning. The second is to improve the outcome of any model merging algorithm by applying a few post-hoc iterations of ATM on a small validation dataset, which is commonly available for hyperparameter tuning. Finally, we provide both empirical and theoretical support for the effectiveness of ATM, demonstrating that it minimizes an upper bound on the loss obtained by jointly finetuning all tasks.
comment: Main paper: 9 Pages, 9 figures, 1 table
♻ ☆ GMTalker: Gaussian Mixture-based Audio-Driven Emotional Talking Video Portraits
Synthesizing high-fidelity and emotion-controllable talking video portraits, with audio-lip sync, vivid expressions, realistic head poses, and eye blinks, has been an important and challenging task in recent years. Most existing methods suffer in achieving personalized and precise emotion control, smooth transitions between different emotion states, and the generation of diverse motions. To tackle these challenges, we present GMTalker, a Gaussian mixture-based emotional talking portraits generation framework. Specifically, we propose a Gaussian mixture-based expression generator that can construct a continuous and disentangled latent space, achieving more flexible emotion manipulation. Furthermore, we introduce a normalizing flow-based motion generator pretrained on a large dataset with a wide-range motion to generate diverse head poses, blinks, and eyeball movements. Finally, we propose a personalized emotion-guided head generator with an emotion mapping network that can synthesize high-fidelity and faithful emotional video portraits. Both quantitative and qualitative experiments demonstrate our method outperforms previous methods in image quality, photo-realism, emotion accuracy, and motion diversity.
comment: Project page: https://bob35buaa.github.io/GMTalker. This work has been submitted to the IEEE journal for possible publication
♻ ☆ Feedback-driven object detection and iterative model improvement
Automated object detection has become increasingly valuable across diverse applications, yet efficient, high-quality annotation remains a persistent challenge. In this paper, we present the development and evaluation of a platform designed to interactively improve object detection models. The platform allows uploading and annotating images as well as fine-tuning object detection models. Users can then manually review and refine annotations, further creating improved snapshots that are used for automatic object detection on subsequent image uploads - a process we refer to as semi-automatic annotation resulting in a significant gain in annotation efficiency. Whereas iterative refinement of model results to speed up annotation has become common practice, we are the first to quantitatively evaluate its benefits with respect to time, effort, and interaction savings. Our experimental results show clear evidence for a significant time reduction of up to 53% for semi-automatic compared to manual annotation. Importantly, these efficiency gains did not compromise annotation quality, while matching or occasionally even exceeding the accuracy of manual annotations. These findings demonstrate the potential of our lightweight annotation platform for creating high-quality object detection datasets and provide best practices to guide future development of annotation platforms. The platform is open-source, with the frontend and backend repositories available on GitHub. To support the understanding of our labeling process, we have created an explanatory video demonstrating the methodology using microscopy images of E. coli bacteria as an example.
comment: Code: https://github.com/ml-lab-htw/iterative-annotate Video: https://www.youtube.com/watch?v=CM9uhE8NN5E
♻ ☆ Not Just Object, But State: Compositional Incremental Learning without Forgetting NeurIPS 2024
Most incremental learners excessively prioritize coarse classes of objects while neglecting various kinds of states (e.g. color and material) attached to the objects. As a result, they are limited in the ability to reason fine-grained compositionality of state-object pairs. To remedy this limitation, we propose a novel task called Compositional Incremental Learning (composition-IL), enabling the model to recognize state-object compositions as a whole in an incremental learning fashion. Since the lack of suitable benchmarks, we re-organize two existing datasets and make them tailored for composition-IL. Then, we propose a prompt-based Composition Incremental Learner (CompILer), to overcome the ambiguous composition boundary problem which challenges composition-IL largely. Specifically, we exploit multi-pool prompt learning, which is regularized by inter-pool prompt discrepancy and intra-pool prompt diversity. Besides, we devise object-injected state prompting by using object prompts to guide the selection of state prompts. Furthermore, we fuse the selected prompts by a generalized-mean strategy, to eliminate irrelevant information learned in the prompts. Extensive experiments on two datasets exhibit state-of-the-art performance achieved by CompILer.
comment: NeurIPS 2024
♻ ☆ Event-boosted Deformable 3D Gaussians for Dynamic Scene Reconstruction
Deformable 3D Gaussian Splatting (3D-GS) is limited by missing intermediate motion information due to the low temporal resolution of RGB cameras. To address this, we introduce the first approach combining event cameras, which capture high-temporal-resolution, continuous motion data, with deformable 3D-GS for dynamic scene reconstruction. We observe that threshold modeling for events plays a crucial role in achieving high-quality reconstruction. Therefore, we propose a GS-Threshold Joint Modeling strategy, creating a mutually reinforcing process that greatly improves both 3D reconstruction and threshold modeling. Moreover, we introduce a Dynamic-Static Decomposition strategy that first identifies dynamic areas by exploiting the inability of static Gaussians to represent motions, then applies a buffer-based soft decomposition to separate dynamic and static areas. This strategy accelerates rendering by avoiding unnecessary deformation in static areas, and focuses on dynamic areas to enhance fidelity. Additionally, we contribute the first event-inclusive 4D benchmark with synthetic and real-world dynamic scenes, on which our method achieves state-of-the-art performance.
♻ ☆ MESA: Effective Matching Redundancy Reduction by Semantic Area Segmentation
We propose MESA and DMESA as novel feature matching methods, which utilize Segment Anything Model (SAM) to effectively mitigate matching redundancy. The key insight of our methods is to establish implicit-semantic area matching prior to point matching, based on advanced image understanding of SAM. Then, informative area matches with consistent internal semantic are able to undergo dense feature comparison, facilitating precise inside-area point matching. Specifically, MESA adopts a sparse matching framework and first obtains candidate areas from SAM results through a novel Area Graph (AG). Then, area matching among the candidates is formulated as graph energy minimization and solved by graphical models derived from AG. To address the efficiency issue of MESA, we further propose DMESA as its dense counterpart, applying a dense matching framework. After candidate areas are identified by AG, DMESA establishes area matches through generating dense matching distributions. The distributions are produced from off-the-shelf patch matching utilizing the Gaussian Mixture Model and refined via the Expectation Maximization. With less repetitive computation, DMESA showcases a speed improvement of nearly five times compared to MESA, while maintaining competitive accuracy. Our methods are extensively evaluated on five datasets encompassing indoor and outdoor scenes. The results illustrate consistent performance improvements from our methods for five distinct point matching baselines across all datasets. Furthermore, our methods exhibit promise generalization and improved robustness against image resolution variations. The code is publicly available at https://github.com/Easonyesheng/A2PM-MESA.
comment: 18pages+suppl
♻ ☆ ProReason: Multi-Modal Proactive Reasoning with Decoupled Eyesight and Wisdom
Large vision-language models (LVLMs) have witnessed significant progress on visual understanding tasks. However, they often prioritize language knowledge over image information on visual reasoning tasks, incurring performance degradation. To tackle this issue, we first identify the drawbacks of existing solutions (i.e., insufficient and irrelevant visual descriptions, and limited multi-modal capacities). We then decompose visual reasoning process into two stages: visual perception (i.e., eyesight) and textual reasoning (i.e., wisdom), and introduce a novel visual reasoning framework named ProReason. This framework features multi-run proactive perception and decoupled vision-reasoning capabilities. Briefly, given a multi-modal question, ProReason iterates proactive information collection and reasoning until the answer can be concluded with necessary and sufficient visual descriptions. Notably, the disassociation of capabilities allows seamless integration of existing large language models (LLMs) to compensate for the reasoning deficits of LVLMs. Our extensive experiments demonstrate that ProReason outperforms both existing multi-step reasoning frameworks and passive peer methods on a wide range of benchmarks for both open-source and closed-source models. In addition, with the assistance of LLMs, ProReason achieves a performance improvement of up to 15% on MMMU benchmark. Our insights into existing solutions and the decoupled perspective for feasible integration of LLMs illuminate future research on visual reasoning techniques, especially LLM-assisted ones.
♻ ☆ Image-to-Text for Medical Reports Using Adaptive Co-Attention and Triple-LSTM Module
Medical report generation requires specialized expertise that general large models often fail to accurately capture. Moreover, the inherent repetition and similarity in medical data make it difficult for models to extract meaningful features, resulting in a tendency to overfit. So in this paper, we propose a multimodal model, Co-Attention Triple-LSTM Network (CA-TriNet), a deep learning model that combines transformer architectures with a Multi-LSTM network. Its Co-Attention module synergistically links a vision transformer with a text transformer to better differentiate medical images with similarities, augmented by an adaptive weight operator to catch and amplify image labels with minor similarities. Furthermore, its Triple-LSTM module refines generated sentences using targeted image objects. Extensive evaluations over three public datasets have demonstrated that CA-TriNet outperforms state-of-the-art models in terms of comprehensive ability, even pre-trained large language models on some metrics.
♻ ☆ Recovering Dynamic 3D Sketches from Videos CVPR 2025
Understanding 3D motion from videos presents inherent challenges due to the diverse types of movement, ranging from rigid and deformable objects to articulated structures. To overcome this, we propose Liv3Stroke, a novel approach for abstracting objects in motion with deformable 3D strokes. The detailed movements of an object may be represented by unstructured motion vectors or a set of motion primitives using a pre-defined articulation from a template model. Just as a free-hand sketch can intuitively visualize scenes or intentions with a sparse set of lines, we utilize a set of parametric 3D curves to capture a set of spatially smooth motion elements for general objects with unknown structures. We first extract noisy, 3D point cloud motion guidance from video frames using semantic features, and our approach deforms a set of curves to abstract essential motion features as a set of explicit 3D representations. Such abstraction enables an understanding of prominent components of motions while maintaining robustness to environmental factors. Our approach allows direct analysis of 3D object movements from video, tackling the uncertainty that typically occurs when translating real-world motion into recorded footage. The project page is accessible via: https://jaeah.me/liv3stroke_web
comment: Accepted to CVPR 2025
♻ ☆ Flip Learning: Weakly Supervised Erase to Segment Nodules in Breast Ultrasound
Accurate segmentation of nodules in both 2D breast ultrasound (BUS) and 3D automated breast ultrasound (ABUS) is crucial for clinical diagnosis and treatment planning. Therefore, developing an automated system for nodule segmentation can enhance user independence and expedite clinical analysis. Unlike fully-supervised learning, weakly-supervised segmentation (WSS) can streamline the laborious and intricate annotation process. However, current WSS methods face challenges in achieving precise nodule segmentation, as many of them depend on inaccurate activation maps or inefficient pseudo-mask generation algorithms. In this study, we introduce a novel multi-agent reinforcement learning-based WSS framework called Flip Learning, which relies solely on 2D/3D boxes for accurate segmentation. Specifically, multiple agents are employed to erase the target from the box to facilitate classification tag flipping, with the erased region serving as the predicted segmentation mask. The key contributions of this research are as follows: (1) Adoption of a superpixel/supervoxel-based approach to encode the standardized environment, capturing boundary priors and expediting the learning process. (2) Introduction of three meticulously designed rewards, comprising a classification score reward and two intensity distribution rewards, to steer the agents' erasing process precisely, thereby avoiding both under- and over-segmentation. (3) Implementation of a progressive curriculum learning strategy to enable agents to interact with the environment in a progressively challenging manner, thereby enhancing learning efficiency. Extensively validated on the large in-house BUS and ABUS datasets, our Flip Learning method outperforms state-of-the-art WSS methods and foundation models, and achieves comparable performance as fully-supervised learning algorithms.
comment: Accepted by Medical Image Analysis. 24 pages, 13 figures, 20 tabels
♻ ☆ StyleStudio: Text-Driven Style Transfer with Selective Control of Style Elements CVPR 2025
Text-driven style transfer aims to merge the style of a reference image with content described by a text prompt. Recent advancements in text-to-image models have improved the nuance of style transformations, yet significant challenges remain, particularly with overfitting to reference styles, limiting stylistic control, and misaligning with textual content. In this paper, we propose three complementary strategies to address these issues. First, we introduce a cross-modal Adaptive Instance Normalization (AdaIN) mechanism for better integration of style and text features, enhancing alignment. Second, we develop a Style-based Classifier-Free Guidance (SCFG) approach that enables selective control over stylistic elements, reducing irrelevant influences. Finally, we incorporate a teacher model during early generation stages to stabilize spatial layouts and mitigate artifacts. Our extensive evaluations demonstrate significant improvements in style transfer quality and alignment with textual prompts. Furthermore, our approach can be integrated into existing style transfer frameworks without fine-tuning.
comment: Accepted by CVPR 2025
♻ ☆ OODFace: Benchmarking Robustness of Face Recognition under Common Corruptions and Appearance Variations
With the rise of deep learning, facial recognition technology has seen extensive research and rapid development. Although facial recognition is considered a mature technology, we find that existing open-source models and commercial algorithms lack robustness in certain complex Out-of-Distribution (OOD) scenarios, raising concerns about the reliability of these systems. In this paper, we introduce OODFace, which explores the OOD challenges faced by facial recognition models from two perspectives: common corruptions and appearance variations. We systematically design 30 OOD scenarios across 9 major categories tailored for facial recognition. By simulating these challenges on public datasets, we establish three robustness benchmarks: LFW-C/V, CFP-FP-C/V, and YTF-C/V. We then conduct extensive experiments on 19 facial recognition models and 3 commercial APIs, along with extended physical experiments on face masks to assess their robustness. Next, we explore potential solutions from two perspectives: defense strategies and Vision-Language Models (VLMs). Based on the results, we draw several key insights, highlighting the vulnerability of facial recognition systems to OOD data and suggesting possible solutions. Additionally, we offer a unified toolkit that includes all corruption and variation types, easily extendable to other datasets. We hope that our benchmarks and findings can provide guidance for future improvements in facial recognition model robustness.
♻ ☆ MouseGPT: A Large-scale Vision-Language Model for Mouse Behavior Analysis
Analyzing animal behavior is crucial in advancing neuroscience, yet quantifying and deciphering its intricate dynamics remains a significant challenge. Traditional machine vision approaches, despite their ability to detect spontaneous behaviors, fall short due to limited interpretability and reliance on manual labeling, which restricts the exploration of the full behavioral spectrum. Here, we introduce MouseGPT, a Vision-Language Model (VLM) that integrates visual cues with natural language to revolutionize mouse behavior analysis. Built upon our first-of-its-kind dataset - incorporating pose dynamics and open-vocabulary behavioral annotations across over 42 million frames of diverse psychiatric conditions - MouseGPT provides a novel, context-rich method for comprehensive behavior interpretation. Our holistic analysis framework enables detailed behavior profiling, clustering, and novel behavior discovery, offering deep insights without the need for labor - intensive manual annotation. Evaluations reveal that MouseGPT surpasses existing models in precision, adaptability, and descriptive richness, positioning it as a transformative tool for ethology and for unraveling complex behavioral dynamics in animal models.
comment: 53 pages, 5 figures, 7 extended figures
♻ ☆ DefectFill: Realistic Defect Generation with Inpainting Diffusion Model for Visual Inspection CVPR 2025
Developing effective visual inspection models remains challenging due to the scarcity of defect data. While image generation models have been used to synthesize defect images, producing highly realistic defects remains difficult. We propose DefectFill, a novel method for realistic defect generation that requires only a few reference defect images. It leverages a fine-tuned inpainting diffusion model, optimized with our custom loss functions incorporating defect, object, and attention terms. It enables precise capture of detailed, localized defect features and their seamless integration into defect-free objects. Additionally, our Low-Fidelity Selection method further enhances the defect sample quality. Experiments show that DefectFill generates high-quality defect images, enabling visual inspection models to achieve state-of-the-art performance on the MVTec AD dataset.
comment: Accepted to CVPR 2025
♻ ☆ MoReVQA: Exploring Modular Reasoning Models for Video Question Answering CVPR 2024
This paper addresses the task of video question answering (videoQA) via a decomposed multi-stage, modular reasoning framework. Previous modular methods have shown promise with a single planning stage ungrounded in visual content. However, through a simple and effective baseline, we find that such systems can lead to brittle behavior in practice for challenging videoQA settings. Thus, unlike traditional single-stage planning methods, we propose a multi-stage system consisting of an event parser, a grounding stage, and a final reasoning stage in conjunction with an external memory. All stages are training-free, and performed using few-shot prompting of large models, creating interpretable intermediate outputs at each stage. By decomposing the underlying planning and task complexity, our method, MoReVQA, improves over prior work on standard videoQA benchmarks (NExT-QA, iVQA, EgoSchema, ActivityNet-QA) with state-of-the-art results, and extensions to related tasks (grounded videoQA, paragraph captioning).
comment: CVPR 2024; updated NExT-GQA results in Appendix
♻ ☆ Frequency-Guided Diffusion Model with Perturbation Training for Skeleton-Based Video Anomaly Detection
Video anomaly detection (VAD) is a vital yet complex open-set task in computer vision, commonly tackled through reconstruction-based methods. However, these methods struggle with two key limitations: (1) insufficient robustness in open-set scenarios, where unseen normal motions are frequently misclassified as anomalies, and (2) an overemphasis on, but restricted capacity for, local motion reconstruction, which are inherently difficult to capture accurately due to their diversity. To overcome these challenges, we introduce a novel frequency-guided diffusion model with perturbation training. First, we enhance robustness by training a generator to produce perturbed samples, which are similar to normal samples and target the weakness of the reconstruction model. This training paradigm expands the reconstruction domain of the model, improving its generalization to unseen normal motions. Second, to address the overemphasis on motion details, we employ the 2D Discrete Cosine Transform (DCT) to separate high-frequency (local) and low-frequency (global) motion components. By guiding the diffusion model with observed high-frequency information, we prioritize the reconstruction of low-frequency components, enabling more accurate and robust anomaly detection. Extensive experiments on five widely used VAD datasets demonstrate that our approach surpasses state-of-the-art methods, underscoring its effectiveness in open-set scenarios and diverse motion contexts. Our project website is https://xiaofeng-tan.github.io/projects/FG-Diff/index.html.
♻ ☆ MotionDiff: Training-free Zero-shot Interactive Motion Editing via Flow-assisted Multi-view Diffusion
Generative models have made remarkable advancements and are capable of producing high-quality content. However, performing controllable editing with generative models remains challenging, due to their inherent uncertainty in outputs. This challenge is praticularly pronounced in motion editing, which involves the processing of spatial information. While some physics-based generative methods have attempted to implement motion editing, they typically operate on single-view images with simple motions, such as translation and dragging. These methods struggle to handle complex rotation and stretching motions and ensure multi-view consistency, often necessitating resource-intensive retraining. To address these challenges, we propose MotionDiff, a training-free zero-shot diffusion method that leverages optical flow for complex multi-view motion editing. Specifically, given a static scene, users can interactively select objects of interest to add motion priors. The proposed Point Kinematic Model (PKM) then estimates corresponding multi-view optical flows during the Multi-view Flow Estimation Stage (MFES). Subsequently, these optical flows are utilized to generate multi-view motion results through decoupled motion representation in the Multi-view Motion Diffusion Stage (MMDS). Extensive experiments demonstrate that MotionDiff outperforms other physics-based generative motion editing methods in achieving high-quality multi-view consistent motion results. Notably, MotionDiff does not require retraining, enabling users to conveniently adapt it for various down-stream tasks.
♻ ☆ Perceptually Accurate 3D Talking Head Generation: New Definitions, Speech-Mesh Representation, and Evaluation Metrics CVPR 2025
Recent advancements in speech-driven 3D talking head generation have made significant progress in lip synchronization. However, existing models still struggle to capture the perceptual alignment between varying speech characteristics and corresponding lip movements. In this work, we claim that three criteria -- Temporal Synchronization, Lip Readability, and Expressiveness -- are crucial for achieving perceptually accurate lip movements. Motivated by our hypothesis that a desirable representation space exists to meet these three criteria, we introduce a speech-mesh synchronized representation that captures intricate correspondences between speech signals and 3D face meshes. We found that our learned representation exhibits desirable characteristics, and we plug it into existing models as a perceptual loss to better align lip movements to the given speech. In addition, we utilize this representation as a perceptual metric and introduce two other physically grounded lip synchronization metrics to assess how well the generated 3D talking heads align with these three criteria. Experiments show that training 3D talking head generation models with our perceptual loss significantly improve all three aspects of perceptually accurate lip synchronization. Codes and datasets are available at https://perceptual-3d-talking-head.github.io/.
comment: CVPR 2025
♻ ☆ SMAFormer: Synergistic Multi-Attention Transformer for Medical Image Segmentation
In medical image segmentation, specialized computer vision techniques, notably transformers grounded in attention mechanisms and residual networks employing skip connections, have been instrumental in advancing performance. Nonetheless, previous models often falter when segmenting small, irregularly shaped tumors. To this end, we introduce SMAFormer, an efficient, Transformer-based architecture that fuses multiple attention mechanisms for enhanced segmentation of small tumors and organs. SMAFormer can capture both local and global features for medical image segmentation. The architecture comprises two pivotal components. First, a Synergistic Multi-Attention (SMA) Transformer block is proposed, which has the benefits of Pixel Attention, Channel Attention, and Spatial Attention for feature enrichment. Second, addressing the challenge of information loss incurred during attention mechanism transitions and feature fusion, we design a Feature Fusion Modulator. This module bolsters the integration between the channel and spatial attention by mitigating reshaping-induced information attrition. To evaluate our method, we conduct extensive experiments on various medical image segmentation tasks, including multi-organ, liver tumor, and bladder tumor segmentation, achieving state-of-the-art results. Code and models are available at: https://github.com/CXH-Research/SMAFormer.
comment: Accepted by IEEE BIBM 2024
♻ ☆ Survey on Monocular Metric Depth Estimation
Monocular Depth Estimation (MDE) is fundamental to computer vision, enabling spatial understanding, 3D reconstruction, and autonomous driving. Deep learning-based MDE predicts relative depth from a single image, but the lack of metric scale introduces inconsistencies, limiting applicability in tasks such as visual SLAM, 3D reconstruction, and novel view synthesis. Monocular Metric Depth Estimation (MMDE) overcomes this limitation by enabling precise scene-scale inference, improving depth consistency, enhancing stability in sequential tasks, and streamlining integration into practical systems. This paper systematically reviews the evolution of depth estimation, from traditional geometric methods to deep learning breakthroughs, emphasizing scale-agnostic approaches in zero-shot generalization which is crucial for advancing MMDE. Recent progress in zero-shot MMDE is examined, focusing on challenges such as model generalization and boundary detail loss. To address these issues, researchers have explored unlabeled data augmentation, image patching, architectural optimization, and generative techniques. This review analyzes these developments, assessing their impact and limitations. Key findings are synthesized, unresolved challenges outlined, and future research direction proposal. By providing a clear technical roadmap and insight into emerging trends, this work aims to drive innovation and expand the real-world applications of MMDE.
♻ ☆ Reason-RFT: Reinforcement Fine-Tuning for Visual Reasoning
Visual reasoning abilities play a crucial role in understanding complex multimodal data, advancing both domain-specific applications and artificial general intelligence (AGI). Existing methods improve VLM reasoning via Chain-of-Thought (CoT) supervised fine-tuning, using meticulously annotated training data to enhance visual reasoning capabilities. However, this training paradigm may lead to overfitting and cognitive rigidity, restricting the model's ability to transfer visual reasoning skills across domains and limiting its real-world applicability. To address these limitations, we propose Reason-RFT, a novel reinforcement fine-tuning framework that significantly enhances generalization capabilities in visual reasoning tasks. Reason-RFT introduces a two-phase training framework for visual reasoning: (1) Supervised Fine-Tuning (SFT) with curated Chain-of-Thought (CoT) data activates the reasoning potential of Vision-Language Models (VLMs), followed by (2) Group Relative Policy Optimization (GRPO)-based reinforcement learning that generates multiple reasoning-response pairs, significantly enhancing generalization in visual reasoning tasks. To evaluate Reason-RFT's visual reasoning capabilities, we reconstructed a comprehensive dataset spanning visual counting, structure perception, and spatial transformation. Experimental results demonstrate Reasoning-RFT's three key advantages: (1) Performance Enhancement: achieving state-of-the-art results across multiple tasks, outperforming most mainstream open-source and proprietary models; (2) Generalization Superiority: consistently maintaining robust performance across diverse tasks and domains, outperforming alternative training paradigms; (3) Data Efficiency: excelling in few-shot learning scenarios while surpassing full-dataset SFT baselines. Project website: https://tanhuajie.github.io/ReasonRFT
comment: 35 pages, 22 figures
♻ ☆ Leveraging 3D Geometric Priors in 2D Rotation Symmetry Detection CVPR 2025
Symmetry plays a vital role in understanding structural patterns, aiding object recognition and scene interpretation. This paper focuses on rotation symmetry, where objects remain unchanged when rotated around a central axis, requiring detection of rotation centers and supporting vertices. Traditional methods relied on hand-crafted feature matching, while recent segmentation models based on convolutional neural networks detect rotation centers but struggle with 3D geometric consistency due to viewpoint distortions. To overcome this, we propose a model that directly predicts rotation centers and vertices in 3D space and projects the results back to 2D while preserving structural integrity. By incorporating a vertex reconstruction stage enforcing 3D geometric priors -- such as equal side lengths and interior angles -- our model enhances robustness and accuracy. Experiments on the DENDI dataset show superior performance in rotation axis detection and validate the impact of 3D priors through ablation studies.
comment: Accepted to CVPR 2025
♻ ☆ Towards Better Alignment: Training Diffusion Models with Reinforcement Learning Against Sparse Rewards CVPR 2025
Diffusion models have achieved remarkable success in text-to-image generation. However, their practical applications are hindered by the misalignment between generated images and corresponding text prompts. To tackle this issue, reinforcement learning (RL) has been considered for diffusion model fine-tuning. Yet, RL's effectiveness is limited by the challenge of sparse reward, where feedback is only available at the end of the generation process. This makes it difficult to identify which actions during the denoising process contribute positively to the final generated image, potentially leading to ineffective or unnecessary denoising policies. To this end, this paper presents a novel RL-based framework that addresses the sparse reward problem when training diffusion models. Our framework, named $\text{B}^2\text{-DiffuRL}$, employs two strategies: \textbf{B}ackward progressive training and \textbf{B}ranch-based sampling. For one thing, backward progressive training focuses initially on the final timesteps of denoising process and gradually extends the training interval to earlier timesteps, easing the learning difficulty from sparse rewards. For another, we perform branch-based sampling for each training interval. By comparing the samples within the same branch, we can identify how much the policies of the current training interval contribute to the final image, which helps to learn effective policies instead of unnecessary ones. $\text{B}^2\text{-DiffuRL}$ is compatible with existing optimization algorithms. Extensive experiments demonstrate the effectiveness of $\text{B}^2\text{-DiffuRL}$ in improving prompt-image alignment and maintaining diversity in generated images. The code for this work is available.
comment: Accepted to CVPR 2025, add references
♻ ☆ VERA: Explainable Video Anomaly Detection via Verbalized Learning of Vision-Language Models CVPR 2025
The rapid advancement of vision-language models (VLMs) has established a new paradigm in video anomaly detection (VAD): leveraging VLMs to simultaneously detect anomalies and provide comprehendible explanations for the decisions. Existing work in this direction often assumes the complex reasoning required for VAD exceeds the capabilities of pretrained VLMs. Consequently, these approaches either incorporate specialized reasoning modules during inference or rely on instruction tuning datasets through additional training to adapt VLMs for VAD. However, such strategies often incur substantial computational costs or data annotation overhead. To address these challenges in explainable VAD, we introduce a verbalized learning framework named VERA that enables VLMs to perform VAD without model parameter modifications. Specifically, VERA automatically decomposes the complex reasoning required for VAD into reflections on simpler, more focused guiding questions capturing distinct abnormal patterns. It treats these reflective questions as learnable parameters and optimizes them through data-driven verbal interactions between learner and optimizer VLMs, using coarsely labeled training data. During inference, VERA embeds the learned questions into model prompts to guide VLMs in generating segment-level anomaly scores, which are then refined into frame-level scores via the fusion of scene and temporal contexts. Experimental results on challenging benchmarks demonstrate that the learned questions of VERA are highly adaptable, significantly improving both detection performance and explainability of VLMs for VAD.
comment: Accepted in CVPR 2025
♻ ☆ SyncDiff: Synchronized Motion Diffusion for Multi-Body Human-Object Interaction Synthesis
Synthesizing realistic human-object interaction motions is a critical problem in VR/AR and human animation. Unlike the commonly studied scenarios involving a single human or hand interacting with one object, we address a more generic multi-body setting with arbitrary numbers of humans, hands, and objects. This complexity introduces significant challenges in synchronizing motions due to the high correlations and mutual influences among bodies. To address these challenges, we introduce SyncDiff, a novel method for multi-body interaction synthesis using a synchronized motion diffusion strategy. SyncDiff employs a single diffusion model to capture the joint distribution of multi-body motions. To enhance motion fidelity, we propose a frequency-domain motion decomposition scheme. Additionally, we introduce a new set of alignment scores to emphasize the synchronization of different body motions. SyncDiff jointly optimizes both data sample likelihood and alignment likelihood through an explicit synchronization strategy. Extensive experiments across four datasets with various multi-body configurations demonstrate the superiority of SyncDiff over existing state-of-the-art motion synthesis methods.
comment: 26 pages, 10 figures
♻ ☆ Mitigating Low-Level Visual Hallucinations Requires Self-Awareness: Database, Model and Training Strategy
The rapid development of multimodal large language models has resulted in remarkable advancements in visual perception and understanding, consolidating several tasks into a single visual question-answering framework. However, these models are prone to hallucinations, which limit their reliability as artificial intelligence systems. While this issue is extensively researched in natural language processing and image captioning, there remains a lack of investigation of hallucinations in Low-level Visual Perception and Understanding (HLPU), especially in the context of image quality assessment tasks. We consider that these hallucinations arise from an absence of clear self-awareness within the models. To address this issue, we first introduce the HLPU instruction database, the first instruction database specifically focused on hallucinations in low-level vision tasks. This database contains approximately 200K question-answer pairs and comprises four subsets, each covering different types of instructions. Subsequently, we propose the Self-Awareness Failure Elimination (SAFEQA) model, which utilizes image features, salient region features and quality features to improve the perception and comprehension abilities of the model in low-level vision tasks. Furthermore, we propose the Enhancing Self-Awareness Preference Optimization (ESA-PO) framework to increase the model's awareness of knowledge boundaries, thereby mitigating the incidence of hallucination. Finally, we conduct comprehensive experiments on low-level vision tasks, with the results demonstrating that our proposed method significantly enhances self-awareness of the model in these tasks and reduces hallucinations. Notably, our proposed method improves both accuracy and self-awareness of the proposed model and outperforms close-source models in terms of various evaluation metrics.
♻ ☆ RatBodyFormer: Rat Body Surface from Keypoints
Analyzing rat behavior lies at the heart of many scientific studies. Past methods for automated rodent modeling have focused on 3D pose estimation from keypoints, e.g., face and appendages. The pose, however, does not capture the rich body surface movement encoding the subtle rat behaviors like curling and stretching. The body surface lacks features that can be visually defined, evading these established keypoint-based methods. In this paper, we introduce the first method for reconstructing the rat body surface as a dense set of points by learning to predict it from the sparse keypoints that can be detected with past methods. Our method consists of two key contributions. The first is RatDome, a novel multi-camera system for rat behavior capture, and a large-scale dataset captured with it that consists of pairs of 3D keypoints and 3D body surface points. The second is RatBodyFormer, a novel network to transform detected keypoints to 3D body surface points. RatBodyFormer is agnostic to the exact locations of the 3D body surface points in the training data and is trained with masked-learning. We experimentally validate our framework with a number of real-world experiments. Our results collectively serve as a novel foundation for automated rat behavior analysis.
comment: https://vision.ist.i.kyoto-u.ac.jp/research/ratbodyformer/
♻ ☆ Towards Complementary Knowledge Distillation for Efficient Dense Image Prediction
It has been revealed that small efficient dense image prediction (EDIP) models, trained using the knowledge distillation (KD) framework, encounter two key challenges, including maintaining boundary region completeness and preserving target region connectivity, despite their favorable capacity to recognize main object regions. In this work, we propose a complementary boundary and context distillation (BCD) method within the KD framework for EDIPs, which facilitates the targeted knowledge transfer from large accurate teacher models to compact efficient student models. Specifically, the boundary distillation component focuses on extracting explicit object-level semantic boundaries from the hierarchical feature maps of the backbone network to enhance the student model's mask quality in boundary regions. Concurrently, the context distillation component leverages self-relations as a bridge to transfer implicit pixel-level contexts from the teacher model to the student model, ensuring strong connectivity in target regions. Our proposed BCD method is specifically designed for EDIP tasks and is characterized by its simplicity and efficiency. Extensive experimental results across semantic segmentation, object detection, and instance segmentation on various representative datasets demonstrate that our method can outperform existing methods without requiring extra supervisions or incurring increased inference costs, resulting in well-defined object boundaries and smooth connecting regions.
comment: under submission
♻ ☆ SplatFlow: Self-Supervised Dynamic Gaussian Splatting in Neural Motion Flow Field for Autonomous Driving
Most existing Dynamic Gaussian Splatting methods for complex dynamic urban scenarios rely on accurate object-level supervision from expensive manual labeling, limiting their scalability in real-world applications. In this paper, we introduce SplatFlow, a Self-Supervised Dynamic Gaussian Splatting within Neural Motion Flow Fields (NMFF) to learn 4D space-time representations without requiring tracked 3D bounding boxes, enabling accurate dynamic scene reconstruction and novel view RGB/depth/flow synthesis. SplatFlow designs a unified framework to seamlessly integrate time-dependent 4D Gaussian representation within NMFF, where NMFF is a set of implicit functions to model temporal motions of both LiDAR points and Gaussians as continuous motion flow fields. Leveraging NMFF, SplatFlow effectively decomposes static background and dynamic objects, representing them with 3D and 4D Gaussian primitives, respectively. NMFF also models the correspondences of each 4D Gaussian across time, which aggregates temporal features to enhance cross-view consistency of dynamic components. SplatFlow further improves dynamic object identification by distilling features from 2D foundation models into 4D space-time representation. Comprehensive evaluations conducted on the Waymo and KITTI Datasets validate SplatFlow's state-of-the-art (SOTA) performance for both image reconstruction and novel view synthesis in dynamic urban scenarios.
♻ ☆ GCA-SUNet: A Gated Context-Aware Swin-UNet for Exemplar-Free Counting ICME 2025
Exemplar-Free Counting aims to count objects of interest without intensive annotations of objects or exemplars. To achieve this, we propose a Gated Context-Aware Swin-UNet (GCA-SUNet) to directly map an input image to the density map of countable objects. Specifically, a set of Swin transformers form an encoder to derive a robust feature representation, and a Gated Context-Aware Modulation block is designed to suppress irrelevant objects or background through a gate mechanism and exploit the attentive support of objects of interest through a self-similarity matrix. The gate strategy is also incorporated into the bottleneck network and the decoder of the Swin-UNet to highlight the features most relevant to objects of interest. By explicitly exploiting the attentive support among countable objects and eliminating irrelevant features through the gate mechanisms, the proposed GCA-SUNet focuses on and counts objects of interest without relying on predefined categories or exemplars. Experimental results on the real-world datasets such as FSC-147 and CARPK demonstrate that GCA-SUNet significantly and consistently outperforms state-of-the-art methods. The code is available at https://github.com/Amordia/GCA-SUNet.
comment: Accepted by ICME 2025
♻ ☆ VideoHandles: Editing 3D Object Compositions in Videos Using Video Generative Priors
Generative methods for image and video editing use generative models as priors to perform edits despite incomplete information, such as changing the composition of 3D objects shown in a single image. Recent methods have shown promising composition editing results in the image setting, but in the video setting, editing methods have focused on editing object's appearance and motion, or camera motion, and as a result, methods to edit object composition in videos are still missing. We propose \name as a method for editing 3D object compositions in videos of static scenes with camera motion. Our approach allows editing the 3D position of a 3D object across all frames of a video in a temporally consistent manner. This is achieved by lifting intermediate features of a generative model to a 3D reconstruction that is shared between all frames, editing the reconstruction, and projecting the features on the edited reconstruction back to each frame. To the best of our knowledge, this is the first generative approach to edit object compositions in videos. Our approach is simple and training-free, while outperforming state-of-the-art image editing baselines.
comment: Project page: https://videohandles.github.io
Machine Learning 157
☆ Test-Time Visual In-Context Tuning CVPR 2025
Visual in-context learning (VICL), as a new paradigm in computer vision, allows the model to rapidly adapt to various tasks with only a handful of prompts and examples. While effective, the existing VICL paradigm exhibits poor generalizability under distribution shifts. In this work, we propose test-time Visual In-Context Tuning (VICT), a method that can adapt VICL models on the fly with a single test sample. Specifically, we flip the role between the task prompts and the test sample and use a cycle consistency loss to reconstruct the original task prompt output. Our key insight is that a model should be aware of a new test distribution if it can successfully recover the original task prompts. Extensive experiments on six representative vision tasks ranging from high-level visual understanding to low-level image processing, with 15 common corruptions, demonstrate that our VICT can improve the generalizability of VICL to unseen new domains. In addition, we show the potential of applying VICT for unseen tasks at test time. Code: https://github.com/Jiahao000/VICT.
comment: CVPR 2025. Code: https://github.com/Jiahao000/VICT
☆ StyleMotif: Multi-Modal Motion Stylization using Style-Content Cross Fusion
We present StyleMotif, a novel Stylized Motion Latent Diffusion model, generating motion conditioned on both content and style from multiple modalities. Unlike existing approaches that either focus on generating diverse motion content or transferring style from sequences, StyleMotif seamlessly synthesizes motion across a wide range of content while incorporating stylistic cues from multi-modal inputs, including motion, text, image, video, and audio. To achieve this, we introduce a style-content cross fusion mechanism and align a style encoder with a pre-trained multi-modal model, ensuring that the generated motion accurately captures the reference style while preserving realism. Extensive experiments demonstrate that our framework surpasses existing methods in stylized motion generation and exhibits emergent capabilities for multi-modal motion stylization, enabling more nuanced motion synthesis. Source code and pre-trained models will be released upon acceptance. Project Page: https://stylemotif.github.io
comment: Project Page: https://stylemotif.github.io
☆ Uni4D: Unifying Visual Foundation Models for 4D Modeling from a Single Video CVPR 2025
This paper presents a unified approach to understanding dynamic scenes from casual videos. Large pretrained vision foundation models, such as vision-language, video depth prediction, motion tracking, and segmentation models, offer promising capabilities. However, training a single model for comprehensive 4D understanding remains challenging. We introduce Uni4D, a multi-stage optimization framework that harnesses multiple pretrained models to advance dynamic 3D modeling, including static/dynamic reconstruction, camera pose estimation, and dense 3D motion tracking. Our results show state-of-the-art performance in dynamic 4D modeling with superior visual quality. Notably, Uni4D requires no retraining or fine-tuning, highlighting the effectiveness of repurposing visual foundation models for 4D understanding.
comment: CVPR 2025. Project page (with code): https://davidyao99.github.io/uni4d
☆ Fwd2Bot: LVLM Visual Token Compression with Double Forward Bottleneck
In this work, we aim to compress the vision tokens of a Large Vision Language Model (LVLM) into a representation that is simultaneously suitable for (a) generative and (b) discriminative tasks, (c) is nearly lossless, and (d) is storage-efficient. We propose a novel compression approach, called Fwd2Bot, that uses the LVLM itself to compress the visual information in a task-agnostic manner. At the core of Fwd2bot there exists a "double-forward pass" training strategy, whereby, during the first forward pass, the LLM (of the LVLM) creates a bottleneck by condensing the visual information into a small number of summary tokens. Then, using the same LLM, the second forward pass processes the language instruction(s) alongside the summary tokens, used as a direct replacement for the image ones. The training signal is provided by two losses: an autoregressive one applied after the second pass that provides a direct optimization objective for compression, and a contrastive loss, applied after the first pass, that further boosts the representation strength, especially for discriminative tasks. The training is further enhanced by stage-specific adapters. We accompany the proposed method by an in-depth ablation study. Overall, Fwd2Bot results in highly-informative compressed representations suitable for both generative and discriminative tasks. For generative tasks, we offer a 2x higher compression rate without compromising the generative capabilities, setting a new state-of-the-art result. For discriminative tasks, we set a new state-of-the-art on image retrieval and compositionality.
☆ A Unified Framework for Diffusion Bridge Problems: Flow Matching and Schrödinger Matching into One
The bridge problem is to find an SDE (or sometimes an ODE) that bridges two given distributions. The application areas of the bridge problem are enormous, among which the recent generative modeling (e.g., conditional or unconditional image generation) is the most popular. Also the famous Schr\"{o}dinger bridge problem, a widely known problem for a century, is a special instance of the bridge problem. Two most popular algorithms to tackle the bridge problems in the deep learning era are: (conditional) flow matching and iterative fitting algorithms, where the former confined to ODE solutions, and the latter specifically for the Schr\"{o}dinger bridge problem. The main contribution of this article is in two folds: i) We provide concise reviews of these algorithms with technical details to some extent; ii) We propose a novel unified perspective and framework that subsumes these seemingly unrelated algorithms (and their variants) into one. In particular, we show that our unified framework can instantiate the Flow Matching (FM) algorithm, the (mini-batch) optimal transport FM algorithm, the (mini-batch) Schr\"{o}dinger bridge FM algorithm, and the deep Schr\"{o}dinger bridge matching (DSBM) algorithm as its special cases. We believe that this unified framework will be useful for viewing the bridge problems in a more general and flexible perspective, and in turn can help researchers and practitioners to develop new bridge algorithms in their fields.
☆ CTRL-O: Language-Controllable Object-Centric Visual Representation Learning CVPR 2025
Object-centric representation learning aims to decompose visual scenes into fixed-size vectors called "slots" or "object files", where each slot captures a distinct object. Current state-of-the-art object-centric models have shown remarkable success in object discovery in diverse domains, including complex real-world scenes. However, these models suffer from a key limitation: they lack controllability. Specifically, current object-centric models learn representations based on their preconceived understanding of objects, without allowing user input to guide which objects are represented. Introducing controllability into object-centric models could unlock a range of useful capabilities, such as the ability to extract instance-specific representations from a scene. In this work, we propose a novel approach for user-directed control over slot representations by conditioning slots on language descriptions. The proposed ConTRoLlable Object-centric representation learning approach, which we term CTRL-O, achieves targeted object-language binding in complex real-world scenes without requiring mask supervision. Next, we apply these controllable slot representations on two downstream vision language tasks: text-to-image generation and visual question answering. The proposed approach enables instance-specific text-to-image generation and also achieves strong performance on visual question answering.
comment: Accepted at CVPR 2025
☆ Effective Skill Unlearning through Intervention and Abstention NAACL 2025
Large language Models (LLMs) have demonstrated remarkable skills across various domains. Understanding the mechanisms behind their abilities and implementing controls over them is becoming increasingly important for developing better models. In this paper, we focus on skill unlearning in LLMs, specifically unlearning a particular skill while retaining their overall capabilities. We introduce two lightweight, training-free machine skill unlearning techniques for LLMs. First, we observe that the pre-activation distribution of neurons in each Feed-Forward Layer (FFL) differs when the model demonstrates different skills. Additionally, we find that queries triggering the same skill cluster within the FFL key space and can be separated from other queries using a hypercube. Based on these observations, we propose two lightweight, training-free skill unlearning methods via \textit{intervention} and \textit{abstention} respectively: \texttt{Neuron Adjust} and \texttt{Key Space Detection}. We evaluate our methods on unlearning math-solving, Python-coding, and comprehension skills across seven different languages. The results demonstrate their strong unlearning capabilities for the designated skills. Specifically, \texttt{Key Space Detection} achieves over 80\% relative performance drop on the forgetting skill and less than 10\% relative performance drop on other skills and the model's general knowledge (MMLU) for most unlearning tasks. Our code is available at https://github.com/Trustworthy-ML-Lab/effective_skill_unlearning
comment: Accepted to NAACL 2025 main conference
☆ Energy Minimization for Participatory Federated Learning in IoT Analyzed via Game Theory
The Internet of Things requires intelligent decision making in many scenarios. To this end, resources available at the individual nodes for sensing or computing, or both, can be leveraged. This results in approaches known as participatory sensing and federated learning, respectively. We investigate the simultaneous implementation of both, through a distributed approach based on empowering local nodes with game theoretic decision making. A global objective of energy minimization is combined with the individual node's optimization of local expenditure for sensing and transmitting data over multiple learning rounds. We present extensive evaluations of this technique, based on both a theoretical framework and experiments in a simulated network scenario with real data. Such a distributed approach can reach a desired level of accuracy for federated learning without a centralized supervision of the data collector. However, depending on the weight attributed to the local costs of the single node, it may also result in a significantly high Price of Anarchy (from 1.28 onwards). Thus, we argue for the need of incentive mechanisms, possibly based on Age of Information of the single nodes.
comment: 6 pages, 6 figures, 2 tables, conference
☆ Elementwise Layer Normalization
A recent paper proposed Dynamic Tanh (DyT) as a drop-in replacement for Layer Normalization. Although the method is empirically well-motivated and appealing from a practical point of view, it lacks a theoretical foundation. In this work, we derive DyT mathematically and show that a well-defined approximation is needed to do so. By dropping said approximation, an alternative element-wise transformation is obtained, which we call Elementwise Layer Normalization (ELN). We demonstrate that ELN resembles Layer Normalization more accurately than DyT does.
comment: 11 pages, 3 figures
☆ Learning to Represent Individual Differences for Choice Decision Making IJCAI
Human decision making can be challenging to predict because decisions are affected by a number of complex factors. Adding to this complexity, decision-making processes can differ considerably between individuals, and methods aimed at predicting human decisions need to take individual differences into account. Behavioral science offers methods by which to measure individual differences (e.g., questionnaires, behavioral models), but these are often narrowed down to low dimensions and not tailored to specific prediction tasks. This paper investigates the use of representation learning to measure individual differences from behavioral experiment data. Representation learning offers a flexible approach to create individual embeddings from data that are both structured (e.g., demographic information) and unstructured (e.g., free text), where the flexibility provides more options for individual difference measures for personalization, e.g., free text responses may allow for open-ended questions that are less privacy-sensitive. In the current paper we use representation learning to characterize individual differences in human performance on an economic decision-making task. We demonstrate that models using representation learning to capture individual differences consistently improve decision predictions over models without representation learning, and even outperform well-known theory-based behavioral models used in these environments. Our results propose that representation learning offers a useful and flexible tool to capture individual differences.
comment: Published in IJCAI MRC 2022
☆ Molecular Quantum Transformer
The Transformer model, renowned for its powerful attention mechanism, has achieved state-of-the-art performance in various artificial intelligence tasks but faces challenges such as high computational cost and memory usage. Researchers are exploring quantum computing to enhance the Transformer's design, though it still shows limited success with classical data. With a growing focus on leveraging quantum machine learning for quantum data, particularly in quantum chemistry, we propose the Molecular Quantum Transformer (MQT) for modeling interactions in molecular quantum systems. By utilizing quantum circuits to implement the attention mechanism on the molecular configurations, MQT can efficiently calculate ground-state energies for all configurations. Numerical demonstrations show that in calculating ground-state energies for H_2, LiH, BeH_2, and H_4, MQT outperforms the classical Transformer, highlighting the promise of quantum effects in Transformer structures. Furthermore, its pretraining capability on diverse molecular data facilitates the efficient learning of new molecules, extending its applicability to complex molecular systems with minimal additional effort. Our method offers an alternative to existing quantum algorithms for estimating ground-state energies, opening new avenues in quantum chemistry and materials science.
comment: 13 pages, 8 figures
☆ A Comprehensive Benchmark for RNA 3D Structure-Function Modeling
The RNA structure-function relationship has recently garnered significant attention within the deep learning community, promising to grow in importance as nucleic acid structure models advance. However, the absence of standardized and accessible benchmarks for deep learning on RNA 3D structures has impeded the development of models for RNA functional characteristics. In this work, we introduce a set of seven benchmarking datasets for RNA structure-function prediction, designed to address this gap. Our library builds on the established Python library rnaglib, and offers easy data distribution and encoding, splitters and evaluation methods, providing a convenient all-in-one framework for comparing models. Datasets are implemented in a fully modular and reproducible manner, facilitating for community contributions and customization. Finally, we provide initial baseline results for all tasks using a graph neural network. Source code: https://github.com/cgoliver/rnaglib Documentation: https://rnaglib.org
☆ A tale of two goals: leveraging sequentiality in multi-goal scenarios
Several hierarchical reinforcement learning methods leverage planning to create a graph or sequences of intermediate goals, guiding a lower-level goal-conditioned (GC) policy to reach some final goals. The low-level policy is typically conditioned on the current goal, with the aim of reaching it as quickly as possible. However, this approach can fail when an intermediate goal can be reached in multiple ways, some of which may make it impossible to continue toward subsequent goals. To address this issue, we introduce two instances of Markov Decision Process (MDP) where the optimization objective favors policies that not only reach the current goal but also subsequent ones. In the first, the agent is conditioned on both the current and final goals, while in the second, it is conditioned on the next two goals in the sequence. We conduct a series of experiments on navigation and pole-balancing tasks in which sequences of intermediate goals are given. By evaluating policies trained with TD3+HER on both the standard GC-MDP and our proposed MDPs, we show that, in most cases, conditioning on the next two goals improves stability and sample efficiency over other approaches.
comment: 14 pages, 5 figures
☆ How do language models learn facts? Dynamics, curricula and hallucinations
Large language models accumulate vast knowledge during pre-training, yet the dynamics governing this acquisition remain poorly understood. This work investigates the learning dynamics of language models on a synthetic factual recall task, uncovering three key findings: First, language models learn in three phases, exhibiting a performance plateau before acquiring precise factual knowledge. Mechanistically, this plateau coincides with the formation of attention-based circuits that support recall. Second, the training data distribution significantly impacts learning dynamics, as imbalanced distributions lead to shorter plateaus. Finally, hallucinations emerge simultaneously with knowledge, and integrating new knowledge into the model through fine-tuning is challenging, as it quickly corrupts its existing parametric memories. Our results emphasize the importance of data distribution in knowledge acquisition and suggest novel data scheduling strategies to accelerate neural network training.
☆ Cognitive Science-Inspired Evaluation of Core Capabilities for Object Understanding in AI
One of the core components of our world models is 'intuitive physics' - an understanding of objects, space, and causality. This capability enables us to predict events, plan action and navigate environments, all of which rely on a composite sense of objecthood. Despite its importance, there is no single, unified account of objecthood, though multiple theoretical frameworks provide insights. In the first part of this paper, we present a comprehensive overview of the main theoretical frameworks in objecthood research - Gestalt psychology, enactive cognition, and developmental psychology - and identify the core capabilities each framework attributes to object understanding, as well as what functional roles they play in shaping world models in biological agents. Given the foundational role of objecthood in world modelling, understanding objecthood is also essential in AI. In the second part of the paper, we evaluate how current AI paradigms approach and test objecthood capabilities compared to those in cognitive science. We define an AI paradigm as a combination of how objecthood is conceptualised, the methods used for studying objecthood, the data utilised, and the evaluation techniques. We find that, whilst benchmarks can detect that AI systems model isolated aspects of objecthood, the benchmarks cannot detect when AI systems lack functional integration across these capabilities, not solving the objecthood challenge fully. Finally, we explore novel evaluation approaches that align with the integrated vision of objecthood outlined in this paper. These methods are promising candidates for advancing from isolated object capabilities toward general-purpose AI with genuine object understanding in real-world contexts.
☆ Model Assembly Learning with Heterogeneous Layer Weight Merging ICLR 2025
Model merging acquires general capabilities without extra data or training by combining multiple models' parameters. Previous approaches achieve linear mode connectivity by aligning parameters into the same loss basin using permutation invariance. In this paper, we introduce Model Assembly Learning (MAL), a novel paradigm for model merging that iteratively integrates parameters from diverse models in an open-ended model zoo to enhance the base model's capabilities. Unlike previous works that require identical architectures, MAL allows the merging of heterogeneous architectures and selective parameters across layers. Specifically, the base model can incorporate parameters from different layers of multiple pre-trained models. We systematically investigate the conditions and fundamental settings of heterogeneous parameter merging, addressing all possible mismatches in layer widths between the base and target models. Furthermore, we establish key laws and provide practical guidelines for effectively implementing MAL.
comment: ICLR 2025 Workshop on Neural Network Weights as a New Data Modality
☆ Towards Fully Automated Decision-Making Systems for Greenhouse Control: Challenges and Opportunities
Machine learning has been successful in building control policies to drive a complex system to desired states in various applications (e.g. games, robotics, etc.). To be specific, a number of parameters of policy can be automatically optimized from the observations of environment to be able to generate a sequence of decisions leading to the best performance. In this survey paper, we particularly explore such policy-learning techniques for another unique, practical use-case scenario--farming, in which critical decisions (e.g., water supply, heating, etc.) must be made in a timely manner to minimize risks (e.g., damage to plants) while maximizing the revenue (e.g., healthy crops) in the end. We first provide a broad overview of latest studies on it to identify not only domain-specific challenges but opportunities with potential solutions, some of which are suggested as promising directions for future research. Also, we then introduce our successful approach to being ranked second among 46 teams at the ''3rd Autonomous Greenhouse Challenge'' to use this specific example to discuss the lessons learned about important considerations for design to create autonomous farm-management systems.
☆ Data-Driven Extreme Response Estimation
A method to rapidly estimate extreme ship response events is developed in this paper. The method involves training by a Long Short-Term Memory (LSTM) neural network to correct a lower-fidelity hydrodynamic model to the level of a higher-fidelity simulation. More focus is placed on larger responses by isolating the time-series near peak events identified in the lower-fidelity simulations and training on only the shorter time-series around the large event. The method is tested on the estimation of pitch time-series maxima in Sea State 5 (significant wave height of 4.0 meters and modal period of 15.0 seconds,) generated by a lower-fidelity hydrodynamic solver known as SimpleCode and a higher-fidelity tool known as the Large Amplitude Motion Program (LAMP). The results are also compared with an LSTM trained without special considerations for large events.
comment: From the 35th Symposium on Naval Hydrodynamics
☆ When Astronomy Meets AI: Manazel For Crescent Visibility Prediction in Morocco
The accurate determination of the beginning of each Hijri month is essential for religious, cultural, and administrative purposes. Manazel (The code and datasets are available at https://github.com/lairgiyassir/manazel) addresses this challenge in Morocco by leveraging 13 years of crescent visibility data to refine the ODEH criterion, a widely used standard for lunar crescent visibility prediction. The study integrates two key features, the Arc of Vision (ARCV) and the total width of the crescent (W), to enhance the accuracy of lunar visibility assessments. A machine learning approach utilizing the Logistic Regression algorithm is employed to classify crescent visibility conditions, achieving a predictive accuracy of 98.83%. This data-driven methodology offers a robust and reliable framework for determining the start of the Hijri month, comparing different data classification tools, and improving the consistency of lunar calendar calculations in Morocco. The findings demonstrate the effectiveness of machine learning in astronomical applications and highlight the potential for further enhancements in the modeling of crescent visibility.
☆ ClusterSC: Advancing Synthetic Control with Donor Selection
In causal inference with observational studies, synthetic control (SC) has emerged as a prominent tool. SC has traditionally been applied to aggregate-level datasets, but more recent work has extended its use to individual-level data. As they contain a greater number of observed units, this shift introduces the curse of dimensionality to SC. To address this, we propose Cluster Synthetic Control (ClusterSC), based on the idea that groups of individuals may exist where behavior aligns internally but diverges between groups. ClusterSC incorporates a clustering step to select only the relevant donors for the target. We provide theoretical guarantees on the improvements induced by ClusterSC, supported by empirical demonstrations on synthetic and real-world datasets. The results indicate that ClusterSC consistently outperforms classical SC approaches.
comment: 35 pages, 11 figures, to be published in Proceedings of The 28th International Conference on Artificial Intelligence and Statistics (AIStats) 2025
☆ Provable Reduction in Communication Rounds for Non-Smooth Convex Federated Learning
Multiple local steps are key to communication-efficient federated learning. However, theoretical guarantees for such algorithms, without data heterogeneity-bounding assumptions, have been lacking in general non-smooth convex problems. Leveraging projection-efficient optimization methods, we propose FedMLS, a federated learning algorithm with provable improvements from multiple local steps. FedMLS attains an $\epsilon$-suboptimal solution in $\mathcal{O}(1/\epsilon)$ communication rounds, requiring a total of $\mathcal{O}(1/\epsilon^2)$ stochastic subgradient oracle calls.
☆ Leveraging Language Models for Analyzing Longitudinal Experiential Data in Education
We propose a novel approach to leveraging pre-trained language models (LMs) for early forecasting of academic trajectories in STEM students using high-dimensional longitudinal experiential data. This data, which captures students' study-related activities, behaviors, and psychological states, offers valuable insights for forecasting-based interventions. Key challenges in handling such data include high rates of missing values, limited dataset size due to costly data collection, and complex temporal variability across modalities. Our approach addresses these issues through a comprehensive data enrichment process, integrating strategies for managing missing values, augmenting data, and embedding task-specific instructions and contextual cues to enhance the models' capacity for learning temporal patterns. Through extensive experiments on a curated student learning dataset, we evaluate both encoder-decoder and decoder-only LMs. While our findings show that LMs effectively integrate data across modalities and exhibit resilience to missing data, they primarily rely on high-level statistical patterns rather than demonstrating a deeper understanding of temporal dynamics. Furthermore, their ability to interpret explicit temporal information remains limited. This work advances educational data science by highlighting both the potential and limitations of LMs in modeling student trajectories for early intervention based on longitudinal experiential data.
☆ Nonlinear Multiple Response Regression and Learning of Latent Spaces
Identifying low-dimensional latent structures within high-dimensional data has long been a central topic in the machine learning community, driven by the need for data compression, storage, transmission, and deeper data understanding. Traditional methods, such as principal component analysis (PCA) and autoencoders (AE), operate in an unsupervised manner, ignoring label information even when it is available. In this work, we introduce a unified method capable of learning latent spaces in both unsupervised and supervised settings. We formulate the problem as a nonlinear multiple-response regression within an index model context. By applying the generalized Stein's lemma, the latent space can be estimated without knowing the nonlinear link functions. Our method can be viewed as a nonlinear generalization of PCA. Moreover, unlike AE and other neural network methods that operate as "black boxes", our approach not only offers better interpretability but also reduces computational complexity while providing strong theoretical guarantees. Comprehensive numerical experiments and real data analyses demonstrate the superior performance of our method.
☆ Critical Iterative Denoising: A Discrete Generative Model Applied to Graphs
Discrete Diffusion and Flow Matching models have significantly advanced generative modeling for discrete structures, including graphs. However, the time dependencies in the noising process of these models lead to error accumulation and propagation during the backward process. This issue, particularly pronounced in mask diffusion, is a known limitation in sequence modeling and, as we demonstrate, also impacts discrete diffusion models for graphs. To address this problem, we propose a novel framework called Iterative Denoising, which simplifies discrete diffusion and circumvents the issue by assuming conditional independence across time. Additionally, we enhance our model by incorporating a Critic, which during generation selectively retains or corrupts elements in an instance based on their likelihood under the data distribution. Our empirical evaluations demonstrate that the proposed method significantly outperforms existing discrete diffusion baselines in graph generation tasks.
☆ Generalizable Implicit Neural Representations via Parameterized Latent Dynamics for Baroclinic Ocean Forecasting
Mesoscale ocean dynamics play a critical role in climate systems, governing heat transport, hurricane genesis, and drought patterns. However, simulating these processes at high resolution remains computationally prohibitive due to their nonlinear, multiscale nature and vast spatiotemporal domains. Implicit neural representations (INRs) reduce the computational costs as resolution-independent surrogates but fail in many-query scenarios (inverse modeling) requiring rapid evaluations across diverse parameters. We present PINROD, a novel framework combining dynamics-aware implicit neural representations with parameterized neural ordinary differential equations to address these limitations. By integrating parametric dependencies into latent dynamics, our method efficiently captures nonlinear oceanic behavior across varying boundary conditions and physical parameters. Experiments on ocean mesoscale activity data show superior accuracy over existing baselines and improved computational efficiency compared to standard numerical simulations.
☆ Probabilistic Functional Neural Networks
High-dimensional functional time series (HDFTS) are often characterized by nonlinear trends and high spatial dimensions. Such data poses unique challenges for modeling and forecasting due to the nonlinearity, nonstationarity, and high dimensionality. We propose a novel probabilistic functional neural network (ProFnet) to address these challenges. ProFnet integrates the strengths of feedforward and deep neural networks with probabilistic modeling. The model generates probabilistic forecasts using Monte Carlo sampling and also enables the quantification of uncertainty in predictions. While capturing both temporal and spatial dependencies across multiple regions, ProFnet offers a scalable and unified solution for large datasets. Applications to Japan's mortality rates demonstrate superior performance. This approach enhances predictive accuracy and provides interpretable uncertainty estimates, making it a valuable tool for forecasting complex high-dimensional functional data and HDFTS.
☆ Fusion of Graph Neural Networks via Optimal Transport
In this paper, we explore the idea of combining GCNs into one model. To that end, we align the weights of different models layer-wise using optimal transport (OT). We present and evaluate three types of transportation costs and show that the studied fusion method consistently outperforms the performance of vanilla averaging. Finally, we present results suggesting that model fusion using OT is harder in the case of GCNs than MLPs and that incorporating the graph structure into the process does not improve the performance of the method.
☆ Consistent Multigroup Low-Rank Approximation
We consider the problem of consistent low-rank approximation for multigroup data: we ask for a sequence of $k$ basis vectors such that projecting the data onto their spanned subspace treats all groups as equally as possible, by minimizing the maximum error among the groups. Additionally, we require that the sequence of basis vectors satisfies the natural consistency property: when looking for the best $k$ vectors, the first $d
☆ SyncSDE: A Probabilistic Framework for Diffusion Synchronization CVPR2025
There have been many attempts to leverage multiple diffusion models for collaborative generation, extending beyond the original domain. A prominent approach involves synchronizing multiple diffusion trajectories by mixing the estimated scores to artificially correlate the generation processes. However, existing methods rely on naive heuristics, such as averaging, without considering task specificity. These approaches do not clarify why such methods work and often fail when a heuristic suitable for one task is blindly applied to others. In this paper, we present a probabilistic framework for analyzing why diffusion synchronization works and reveal where heuristics should be focused - modeling correlations between multiple trajectories and adapting them to each specific task. We further identify optimal correlation models per task, achieving better results than previous approaches that apply a single heuristic across all tasks without justification.
comment: Accepted to CVPR2025
☆ SWI: Speaking with Intent in Large Language Models
Intent, typically clearly formulated and planned, functions as a cognitive framework for reasoning and problem-solving. This paper introduces the concept of Speaking with Intent (SWI) in large language models (LLMs), where the explicitly generated intent encapsulates the model's underlying intention and provides high-level planning to guide subsequent analysis and communication. By emulating deliberate and purposeful thoughts in the human mind, SWI is hypothesized to enhance the reasoning capabilities and generation quality of LLMs. Extensive experiments on mathematical reasoning benchmarks consistently demonstrate the superiority of Speaking with Intent over Baseline (i.e., generation without explicit intent). Moreover, SWI outperforms answer-trigger prompting methods Chain-of-Thought and Plan-and-Solve and maintains competitive performance with the strong method ARR (Analyzing, Retrieving, and Reasoning). Additionally, the effectiveness and generalizability of SWI are solidified on reasoning-intensive question answering (QA) and text summarization benchmarks, where SWI brings consistent improvement to the Baseline generation. In text summarization, SWI-generated summaries exhibit greater accuracy, conciseness, and factual correctness, with fewer hallucinations. Furthermore, human evaluations verify the coherence, effectiveness, and interpretability of the intent produced by SWI. This proof-of-concept study creates a novel avenue for enhancing LLMs' reasoning abilities with cognitive notions.
comment: 24 pages. Code: https://github.com/YuweiYin/SWI
☆ Formation Shape Control using the Gromov-Wasserstein Metric
This article introduces a formation shape control algorithm, in the optimal control framework, for steering an initial population of agents to a desired configuration via employing the Gromov-Wasserstein distance. The underlying dynamical system is assumed to be a constrained linear system and the objective function is a sum of quadratic control-dependent stage cost and a Gromov-Wasserstein terminal cost. The inclusion of the Gromov-Wasserstein cost transforms the resulting optimal control problem into a well-known NP-hard problem, making it both numerically demanding and difficult to solve with high accuracy. Towards that end, we employ a recent semi-definite relaxation-driven technique to tackle the Gromov-Wasserstein distance. A numerical example is provided to illustrate our results.
comment: To appear in the proceedings of Learning for Dynamics and Control (L4DC) conference, PMLR, 2025
☆ Exploring the Energy Landscape of RBMs: Reciprocal Space Insights into Bosons, Hierarchical Learning and Symmetry Breaking
Deep generative models have become ubiquitous due to their ability to learn and sample from complex distributions. Despite the proliferation of various frameworks, the relationships among these models remain largely unexplored, a gap that hinders the development of a unified theory of AI learning. We address two central challenges: clarifying the connections between different deep generative models and deepening our understanding of their learning mechanisms. We focus on Restricted Boltzmann Machines (RBMs), known for their universal approximation capabilities for discrete distributions. By introducing a reciprocal space formulation, we reveal a connection between RBMs, diffusion processes, and coupled Bosons. We show that at initialization, the RBM operates at a saddle point, where the local curvature is determined by the singular values, whose distribution follows the Marcenko-Pastur law and exhibits rotational symmetry. During training, this rotational symmetry is broken due to hierarchical learning, where different degrees of freedom progressively capture features at multiple levels of abstraction. This leads to a symmetry breaking in the energy landscape, reminiscent of Landau theory. This symmetry breaking in the energy landscape is characterized by the singular values and the weight matrix eigenvector matrix. We derive the corresponding free energy in a mean-field approximation. We show that in the limit of infinite size RBM, the reciprocal variables are Gaussian distributed. Our findings indicate that in this regime, there will be some modes for which the diffusion process will not converge to the Boltzmann distribution. To illustrate our results, we trained replicas of RBMs with different hidden layer sizes using the MNIST dataset. Our findings bridge the gap between disparate generative frameworks and also shed light on the processes underpinning learning in generative models.
comment: 19pp, 8figs, research article
☆ Bayesian Pseudo Posterior Mechanism for Differentially Private Machine Learning
Differential privacy (DP) is becoming increasingly important for deployed machine learning applications because it provides strong guarantees for protecting the privacy of individuals whose data is used to train models. However, DP mechanisms commonly used in machine learning tend to struggle on many real world distributions, including highly imbalanced or small labeled training sets. In this work, we propose a new scalable DP mechanism for deep learning models, SWAG-PPM, by using a pseudo posterior distribution that downweights by-record likelihood contributions proportionally to their disclosure risks as the randomized mechanism. As a motivating example from official statistics, we demonstrate SWAG-PPM on a workplace injury text classification task using a highly imbalanced public dataset published by the U.S. Occupational Safety and Health Administration (OSHA). We find that SWAG-PPM exhibits only modest utility degradation against a non-private comparator while greatly outperforming the industry standard DP-SGD for a similar privacy budget.
☆ Constraint-based causal discovery with tiered background knowledge and latent variables in single or overlapping datasets
In this paper we consider the use of tiered background knowledge within constraint based causal discovery. Our focus is on settings relaxing causal sufficiency, i.e. allowing for latent variables which may arise because relevant information could not be measured at all, or not jointly, as in the case of multiple overlapping datasets. We first present novel insights into the properties of the 'tiered FCI' (tFCI) algorithm. Building on this, we introduce a new extension of the IOD (integrating overlapping datasets) algorithm incorporating tiered background knowledge, the 'tiered IOD' (tIOD) algorithm. We show that under full usage of the tiered background knowledge tFCI and tIOD are sound, while simple versions of the tIOD and tFCI are sound and complete. We further show that the tIOD algorithm can often be expected to be considerably more efficient and informative than the IOD algorithm even beyond the obvious restriction of the Markov equivalence classes. We provide a formal result on the conditions for this gain in efficiency and informativeness. Our results are accompanied by a series of examples illustrating the exact role and usefulness of tiered background knowledge.
comment: Accepted for the 4th Conference on Causal Learning and Reasoning (CLeaR 2025)
☆ Quantitative Evaluation of Quantum/Classical Neural Network Using a Game Solver Metric
To evaluate the performance of quantum computing systems relative to classical counterparts and explore the potential for quantum advantage, we propose a game-solving benchmark based on Elo ratings in the game of tic-tac-toe. We compare classical convolutional neural networks (CNNs), quantum convolutional neural networks (QCNNs), and hybrid classical-quantum models by assessing their performance against a random-move agent in automated matches. Additionally, we implement a QCNN integrated with quantum communication and evaluate its performance to quantify the overhead introduced by noisy quantum channels. Our results show that the classical-quantum hybrid model achieves Elo ratings comparable to those of classical CNNs, while the standalone QCNN underperforms under current hardware constraints. The communication overhead was found to be modest. These findings demonstrate the viability of using game-based benchmarks for evaluating quantum computing systems and suggest that quantum communication can be incorporated with limited impact on performance, providing a foundation for future hybrid quantum applications.
comment: 11 pages, 16 figures
☆ Uncertainty-aware Bayesian machine learning modelling of land cover classification
Land cover classification involves the production of land cover maps, which determine the type of land through remote sensing imagery. Over recent years, such classification is being performed by machine learning classification models, which can give highly accurate predictions on land cover per pixel using large quantities of input training data. However, such models do not currently take account of input measurement uncertainty, which is vital for traceability in metrology. In this work we propose a Bayesian classification framework using generative modelling to take account of input measurement uncertainty. We take the specific case of Bayesian quadratic discriminant analysis, and apply it to land cover datasets from Copernicus Sentinel-2 in 2020 and 2021. We benchmark the performance of the model against more popular classification models used in land cover maps such as random forests and neural networks. We find that such Bayesian models are more trustworthy, in the sense that they are more interpretable, explicitly model the input measurement uncertainty, and maintain predictive performance of class probability outputs across datasets of different years and sizes, whilst also being computationally efficient.
comment: 31 pages, 10 figures
☆ F-INR: Functional Tensor Decomposition for Implicit Neural Representations
Implicit Neural Representation (INR) has emerged as a powerful tool for encoding discrete signals into continuous, differentiable functions using neural networks. However, these models often have an unfortunate reliance on monolithic architectures to represent high-dimensional data, leading to prohibitive computational costs as dimensionality grows. We propose F-INR, a framework that reformulates INR learning through functional tensor decomposition, breaking down high-dimensional tasks into lightweight, axis-specific sub-networks. Each sub-network learns a low-dimensional data component (e.g., spatial or temporal). Then, we combine these components via tensor operations, reducing forward pass complexity while improving accuracy through specialized learning. F-INR is modular and, therefore, architecture-agnostic, compatible with MLPs, SIREN, WIRE, or other state-of-the-art INR architecture. It is also decomposition-agnostic, supporting CP, TT, and Tucker modes with user-defined rank for speed-accuracy control. In our experiments, F-INR trains $100\times$ faster than existing approaches on video tasks while achieving higher fidelity (+3.4 dB PSNR). Similar gains hold for image compression, physics simulations, and 3D geometry reconstruction. Through this, F-INR offers a new scalable, flexible solution for high-dimensional signal modeling.
comment: 26 pages, 33 figures, 12 tables
☆ Adaptive Resampling with Bootstrap for Noisy Multi-Objective Optimization Problems
The challenge of noisy multi-objective optimization lies in the constant trade-off between exploring new decision points and improving the precision of known points through resampling. This decision should take into account both the variability of the objective functions and the current estimate of a point in relation to the Pareto front. Since the amount and distribution of noise are generally unknown, it is desirable for a decision function to be highly adaptive to the properties of the optimization problem. This paper presents a resampling decision function that incorporates the stochastic nature of the optimization problem by using bootstrapping and the probability of dominance. The distribution-free estimation of the probability of dominance is achieved using bootstrap estimates of the means. To make the procedure applicable even with very few observations, we transfer the distribution observed at other decision points. The efficiency of this resampling approach is demonstrated by applying it in the NSGA-II algorithm with a sequential resampling procedure under multiple noise variations.
comment: 14 pages. 5 figures
☆ Robust DNN Partitioning and Resource Allocation Under Uncertain Inference Time
In edge intelligence systems, deep neural network (DNN) partitioning and data offloading can provide real-time task inference for resource-constrained mobile devices. However, the inference time of DNNs is typically uncertain and cannot be precisely determined in advance, presenting significant challenges in ensuring timely task processing within deadlines. To address the uncertain inference time, we propose a robust optimization scheme to minimize the total energy consumption of mobile devices while meeting task probabilistic deadlines. The scheme only requires the mean and variance information of the inference time, without any prediction methods or distribution functions. The problem is formulated as a mixed-integer nonlinear programming (MINLP) that involves jointly optimizing the DNN model partitioning and the allocation of local CPU/GPU frequencies and uplink bandwidth. To tackle the problem, we first decompose the original problem into two subproblems: resource allocation and DNN model partitioning. Subsequently, the two subproblems with probability constraints are equivalently transformed into deterministic optimization problems using the chance-constrained programming (CCP) method. Finally, the convex optimization technique and the penalty convex-concave procedure (PCCP) technique are employed to obtain the optimal solution of the resource allocation subproblem and a stationary point of the DNN model partitioning subproblem, respectively. The proposed algorithm leverages real-world data from popular hardware platforms and is evaluated on widely used DNN models. Extensive simulations show that our proposed algorithm effectively addresses the inference time uncertainty with probabilistic deadline guarantees while minimizing the energy consumption of mobile devices.
☆ The Procedural Content Generation Benchmark: An Open-source Testbed for Generative Challenges in Games
This paper introduces the Procedural Content Generation Benchmark for evaluating generative algorithms on different game content creation tasks. The benchmark comes with 12 game-related problems with multiple variants on each problem. Problems vary from creating levels of different kinds to creating rule sets for simple arcade games. Each problem has its own content representation, control parameters, and evaluation metrics for quality, diversity, and controllability. This benchmark is intended as a first step towards a standardized way of comparing generative algorithms. We use the benchmark to score three baseline algorithms: a random generator, an evolution strategy, and a genetic algorithm. Results show that some problems are easier to solve than others, as well as the impact the chosen objective has on quality, diversity, and controllability of the generated artifacts.
comment: 12 pages, 4 figures, 2 tables, published at FDG2025
☆ DeepRV: pre-trained spatial priors for accelerated disease mapping
Recently introduced prior-encoding deep generative models (e.g., PriorVAE, $\pi$VAE, and PriorCVAE) have emerged as powerful tools for scalable Bayesian inference by emulating complex stochastic processes like Gaussian processes (GPs). However, these methods remain largely a proof-of-concept and inaccessible to practitioners. We propose DeepRV, a lightweight, decoder-only approach that accelerates training, and enhances real-world applicability in comparison to current VAE-based prior encoding approaches. Leveraging probabilistic programming frameworks (e.g., NumPyro) for inference, DeepRV achieves significant speedups while also improving the quality of parameter inference, closely matching full MCMC sampling. We showcase its effectiveness in process emulation and spatial analysis of the UK using simulated data, gender-wise cancer mortality rates for individuals under 50, and HIV prevalence in Zimbabwe. To bridge the gap between theory and practice, we provide a user-friendly API, enabling scalable and efficient Bayesian inference.
☆ Retinal Fundus Multi-Disease Image Classification using Hybrid CNN-Transformer-Ensemble Architectures
Our research is motivated by the urgent global issue of a large population affected by retinal diseases, which are evenly distributed but underserved by specialized medical expertise, particularly in non-urban areas. Our primary objective is to bridge this healthcare gap by developing a comprehensive diagnostic system capable of accurately predicting retinal diseases solely from fundus images. However, we faced significant challenges due to limited, diverse datasets and imbalanced class distributions. To overcome these issues, we have devised innovative strategies. Our research introduces novel approaches, utilizing hybrid models combining deeper Convolutional Neural Networks (CNNs), Transformer encoders, and ensemble architectures sequentially and in parallel to classify retinal fundus images into 20 disease labels. Our overarching goal is to assess these advanced models' potential in practical applications, with a strong focus on enhancing retinal disease diagnosis accuracy across a broader spectrum of conditions. Importantly, our efforts have surpassed baseline model results, with the C-Tran ensemble model emerging as the leader, achieving a remarkable model score of 0.9166, surpassing the baseline score of 0.9. Additionally, experiments with the IEViT model showcased equally promising outcomes with improved computational efficiency. We've also demonstrated the effectiveness of dynamic patch extraction and the integration of domain knowledge in computer vision tasks. In summary, our research strives to contribute significantly to retinal disease diagnosis, addressing the critical need for accessible healthcare solutions in underserved regions while aiming for comprehensive and accurate disease prediction.
comment: 17 pages, 3 figures, 7 tables. Conference paper presented at the International Health Informatics Conference (IHIC 2023)
☆ DATA-WA: Demand-based Adaptive Task Assignment with Dynamic Worker Availability Windows
With the rapid advancement of mobile networks and the widespread use of mobile devices, spatial crowdsourcing, which involves assigning location-based tasks to mobile workers, has gained significant attention. However, most existing research focuses on task assignment at the current moment, overlooking the fluctuating demand and supply between tasks and workers over time. To address this issue, we introduce an adaptive task assignment problem, which aims to maximize the number of assigned tasks by dynamically adjusting task assignments in response to changing demand and supply. We develop a spatial crowdsourcing framework, namely demand-based adaptive task assignment with dynamic worker availability windows, which consists of two components including task demand prediction and task assignment. In the first component, we construct a graph adjacency matrix representing the demand dependency relationships in different regions and employ a multivariate time series learning approach to predict future task demands. In the task assignment component, we adjust tasks to workers based on these predictions, worker availability windows, and the current task assignments, where each worker has an availability window that indicates the time periods they are available for task assignments. To reduce the search space of task assignments and be efficient, we propose a worker dependency separation approach based on graph partition and a task value function with reinforcement learning. Experiments on real data demonstrate that our proposals are both effective and efficient.
☆ Stochastic Engrams for Efficient Continual Learning with Binarized Neural Networks
The ability to learn continuously in artificial neural networks (ANNs) is often limited by catastrophic forgetting, a phenomenon in which new knowledge becomes dominant. By taking mechanisms of memory encoding in neuroscience (aka. engrams) as inspiration, we propose a novel approach that integrates stochastically-activated engrams as a gating mechanism for metaplastic binarized neural networks (mBNNs). This method leverages the computational efficiency of mBNNs combined with the robustness of probabilistic memory traces to mitigate forgetting and maintain the model's reliability. Previously validated metaplastic optimization techniques have been incorporated to enhance synaptic stability further. Compared to baseline binarized models and benchmark fully connected continual learning approaches, our method is the only strategy capable of reaching average accuracies over 20% in class-incremental scenarios and achieving comparable domain-incremental results to full precision state-of-the-art methods. Furthermore, we achieve a significant reduction in peak GPU and RAM usage, under 5% and 20%, respectively. Our findings demonstrate (A) an improved stability vs. plasticity trade-off, (B) a reduced memory intensiveness, and (C) an enhanced performance in binarized architectures. By uniting principles of neuroscience and efficient computing, we offer new insights into the design of scalable and robust deep learning systems.
☆ Exploring the flavor structure of leptons via diffusion models
We propose a method to explore the flavor structure of leptons using diffusion models, which are known as one of generative artificial intelligence (generative AI). We consider a simple extension of the Standard Model with the type I seesaw mechanism and train a neural network to generate the neutrino mass matrix. By utilizing transfer learning, the diffusion model generates 104 solutions that are consistent with the neutrino mass squared differences and the leptonic mixing angles. The distributions of the CP phases and the sums of neutrino masses, which are not included in the conditional labels but are calculated from the solutions, exhibit non-trivial tendencies. In addition, the effective mass in neutrinoless double beta decay is concentrated near the boundaries of the existing confidence intervals, allowing us to verify the obtained solutions through future experiments. An inverse approach using the diffusion model is expected to facilitate the experimental verification of flavor models from a perspective distinct from conventional analytical methods.
comment: 23 pages, 5 figures
☆ Nearest Neighbour Equilibrium Clustering
A novel and intuitive nearest neighbours based clustering algorithm is introduced, in which a cluster is defined in terms of an equilibrium condition which balances its size and cohesiveness. The formulation of the equilibrium condition allows for a quantification of the strength of alignment of each point to a cluster, with these cluster alignment strengths leading naturally to a model selection criterion which renders the proposed approach fully automatable. The algorithm is simple to implement and computationally efficient, and produces clustering solutions of extremely high quality in comparison with relevant benchmarks from the literature. R code to implement the approach is available from https://github.com/DavidHofmeyr/NNEC.
comment: Currently being considered for publication by IEEE
☆ AdvSGM: Differentially Private Graph Learning via Adversarial Skip-gram Model ICDE 2025
The skip-gram model (SGM), which employs a neural network to generate node vectors, serves as the basis for numerous popular graph embedding techniques. However, since the training datasets contain sensitive linkage information, the parameters of a released SGM may encode private information and pose significant privacy risks. Differential privacy (DP) is a rigorous standard for protecting individual privacy in data analysis. Nevertheless, when applying differential privacy to skip-gram in graphs, it becomes highly challenging due to the complex link relationships, which potentially result in high sensitivity and necessitate substantial noise injection. To tackle this challenge, we present AdvSGM, a differentially private skip-gram for graphs via adversarial training. Our core idea is to leverage adversarial training to privatize skip-gram while improving its utility. Towards this end, we develop a novel adversarial training module by devising two optimizable noise terms that correspond to the parameters of a skip-gram. By fine-tuning the weights between modules within AdvSGM, we can achieve differentially private gradient updates without additional noise injection. Extensive experimental results on six real-world graph datasets show that AdvSGM preserves high data utility across different downstream tasks.
comment: Accepted by ICDE 2025
☆ From Deep Learning to LLMs: A survey of AI in Quantitative Investment
Quantitative investment (quant) is an emerging, technology-driven approach in asset management, increasingy shaped by advancements in artificial intelligence. Recent advances in deep learning and large language models (LLMs) for quant finance have improved predictive modeling and enabled agent-based automation, suggesting a potential paradigm shift in this field. In this survey, taking alpha strategy as a representative example, we explore how AI contributes to the quantitative investment pipeline. We first examine the early stage of quant research, centered on human-crafted features and traditional statistical models with an established alpha pipeline. We then discuss the rise of deep learning, which enabled scalable modeling across the entire pipeline from data processing to order execution. Building on this, we highlight the emerging role of LLMs in extending AI beyond prediction, empowering autonomous agents to process unstructured data, generate alphas, and support self-iterative workflows.
☆ Neuro-Symbolic Imitation Learning: Discovering Symbolic Abstractions for Skill Learning ICRA
Imitation learning is a popular method for teaching robots new behaviors. However, most existing methods focus on teaching short, isolated skills rather than long, multi-step tasks. To bridge this gap, imitation learning algorithms must not only learn individual skills but also an abstract understanding of how to sequence these skills to perform extended tasks effectively. This paper addresses this challenge by proposing a neuro-symbolic imitation learning framework. Using task demonstrations, the system first learns a symbolic representation that abstracts the low-level state-action space. The learned representation decomposes a task into easier subtasks and allows the system to leverage symbolic planning to generate abstract plans. Subsequently, the system utilizes this task decomposition to learn a set of neural skills capable of refining abstract plans into actionable robot commands. Experimental results in three simulated robotic environments demonstrate that, compared to baselines, our neuro-symbolic approach increases data efficiency, improves generalization capabilities, and facilitates interpretability.
comment: IEEE International Conference on Robotics and Automation (ICRA) 2025
☆ AcL: Action Learner for Fault-Tolerant Quadruped Locomotion Control
Quadrupedal robots can learn versatile locomotion skills but remain vulnerable when one or more joints lose power. In contrast, dogs and cats can adopt limping gaits when injured, demonstrating their remarkable ability to adapt to physical conditions. Inspired by such adaptability, this paper presents Action Learner (AcL), a novel teacher-student reinforcement learning framework that enables quadrupeds to autonomously adapt their gait for stable walking under multiple joint faults. Unlike conventional teacher-student approaches that enforce strict imitation, AcL leverages teacher policies to generate style rewards, guiding the student policy without requiring precise replication. We train multiple teacher policies, each corresponding to a different fault condition, and subsequently distill them into a single student policy with an encoder-decoder architecture. While prior works primarily address single-joint faults, AcL enables quadrupeds to walk with up to four faulty joints across one or two legs, autonomously switching between different limping gaits when faults occur. We validate AcL on a real Go2 quadruped robot under single- and double-joint faults, demonstrating fault-tolerant, stable walking, smooth gait transitions between normal and lamb gaits, and robustness against external disturbances.
☆ ProHOC: Probabilistic Hierarchical Out-of-Distribution Classification via Multi-Depth Networks CVPR2025
Out-of-distribution (OOD) detection in deep learning has traditionally been framed as a binary task, where samples are either classified as belonging to the known classes or marked as OOD, with little attention given to the semantic relationships between OOD samples and the in-distribution (ID) classes. We propose a framework for detecting and classifying OOD samples in a given class hierarchy. Specifically, we aim to predict OOD data to their correct internal nodes of the class hierarchy, whereas the known ID classes should be predicted as their corresponding leaf nodes. Our approach leverages the class hierarchy to create a probabilistic model and we implement this model by using networks trained for ID classification at multiple hierarchy depths. We conduct experiments on three datasets with predefined class hierarchies and show the effectiveness of our method. Our code is available at https://github.com/walline/prohoc.
comment: CVPR2025
☆ Controlling Large Language Model with Latent Actions
Adapting Large Language Models (LLMs) to downstream tasks using Reinforcement Learning (RL) has proven to be an effective approach. However, LLMs do not inherently define the structure of an agent for RL training, particularly in terms of defining the action space. This paper studies learning a compact latent action space to enhance the controllability and exploration of RL for LLMs. We propose Controlling Large Language Models with Latent Actions (CoLA), a framework that integrates a latent action space into pre-trained LLMs. We apply CoLA to the Llama-3.1-8B model. Our experiments demonstrate that, compared to RL with token-level actions, CoLA's latent action enables greater semantic diversity in text generation. For enhancing downstream tasks, we show that CoLA with RL achieves a score of 42.4 on the math500 benchmark, surpassing the baseline score of 38.2, and reaches 68.2 when augmented with a Monte Carlo Tree Search variant. Furthermore, CoLA with RL consistently improves performance on agent-based tasks without degrading the pre-trained LLM's capabilities, unlike the baseline. Finally, CoLA reduces computation time by half in tasks involving enhanced thinking prompts for LLMs by RL. These results highlight CoLA's potential to advance RL-based adaptation of LLMs for downstream applications.
☆ Investigating the Duality of Interpretability and Explainability in Machine Learning
The rapid evolution of machine learning (ML) has led to the widespread adoption of complex "black box" models, such as deep neural networks and ensemble methods. These models exhibit exceptional predictive performance, making them invaluable for critical decision-making across diverse domains within society. However, their inherently opaque nature raises concerns about transparency and interpretability, making them untrustworthy decision support systems. To alleviate such a barrier to high-stakes adoption, research community focus has been on developing methods to explain black box models as a means to address the challenges they pose. Efforts are focused on explaining these models instead of developing ones that are inherently interpretable. Designing inherently interpretable models from the outset, however, can pave the path towards responsible and beneficial applications in the field of ML. In this position paper, we clarify the chasm between explaining black boxes and adopting inherently interpretable models. We emphasize the imperative need for model interpretability and, following the purpose of attaining better (i.e., more effective or efficient w.r.t. predictive performance) and trustworthy predictors, provide an experimental evaluation of latest hybrid learning methods that integrates symbolic knowledge into neural network predictors. We demonstrate how interpretable hybrid models could potentially supplant black box ones in different domains.
☆ Fine-Tuning LLMs on Small Medical Datasets: Text Classification and Normalization Effectiveness on Cardiology reports and Discharge records
We investigate the effectiveness of fine-tuning large language models (LLMs) on small medical datasets for text classification and named entity recognition tasks. Using a German cardiology report dataset and the i2b2 Smoking Challenge dataset, we demonstrate that fine-tuning small LLMs locally on limited training data can improve performance achieving comparable results to larger models. Our experiments show that fine-tuning improves performance on both tasks, with notable gains observed with as few as 200-300 training examples. Overall, the study highlights the potential of task-specific fine-tuning of LLMs for automating clinical workflows and efficiently extracting structured data from unstructured medical text.
comment: 4 pages, 2 tables,
☆ Scalable Expectation Estimation with Subtractive Mixture Models
Many Monte Carlo (MC) and importance sampling (IS) methods use mixture models (MMs) for their simplicity and ability to capture multimodal distributions. Recently, subtractive mixture models (SMMs), i.e. MMs with negative coefficients, have shown greater expressiveness and success in generative modeling. However, their negative parameters complicate sampling, requiring costly auto-regressive techniques or accept-reject algorithms that do not scale in high dimensions. In this work, we use the difference representation of SMMs to construct an unbiased IS estimator ($\Delta\text{Ex}$) that removes the need to sample from the SMM, enabling high-dimensional expectation estimation with SMMs. In our experiments, we show that $\Delta\text{Ex}$ can achieve comparable estimation quality to auto-regressive sampling while being considerably faster in MC estimation. Moreover, we conduct initial experiments with $\Delta\text{Ex}$ using hand-crafted proposals, gaining first insights into how to construct safe proposals for $\Delta\text{Ex}$.
☆ DuckSegmentation: A segmentation model based on the AnYue Hemp Duck Dataset
The modernization of smart farming is a way to improve agricultural production efficiency, and improve the agricultural production environment. Although many large models have achieved high accuracy in the task of object recognition and segmentation, they cannot really be put into use in the farming industry due to their own poor interpretability and limitations in computational volume. In this paper, we built AnYue Shelduck Dateset, which contains a total of 1951 Shelduck datasets, and performed target detection and segmentation annotation with the help of professional annotators. Based on AnYue ShelduckDateset, this paper describes DuckProcessing, an efficient and powerful module for duck identification based on real shelduckfarms. First of all, using the YOLOv8 module designed to divide the mahjong between them, Precision reached 98.10%, Recall reached 96.53% and F1 score reached 0.95 on the test set. Again using the DuckSegmentation segmentation model, DuckSegmentation reached 96.43% mIoU. Finally, the excellent DuckSegmentation was used as the teacher model, and through knowledge distillation, Deeplabv3 r50 was used as the student model, and the final student model achieved 94.49% mIoU on the test set. The method provides a new way of thinking in practical sisal duck smart farming.
☆ Explainable Boosting Machine for Predicting Claim Severity and Frequency in Car Insurance
In a context of constant increase in competition and heightened regulatory pressure, accuracy, actuarial precision, as well as transparency and understanding of the tariff, are key issues in non-life insurance. Traditionally used generalized linear models (GLM) result in a multiplicative tariff that favors interpretability. With the rapid development of machine learning and deep learning techniques, actuaries and the rest of the insurance industry have adopted these techniques widely. However, there is a need to associate them with interpretability techniques. In this paper, our study focuses on introducing an Explainable Boosting Machine (EBM) model that combines intrinsically interpretable characteristics and high prediction performance. This approach is described as a glass-box model and relies on the use of a Generalized Additive Model (GAM) and a cyclic gradient boosting algorithm. It accounts for univariate and pairwise interaction effects between features and provides naturally explanations on them. We implement this approach on car insurance frequency and severity data and extensively compare the performance of this approach with classical competitors: a GLM, a GAM, a CART model and an Extreme Gradient Boosting (XGB) algorithm. Finally, we examine the interpretability of these models to capture the main determinants of claim costs.
☆ Tricking Retrievers with Influential Tokens: An Efficient Black-Box Corpus Poisoning Attack NAACL 2025
Retrieval-augmented generation (RAG) systems enhance large language models by incorporating external knowledge, addressing issues like outdated internal knowledge and hallucination. However, their reliance on external knowledge bases makes them vulnerable to corpus poisoning attacks, where adversarial passages can be injected to manipulate retrieval results. Existing methods for crafting such passages, such as random token replacement or training inversion models, are often slow and computationally expensive, requiring either access to retriever's gradients or large computational resources. To address these limitations, we propose Dynamic Importance-Guided Genetic Algorithm (DIGA), an efficient black-box method that leverages two key properties of retrievers: insensitivity to token order and bias towards influential tokens. By focusing on these characteristics, DIGA dynamically adjusts its genetic operations to generate effective adversarial passages with significantly reduced time and memory usage. Our experimental evaluation shows that DIGA achieves superior efficiency and scalability compared to existing methods, while maintaining comparable or better attack success rates across multiple datasets.
comment: Accepted to NAACL 2025 Main Track
☆ Simulation-informed deep learning for enhanced SWOT observations of fine-scale ocean dynamics
Oceanic processes at fine scales are crucial yet difficult to observe accurately due to limitations in satellite and in-situ measurements. The Surface Water and Ocean Topography (SWOT) mission provides high-resolution Sea Surface Height (SSH) data, though noise patterns often obscure fine scale structures. Current methods struggle with noisy data or require extensive supervised training, limiting their effectiveness on real-world observations. We introduce SIMPGEN (Simulation-Informed Metric and Prior for Generative Ensemble Networks), an unsupervised adversarial learning framework combining real SWOT observations with simulated reference data. SIMPGEN leverages wavelet-informed neural metrics to distinguish noisy from clean fields, guiding realistic SSH reconstructions. Applied to SWOT data, SIMPGEN effectively removes noise, preserving fine-scale features better than existing neural methods. This robust, unsupervised approach not only improves SWOT SSH data interpretation but also demonstrates strong potential for broader oceanographic applications, including data assimilation and super-resolution.
☆ HOT: Hadamard-based Optimized Training CVPR 2025
It has become increasingly important to optimize backpropagation to reduce memory usage and computational overhead. Achieving this goal is highly challenging, as multiple objectives must be considered jointly while maintaining training quality. In this paper, we focus on matrix multiplication, which accounts for the largest portion of training costs, and analyze its backpropagation in detail to identify lightweight techniques that offer the best benefits. Based on this analysis, we introduce a novel method, Hadamard-based Optimized Training (HOT). In this approach, we apply Hadamard-based optimizations, such as Hadamard quantization and Hadamard low-rank approximation, selectively and with awareness of the suitability of each optimization for different backward paths. Additionally, we introduce two enhancements: activation buffer compression and layer-wise quantizer selection. Our extensive analysis shows that HOT achieves up to 75% memory savings and a 2.6 times acceleration on real GPUs, with negligible accuracy loss compared to FP32 precision.
comment: Accepted in CVPR 2025
☆ Dual-Splitting Conformal Prediction for Multi-Step Time Series Forecasting
Time series forecasting is crucial for applications like resource scheduling and risk management, where multi-step predictions provide a comprehensive view of future trends. Uncertainty Quantification (UQ) is a mainstream approach for addressing forecasting uncertainties, with Conformal Prediction (CP) gaining attention due to its model-agnostic nature and statistical guarantees. However, most variants of CP are designed for single-step predictions and face challenges in multi-step scenarios, such as reliance on real-time data and limited scalability. This highlights the need for CP methods specifically tailored to multi-step forecasting. We propose the Dual-Splitting Conformal Prediction (DSCP) method, a novel CP approach designed to capture inherent dependencies within time-series data for multi-step forecasting. Experimental results on real-world datasets from four different domains demonstrate that the proposed DSCP significantly outperforms existing CP variants in terms of the Winkler Score, achieving a performance improvement of up to 23.59% compared to state-of-the-art methods. Furthermore, we deployed the DSCP approach for renewable energy generation and IT load forecasting in power management of a real-world trajectory-based application, achieving an 11.25% reduction in carbon emissions through predictive optimization of data center operations and controls.
comment: 28 pages, 13 figures, 3 tables. Submitted to Applied Soft Computing. With Editor This is the first public release of the work
☆ Improving $(α, f)$-Byzantine Resilience in Federated Learning via layerwise aggregation and cosine distance
The rapid development of artificial intelligence systems has amplified societal concerns regarding their usage, necessitating regulatory frameworks that encompass data privacy. Federated Learning (FL) is posed as potential solution to data privacy challenges in distributed machine learning by enabling collaborative model training {without data sharing}. However, FL systems remain vulnerable to Byzantine attacks, where malicious nodes contribute corrupted model updates. While Byzantine Resilient operators have emerged as a widely adopted robust aggregation algorithm to mitigate these attacks, its efficacy diminishes significantly in high-dimensional parameter spaces, sometimes leading to poor performing models. This paper introduces Layerwise Cosine Aggregation, a novel aggregation scheme designed to enhance robustness of these rules in such high-dimensional settings while preserving computational efficiency. A theoretical analysis is presented, demonstrating the superior robustness of the proposed Layerwise Cosine Aggregation compared to original robust aggregation operators. Empirical evaluation across diverse image classification datasets, under varying data distributions and Byzantine attack scenarios, consistently demonstrates the improved performance of Layerwise Cosine Aggregation, achieving up to a 16% increase in model accuracy.
comment: Submitted to Knowledge-Based Systems
☆ Feature-Enhanced Machine Learning for All-Cause Mortality Prediction in Healthcare Data
Accurate patient mortality prediction enables effective risk stratification, leading to personalized treatment plans and improved patient outcomes. However, predicting mortality in healthcare remains a significant challenge, with existing studies often focusing on specific diseases or limited predictor sets. This study evaluates machine learning models for all-cause in-hospital mortality prediction using the MIMIC-III database, employing a comprehensive feature engineering approach. Guided by clinical expertise and literature, we extracted key features such as vital signs (e.g., heart rate, blood pressure), laboratory results (e.g., creatinine, glucose), and demographic information. The Random Forest model achieved the highest performance with an AUC of 0.94, significantly outperforming other machine learning and deep learning approaches. This demonstrates Random Forest's robustness in handling high-dimensional, noisy clinical data and its potential for developing effective clinical decision support tools. Our findings highlight the importance of careful feature engineering for accurate mortality prediction. We conclude by discussing implications for clinical adoption and propose future directions, including enhancing model robustness and tailoring prediction models for specific diseases.
☆ Efficient Learning for Entropy-regularized Markov Decision Processes via Multilevel Monte Carlo
Designing efficient learning algorithms with complexity guarantees for Markov decision processes (MDPs) with large or continuous state and action spaces remains a fundamental challenge. We address this challenge for entropy-regularized MDPs with Polish state and action spaces, assuming access to a generative model of the environment. We propose a novel family of multilevel Monte Carlo (MLMC) algorithms that integrate fixed-point iteration with MLMC techniques and a generic stochastic approximation of the Bellman operator. We quantify the precise impact of the chosen approximate Bellman operator on the accuracy of the resulting MLMC estimator. Leveraging this error analysis, we show that using a biased plain MC estimate for the Bellman operator results in quasi-polynomial sample complexity, whereas an unbiased randomized multilevel approximation of the Bellman operator achieves polynomial sample complexity in expectation. Notably, these complexity bounds are independent of the dimensions or cardinalities of the state and action spaces, distinguishing our approach from existing algorithms whose complexities scale with the sizes of these spaces. We validate these theoretical performance guarantees through numerical experiments.
comment: 46 pages, 6 figures
☆ Rethinking Graph Structure Learning in the Era of LLMs
Recently, the emergence of large language models (LLMs) has prompted researchers to explore the integration of language descriptions into graphs, aiming to enhance model encoding capabilities from a data-centric perspective. This graph representation is called text-attributed graphs (TAGs). A review of prior advancements highlights that graph structure learning (GSL) is a pivotal technique for improving data utility, making it highly relevant to efficient TAG learning. However, most GSL methods are tailored for traditional graphs without textual information, underscoring the necessity of developing a new GSL paradigm. Despite clear motivations, it remains challenging: (1) How can we define a reasonable optimization objective for GSL in the era of LLMs, considering the massive parameters in LLM? (2) How can we design an efficient model architecture that enables seamless integration of LLM for this optimization objective? For Question 1, we reformulate existing GSL optimization objectives as a tree optimization framework, shifting the focus from obtaining a well-trained edge predictor to a language-aware tree sampler. For Question 2, we propose decoupled and training-free model design principles for LLM integration, shifting the focus from computation-intensive fine-tuning to more efficient inference. Based on this, we propose Large Language and Tree Assistant (LLaTA), which leverages tree-based LLM in-context learning to enhance the understanding of topology and text, enabling reliable inference and generating improved graph structure. Extensive experiments on 10 TAG datasets demonstrate that LLaTA enjoys flexibility - incorporated with any backbone; scalability - outperforms other LLM-based GSL methods in terms of running efficiency; effectiveness - achieves SOTA performance.
comment: 17 pages, 8 figures
☆ Resource-Efficient Federated Fine-Tuning Large Language Models for Heterogeneous Data
Fine-tuning large language models (LLMs) via federated learning, i.e., FedLLM, has been proposed to adapt LLMs for various downstream applications in a privacy-preserving way. To reduce the fine-tuning costs on resource-constrained devices, FedLoRA is proposed to fine-tune only a small subset of model parameters by integrating low-rank adaptation (LoRA) into FedLLM. However, apart from resource constraints, there is still another critical challenge, i.e., data heterogeneity, severely hindering the implementation of FedLoRA in practical applications. Herein, inspired by the previous group-based federated learning paradigm, we propose a hierarchical FedLoRA framework, termed HierFedLoRA, to address these challenges. Specifically, HierFedLoRA partitions all devices into multiple near-IID groups and adjusts the intra-group aggregation frequency for each group to eliminate the negative effects of non-IID data. Meanwhile, to reduce the computation and communication cost, HierFedLoRA dynamically assigns diverse and suitable fine-tuning depth (i.e., the number of continuous fine-tuning layers from the output) for each group. HierFedLoRA explores jointly optimizing aggregation frequency and depth upon their coupled relationship to better enhance the performance of FedLoRA. Extensive experiments are conducted on a physical platform with 80 commercial devices. The results show that HierFedLoRA improves the final model accuracy by 1.6% to 4.2%, speeding up the fine-tuning process by at least 2.1$\times$, compared to the strong baselines.
☆ Interpretable Cross-Sphere Multiscale Deep Learning Predicts ENSO Skilfully Beyond 2 Years
El Ni\~no-Southern Oscillation (ENSO) exerts global climate and societal impacts, but real-time prediction with lead times beyond one year remains challenging. Dynamical models suffer from large biases and uncertainties, while deep learning struggles with interpretability and multi-scale dynamics. Here, we introduce PTSTnet, an interpretable model that unifies dynamical processes and cross-scale spatiotemporal learning in an innovative neural-network framework with physics-encoding learning. PTSTnet produces interpretable predictions significantly outperforming state-of-the-art benchmarks with lead times beyond 24 months, providing physical insights into error propagation in ocean-atmosphere interactions. PTSTnet learns feature representations with physical consistency from sparse data to tackle inherent multi-scale and multi-physics challenges underlying ocean-atmosphere processes, thereby inherently enhancing long-term prediction skill. Our successful realizations mark substantial steps forward in interpretable insights into innovative neural ocean modelling.
comment: 13 pages, 4 figures
☆ Learning Generalizable Skills from Offline Multi-Task Data for Multi-Agent Cooperation
Learning cooperative multi-agent policy from offline multi-task data that can generalize to unseen tasks with varying numbers of agents and targets is an attractive problem in many scenarios. Although aggregating general behavior patterns among multiple tasks as skills to improve policy transfer is a promising approach, two primary challenges hinder the further advancement of skill learning in offline multi-task MARL. Firstly, extracting general cooperative behaviors from various action sequences as common skills lacks bringing cooperative temporal knowledge into them. Secondly, existing works only involve common skills and can not adaptively choose independent knowledge as task-specific skills in each task for fine-grained action execution. To tackle these challenges, we propose Hierarchical and Separate Skill Discovery (HiSSD), a novel approach for generalizable offline multi-task MARL through skill learning. HiSSD leverages a hierarchical framework that jointly learns common and task-specific skills. The common skills learn cooperative temporal knowledge and enable in-sample exploitation for offline multi-task MARL. The task-specific skills represent the priors of each task and achieve a task-guided fine-grained action execution. To verify the advancement of our method, we conduct experiments on multi-agent MuJoCo and SMAC benchmarks. After training the policy using HiSSD on offline multi-task data, the empirical results show that HiSSD assigns effective cooperative behaviors and obtains superior performance in unseen tasks.
☆ Unveiling the Potential of Superexpressive Networks in Implicit Neural Representations ICLR 2025
In this study, we examine the potential of one of the ``superexpressive'' networks in the context of learning neural functions for representing complex signals and performing machine learning downstream tasks. Our focus is on evaluating their performance on computer vision and scientific machine learning tasks including signal representation/inverse problems and solutions of partial differential equations. Through an empirical investigation in various benchmark tasks, we demonstrate that superexpressive networks, as proposed by [Zhang et al. NeurIPS, 2022], which employ a specialized network structure characterized by having an additional dimension, namely width, depth, and ``height'', can surpass recent implicit neural representations that use highly-specialized nonlinear activation functions.
comment: Accepted at ICLR 2025 Workshop on Neural Network Weights as a New Data Modality
☆ Adversarial Wear and Tear: Exploiting Natural Damage for Generating Physical-World Adversarial Examples
The presence of adversarial examples in the physical world poses significant challenges to the deployment of Deep Neural Networks in safety-critical applications such as autonomous driving. Most existing methods for crafting physical-world adversarial examples are ad-hoc, relying on temporary modifications like shadows, laser beams, or stickers that are tailored to specific scenarios. In this paper, we introduce a new class of physical-world adversarial examples, AdvWT, which draws inspiration from the naturally occurring phenomenon of `wear and tear', an inherent property of physical objects. Unlike manually crafted perturbations, `wear and tear' emerges organically over time due to environmental degradation, as seen in the gradual deterioration of outdoor signboards. To achieve this, AdvWT follows a two-step approach. First, a GAN-based, unsupervised image-to-image translation network is employed to model these naturally occurring damages, particularly in the context of outdoor signboards. The translation network encodes the characteristics of damaged signs into a latent `damage style code'. In the second step, we introduce adversarial perturbations into the style code, strategically optimizing its transformation process. This manipulation subtly alters the damage style representation, guiding the network to generate adversarial images where the appearance of damages remains perceptually realistic, while simultaneously ensuring their effectiveness in misleading neural networks. Through comprehensive experiments on two traffic sign datasets, we show that AdvWT effectively misleads DNNs in both digital and physical domains. AdvWT achieves an effective attack success rate, greater robustness, and a more natural appearance compared to existing physical-world adversarial examples. Additionally, integrating AdvWT into training enhances a model's generalizability to real-world damaged signs.
comment: 11 pages, 9 figures
☆ A Data Balancing and Ensemble Learning Approach for Credit Card Fraud Detection
This research introduces an innovative method for identifying credit card fraud by combining the SMOTE-KMEANS technique with an ensemble machine learning model. The proposed model was benchmarked against traditional models such as logistic regression, decision trees, random forests, and support vector machines. Performance was evaluated using metrics, including accuracy, recall, and area under the curve (AUC). The results demonstrated that the proposed model achieved superior performance, with an AUC of 0.96 when combined with the SMOTE-KMEANS algorithm. This indicates a significant improvement in detecting fraudulent transactions while maintaining high precision and recall. The study also explores the application of different oversampling techniques to enhance the performance of various classifiers. The findings suggest that the proposed method is robust and effective for classification tasks on balanced datasets. Future research directions include further optimization of the SMOTE-KMEANS approach and its integration into existing fraud detection systems to enhance financial security and consumer protection.
☆ Multi-Objective Optimization for Privacy-Utility Balance in Differentially Private Federated Learning
Federated learning (FL) enables collaborative model training across distributed clients without sharing raw data, making it a promising approach for privacy-preserving machine learning. However, ensuring differential privacy (DP) in FL presents challenges due to the trade-off between model utility and privacy protection. Clipping gradients before aggregation is a common strategy to limit privacy loss, but selecting an optimal clipping norm is non-trivial, as excessively high values compromise privacy, while overly restrictive clipping degrades model performance. In this work, we propose an adaptive clipping mechanism that dynamically adjusts the clipping norm using a multi-objective optimization framework. By integrating privacy and utility considerations into the optimization objective, our approach balances privacy preservation with model accuracy. We theoretically analyze the convergence properties of our method and demonstrate its effectiveness through extensive experiments on MNIST, Fashion-MNIST, and CIFAR-10 datasets. Our results show that adaptive clipping consistently outperforms fixed-clipping baselines, achieving improved accuracy under the same privacy constraints. This work highlights the potential of dynamic clipping strategies to enhance privacy-utility trade-offs in differentially private federated learning.
☆ Real-Time Evaluation Models for RAG: Who Detects Hallucinations Best?
This article surveys Evaluation models to automatically detect hallucinations in Retrieval-Augmented Generation (RAG), and presents a comprehensive benchmark of their performance across six RAG applications. Methods included in our study include: LLM-as-a-Judge, Prometheus, Lynx, the Hughes Hallucination Evaluation Model (HHEM), and the Trustworthy Language Model (TLM). These approaches are all reference-free, requiring no ground-truth answers/labels to catch incorrect LLM responses. Our study reveals that, across diverse RAG applications, some of these approaches consistently detect incorrect RAG responses with high precision/recall.
comment: 11 pages, 8 figures
☆ Embedding Domain-Specific Knowledge from LLMs into the Feature Engineering Pipeline
Feature engineering is mandatory in the machine learning pipeline to obtain robust models. While evolutionary computation is well-known for its great results both in feature selection and feature construction, its methods are computationally expensive due to the large number of evaluations required to induce the final model. Part of the reason why these algorithms require a large number of evaluations is their lack of domain-specific knowledge, resulting in a lot of random guessing during evolution. In this work, we propose using Large Language Models (LLMs) as an initial feature construction step to add knowledge to the dataset. By doing so, our results show that the evolution can converge faster, saving us computational resources. The proposed approach only provides the names of the features in the dataset and the target objective to the LLM, making it usable even when working with datasets containing private data. While consistent improvements to test performance were only observed for one-third of the datasets (CSS, PM, and IM10), possibly due to problems being easily explored by LLMs, this approach only decreased the model performance in 1/77 test cases. Additionally, this work introduces the M6GP feature engineering algorithm to symbolic regression, showing it can improve the results of the random forest regressor and produce competitive results with its predecessor, M3GP.
comment: 9 pages, 4 figures, 5 tables
☆ Federated Learning with Differential Privacy: An Utility-Enhanced Approach
Federated learning has emerged as an attractive approach to protect data privacy by eliminating the need for sharing clients' data while reducing communication costs compared with centralized machine learning algorithms. However, recent studies have shown that federated learning alone does not guarantee privacy, as private data may still be inferred from the uploaded parameters to the central server. In order to successfully avoid data leakage, adopting differential privacy (DP) in the local optimization process or in the local update aggregation process has emerged as two feasible ways for achieving sample-level or user-level privacy guarantees respectively, in federated learning models. However, compared to their non-private equivalents, these approaches suffer from a poor utility. To improve the privacy-utility trade-off, we present a modification to these vanilla differentially private algorithms based on a Haar wavelet transformation step and a novel noise injection scheme that significantly lowers the asymptotic bound of the noise variance. We also present a holistic convergence analysis of our proposed algorithm, showing that our method yields better convergence performance than the vanilla DP algorithms. Numerical experiments on real-world datasets demonstrate that our method outperforms existing approaches in model utility while maintaining the same privacy guarantees.
☆ A computational theory of evaluation for parameterisable subject
Evaluation is critical to advance decision making across domains, yet existing methodologies often struggle to balance theoretical rigor and practical scalability. In order to reduce the cost of experimental evaluation, we introduce a computational theory of evaluation for parameterisable subjects. We prove upper bounds of generalized evaluation error and generalized causal effect error of evaluation metric on subject. We also prove efficiency, and consistency to estimated causal effect of subject on metric by prediction. To optimize evaluation models, we propose a meta-learner to handle heterogeneous evaluation subjects space. Comparing with other computational approaches, our (conditional) evaluation model reduced 24.1%-99.0% evaluation errors across 12 scenes, including individual medicine, scientific simulation, business activities, and quantum trade. The evaluation time is reduced 3-7 order of magnitude comparing with experiments or simulations.
☆ MoQa: Rethinking MoE Quantization with Multi-stage Data-model Distribution Awareness
With the advances in artificial intelligence, Mix-of-Experts (MoE) has become the main form of Large Language Models (LLMs), and its demand for model compression is increasing. Quantization is an effective method that not only compresses the models but also significantly accelerates their performance. Existing quantization methods have gradually shifted the focus from parameter scaling to the analysis of data distributions. However, their analysis is designed for dense LLMs and relies on the simple one-model-all-data mapping, which is unsuitable for MoEs. This paper proposes a new quantization framework called MoQa. MoQa decouples the data-model distribution complexity of MoEs in multiple analysis stages, quantitively revealing the dynamics during sparse data activation, data-parameter mapping, and inter-expert correlations. Based on these, MoQa identifies particular experts' and parameters' significance with optimal data-model distribution awareness and proposes a series of fine-grained mix-quantization strategies adaptive to various data activation and expert combination scenarios. Moreover, MoQa discusses the limitations of existing quantization and analyzes the impact of each stage analysis, showing novel insights for MoE quantization. Experiments show that MoQa achieves a 1.69~2.18 perplexity decrease in language modeling tasks and a 1.58%~8.91% accuracy improvement in zero-shot inference tasks. We believe MoQa will play a role in future MoE construction, optimization, and compression.
comment: 6 pages, 6 figures and 3 tables
☆ Squared families: Searching beyond regular probability models
We introduce squared families, which are families of probability densities obtained by squaring a linear transformation of a statistic. Squared families are singular, however their singularity can easily be handled so that they form regular models. After handling the singularity, squared families possess many convenient properties. Their Fisher information is a conformal transformation of the Hessian metric induced from a Bregman generator. The Bregman generator is the normalising constant, and yields a statistical divergence on the family. The normalising constant admits a helpful parameter-integral factorisation, meaning that only one parameter-independent integral needs to be computed for all normalising constants in the family, unlike in exponential families. Finally, the squared family kernel is the only integral that needs to be computed for the Fisher information, statistical divergence and normalising constant. We then describe how squared families are special in the broader class of $g$-families, which are obtained by applying a sufficiently regular function $g$ to a linear transformation of a statistic. After removing special singularities, positively homogeneous families and exponential families are the only $g$-families for which the Fisher information is a conformal transformation of the Hessian metric, where the generator depends on the parameter only through the normalising constant. Even-order monomial families also admit parameter-integral factorisations, unlike exponential families. We study parameter estimation and density estimation in squared families, in the well-specified and misspecified settings. We use a universal approximation property to show that squared families can learn sufficiently well-behaved target densities at a rate of $\mathcal{O}(N^{-1/2})+C n^{-1/4}$, where $N$ is the number of datapoints, $n$ is the number of parameters, and $C$ is some constant.
comment: 43 pages. Preprint
☆ AugWard: Augmentation-Aware Representation Learning for Accurate Graph Classification PAKDD 2025
How can we accurately classify graphs? Graph classification is a pivotal task in data mining with applications in social network analysis, web analysis, drug discovery, molecular property prediction, etc. Graph neural networks have achieved the state-of-the-art performance in graph classification, but they consistently struggle with overfitting. To mitigate overfitting, researchers have introduced various representation learning methods utilizing graph augmentation. However, existing methods rely on simplistic use of graph augmentation, which loses augmentation-induced differences and limits the expressiveness of representations. In this paper, we propose AugWard (Augmentation-Aware Training with Graph Distance and Consistency Regularization), a novel graph representation learning framework that carefully considers the diversity introduced by graph augmentation. AugWard applies augmentation-aware training to predict the graph distance between the augmented graph and its original one, aligning the representation difference directly with graph distance at both feature and structure levels. Furthermore, AugWard employs consistency regularization to encourage the classifier to handle richer representations. Experimental results show that AugWard gives the state-of-the-art performance in supervised, semi-supervised graph classification, and transfer learning.
comment: Accepted to PAKDD 2025 (Oral Presentation)
☆ Low Stein Discrepancy via Message-Passing Monte Carlo ICLR 2025
Message-Passing Monte Carlo (MPMC) was recently introduced as a novel low-discrepancy sampling approach leveraging tools from geometric deep learning. While originally designed for generating uniform point sets, we extend this framework to sample from general multivariate probability distributions with known probability density function. Our proposed method, Stein-Message-Passing Monte Carlo (Stein-MPMC), minimizes a kernelized Stein discrepancy, ensuring improved sample quality. Finally, we show that Stein-MPMC outperforms competing methods, such as Stein Variational Gradient Descent and (greedy) Stein Points, by achieving a lower Stein discrepancy.
comment: 8 pages, 2 figures, Accepted at the ICLR 2025 Workshop on Frontiers in Probabilistic Inference
☆ Confidence Adjusted Surprise Measure for Active Resourceful Trials (CA-SMART): A Data-driven Active Learning Framework for Accelerating Material Discovery under Resource Constraints
Accelerating the discovery and manufacturing of advanced materials with specific properties is a critical yet formidable challenge due to vast search space, high costs of experiments, and time-intensive nature of material characterization. In recent years, active learning, where a surrogate machine learning (ML) model mimics the scientific discovery process of a human scientist, has emerged as a promising approach to address these challenges by guiding experimentation toward high-value outcomes with a limited budget. Among the diverse active learning philosophies, the concept of surprise (capturing the divergence between expected and observed outcomes) has demonstrated significant potential to drive experimental trials and refine predictive models. Scientific discovery often stems from surprise thereby making it a natural driver to guide the search process. Despite its promise, prior studies leveraging surprise metrics such as Shannon and Bayesian surprise lack mechanisms to account for prior confidence, leading to excessive exploration of uncertain regions that may not yield useful information. To address this, we propose the Confidence-Adjusted Surprise Measure for Active Resourceful Trials (CA-SMART), a novel Bayesian active learning framework tailored for optimizing data-driven experimentation. On a high level, CA-SMART incorporates Confidence-Adjusted Surprise (CAS) to dynamically balance exploration and exploitation by amplifying surprises in regions where the model is more certain while discounting them in highly uncertain areas. We evaluated CA-SMART on two benchmark functions (Six-Hump Camelback and Griewank) and in predicting the fatigue strength of steel. The results demonstrate superior accuracy and efficiency compared to traditional surprise metrics, standard Bayesian Optimization (BO) acquisition functions and conventional ML methods.
☆ ZJUKLAB at SemEval-2025 Task 4: Unlearning via Model Merging
This paper presents the ZJUKLAB team's submission for SemEval-2025 Task 4: Unlearning Sensitive Content from Large Language Models. This task aims to selectively erase sensitive knowledge from large language models, avoiding both over-forgetting and under-forgetting issues. We propose an unlearning system that leverages Model Merging (specifically TIES-Merging), combining two specialized models into a more balanced unlearned model. Our system achieves competitive results, ranking second among 26 teams, with an online score of 0.944 for Task Aggregate and 0.487 for overall Aggregate. In this paper, we also conduct local experiments and perform a comprehensive analysis of the unlearning process, examining performance trajectories, loss dynamics, and weight perspectives, along with several supplementary experiments, to understand the effectiveness of our method. Furthermore, we analyze the shortcomings of our method and evaluation metrics, emphasizing that MIA scores and ROUGE-based metrics alone are insufficient to fully evaluate successful unlearning. Finally, we emphasize the need for more comprehensive evaluation methodologies and rethinking of unlearning objectives in future research. Code is available at https://github.com/zjunlp/unlearn/tree/main/semeval25.
comment: Work in progress
☆ Geographical hotspot prediction based on point cloud-voxel-community partition clustering
Existing solutions to the hotspot prediction problem in the field of geographic information remain at a relatively preliminary stage. This study presents a novel approach for detecting and predicting geographical hotspots, utilizing point cloud-voxel-community partition clustering. By analyzing high-dimensional data, we represent spatial information through point clouds, which are then subdivided into multiple voxels to enhance analytical efficiency. Our method identifies spatial voxels with similar characteristics through community partitioning, thereby revealing underlying patterns in hotspot distributions. Experimental results indicate that when applied to a dataset of archaeological sites in Turkey, our approach achieves a 19.31% increase in processing speed, with an accuracy loss of merely 6%, outperforming traditional clustering methods. This method not only provides a fresh perspective for hotspot prediction but also serves as an effective tool for high-dimensional data analysis.
☆ KAC: Kolmogorov-Arnold Classifier for Continual Learning CVPR 2025
Continual learning requires models to train continuously across consecutive tasks without forgetting. Most existing methods utilize linear classifiers, which struggle to maintain a stable classification space while learning new tasks. Inspired by the success of Kolmogorov-Arnold Networks (KAN) in preserving learning stability during simple continual regression tasks, we set out to explore their potential in more complex continual learning scenarios. In this paper, we introduce the Kolmogorov-Arnold Classifier (KAC), a novel classifier developed for continual learning based on the KAN structure. We delve into the impact of KAN's spline functions and introduce Radial Basis Functions (RBF) for improved compatibility with continual learning. We replace linear classifiers with KAC in several recent approaches and conduct experiments across various continual learning benchmarks, all of which demonstrate performance improvements, highlighting the effectiveness and robustness of KAC in continual learning. The code is available at https://github.com/Ethanhuhuhu/KAC.
comment: CVPR 2025
☆ Rerouting Connection: Hybrid Computer Vision Analysis Reveals Visual Similarity Between Indus and Tibetan-Yi Corridor Writing Systems
This thesis employs a hybrid CNN-Transformer architecture, in conjunction with a detailed anthropological framework, to investigate potential historical connections between the visual morphology of the Indus Valley script and pictographic systems of the Tibetan-Yi Corridor. Through an ensemble methodology of three target scripts across 15 independently trained models, we demonstrate that Tibetan-Yi Corridor scripts exhibit approximately six-fold higher visual similarity to the Indus script (61.7%-63.5%) than to the Bronze Age Proto-Cuneiform (10.2%-10.9%) or Proto-Elamite (7.6%-8.7%) systems. Additionally and contrarily to our current understanding of the networks of the Indus Valley Civilization, the Indus script unexpectedly maps closer to Tibetan-Yi Corridor scripts, with a mean cosine similarity of 0.629, than to the aforementioned contemporaneous West Asian signaries, both of which recorded mean cosine similarities of 0.104 and 0.080 despite their close geographic proximity and evident trade relations. Across various dimensionality reduction practices and clustering methodologies, the Indus script consistently clusters closest to Tibetan-Yi Corridor scripts. Our computational results align with qualitative observations of specific pictorial parallels in numeral systems, gender markers, and key iconographic elements; this is further supported by archaeological evidence of sustained contact networks along the ancient Shu-Shendu road in tandem with the Indus Valley Civilization's decline, providing a plausible transmission pathway. While alternative explanations cannot be ruled out, the specificity and consistency of observed similarities challenge conventional narratives of isolated script development and suggest more complex ancient cultural transmission networks between South and East Asia than previously recognized.
comment: 106 pages total (main text: 42, 48 w/refs, 100 w/appendices). 21 figures, 4 tables in main; 106 figs, 8 tables total. Code and data at this URL: https://github.com/oohalakkadi/ivc2tyc. Submitted as undergrad thesis at Duke Kunshan University; accepted for presentation at the 2025 Computer Applications and Quantitative Methods in Archaeology Conference, Athens
☆ Shared Global and Local Geometry of Language Model Embeddings
Researchers have recently suggested that models share common representations. In this work, we find that the token embeddings of language models exhibit common geometric structure. First, we find ``global'' similarities: token embeddings often share similar relative orientations. Next, we characterize local geometry in two ways: (1) by using Locally Linear Embeddings, and (2) by defining a simple measure for the intrinsic dimension of each token embedding. Our intrinsic dimension measure demonstrates that token embeddings lie on a lower dimensional manifold. We qualitatively show that tokens with lower intrinsic dimensions often have semantically coherent clusters, while those with higher intrinsic dimensions do not. Both characterizations allow us to find similarities in the local geometry of token embeddings. Perhaps most surprisingly, we find that alignment in token embeddings persists through the hidden states of language models, allowing us to develop an application for interpretability. Namely, we empirically demonstrate that steering vectors from one language model can be transferred to another, despite the two models having different dimensions.
☆ Purifying Approximate Differential Privacy with Randomized Post-processing
We propose a framework to convert $(\varepsilon, \delta)$-approximate Differential Privacy (DP) mechanisms into $(\varepsilon, 0)$-pure DP mechanisms, a process we call ``purification''. This algorithmic technique leverages randomized post-processing with calibrated noise to eliminate the $\delta$ parameter while preserving utility. By combining the tighter utility bounds and computational efficiency of approximate DP mechanisms with the stronger guarantees of pure DP, our approach achieves the best of both worlds. We illustrate the applicability of this framework in various settings, including Differentially Private Empirical Risk Minimization (DP-ERM), data-dependent DP mechanisms such as Propose-Test-Release (PTR), and query release tasks. To the best of our knowledge, this is the first work to provide a systematic method for transforming approximate DP into pure DP while maintaining competitive accuracy and computational efficiency.
☆ Uncertainty propagation in feed-forward neural network models
We develop new uncertainty propagation methods for feed-forward neural network architectures with leaky ReLu activation functions subject to random perturbations in the input vectors. In particular, we derive analytical expressions for the probability density function (PDF) of the neural network output and its statistical moments as a function of the input uncertainty and the parameters of the network, i.e., weights and biases. A key finding is that an appropriate linearization of the leaky ReLu activation function yields accurate statistical results even for large perturbations in the input vectors. This can be attributed to the way information propagates through the network. We also propose new analytically tractable Gaussian copula surrogate models to approximate the full joint PDF of the neural network output. To validate our theorical results, we conduct Monte Carlo simulations and a thorough error analysis on a multi-layer neural network representing a nonlinear integro-differential operator between two polynomial function spaces. Our findings demonstrate excellent agreement between the theoretical predictions and Monte Carlo simulations.
comment: 21 pages, 13 figures
♻ ☆ Partial Gromov-Wasserstein Metric ICLR 2025
The Gromov-Wasserstein (GW) distance has gained increasing interest in the machine learning community in recent years, as it allows for the comparison of measures in different metric spaces. To overcome the limitations imposed by the equal mass requirements of the classical GW problem, researchers have begun exploring its application in unbalanced settings. However, Unbalanced GW (UGW) can only be regarded as a discrepancy rather than a rigorous metric/distance between two metric measure spaces (mm-spaces). In this paper, we propose a particular case of the UGW problem, termed Partial Gromov-Wasserstein (PGW). We establish that PGW is a well-defined metric between mm-spaces and discuss its theoretical properties, including the existence of a minimizer for the PGW problem and the relationship between PGW and GW, among others. We then propose two variants of the Frank-Wolfe algorithm for solving the PGW problem and show that they are mathematically and computationally equivalent. Moreover, based on our PGW metric, we introduce the analogous concept of barycenters for mm-spaces. Finally, we validate the effectiveness of our PGW metric and related solvers in applications such as shape matching, shape retrieval, and shape interpolation, comparing them against existing baselines. Our code is available at https://github.com/mint-vu/PGW_Metric.
comment: Published at ICLR 2025
♻ ☆ GenoTEX: A Benchmark for Automated Gene Expression Data Analysis in Alignment with Bioinformaticians
Recent advancements in machine learning have significantly improved the identification of disease-associated genes from gene expression datasets. However, these processes often require extensive expertise and manual effort, limiting their scalability. Large Language Model (LLM)-based agents have shown promise in automating these tasks due to their increasing problem-solving abilities. To support the evaluation and development of such methods, we introduce GenoTEX, a benchmark dataset for the automated analysis of gene expression data. GenoTEX provides annotated code and results for solving a wide range of gene identification problems, encompassing dataset selection, preprocessing, and statistical analysis, in a pipeline that follows computational genomics standards. The benchmark includes expert-curated annotations from bioinformaticians to ensure accuracy and reliability. To provide baselines for these tasks, we present GenoAgent, a team of LLM-based agents that adopt a multi-step programming workflow with flexible self-correction, to collaboratively analyze gene expression datasets. Our experiments demonstrate the potential of LLM-based methods in analyzing genomic data, while error analysis highlights the challenges and areas for future improvement. We propose GenoTEX as a promising resource for benchmarking and enhancing automated methods for gene expression data analysis. The benchmark is available at https://github.com/Liu-Hy/GenoTex.
comment: 29 pages, 3 figures
♻ ☆ A Selective Homomorphic Encryption Approach for Faster Privacy-Preserving Federated Learning
Federated learning is a machine learning method that supports training models on decentralized devices or servers, where each holds its local data, removing the need for data exchange. This approach is especially useful in healthcare, as it enables training on sensitive data without needing to share them. The nature of federated learning necessitates robust security precautions due to data leakage concerns during communication. To address this issue, we propose a new approach that employs selective encryption, homomorphic encryption, differential privacy, and bit-wise scrambling to minimize data leakage while achieving good execution performance. Our technique , FAS (fast and secure federated learning) is used to train deep learning models on medical imaging data. We implemented our technique using the Flower framework and compared with a state-of-the-art federated learning approach that also uses selective homomorphic encryption. Our experiments were run in a cluster of eleven physical machines to create a real-world federated learning scenario on different datasets. We observed that our approach is up to 90\% faster than applying fully homomorphic encryption on the model weights. In addition, we can avoid the pretraining step that is required by our competitor and can save up to 46% in terms of total execution time. While our approach was faster, it obtained similar security results as the competitor.
comment: 23 pages, 32 figures
♻ ☆ Multi-View and Multi-Scale Alignment for Contrastive Language-Image Pre-training in Mammography
Contrastive Language-Image Pre-training (CLIP) demonstrates strong potential in medical image analysis but requires substantial data and computational resources. Due to these restrictions, existing CLIP applications in medical imaging focus mainly on modalities like chest X-rays that have abundant image-report data available, leaving many other important modalities underexplored. Here, we propose one of the first adaptations of the full CLIP model to mammography, which presents significant challenges due to labeled data scarcity, high-resolution images with small regions of interest, and class-wise imbalance. We first develop a specialized supervision framework for mammography that leverages its multi-view nature. Furthermore, we design a symmetric local alignment module to better focus on detailed features in high-resolution images. Lastly, we incorporate a parameter-efficient fine-tuning approach for large language models pre-trained with medical knowledge to address data limitations. Our multi-view and multi-scale alignment (MaMA) method outperforms state-of-the-art baselines for three different tasks on two large real-world mammography datasets, EMBED and RSNA-Mammo, with only 52% model size compared with the largest baseline. The code is available at https://github.com/XYPB/MaMA
comment: This paper is accepted by IPMI 2025 for Oral Presentation
♻ ☆ PVLens: Enhancing Pharmacovigilance Through Automated Label Extraction
Reliable drug safety reference databases are essential for pharmacovigilance, yet existing resources like SIDER are outdated and static. We introduce PVLens, an automated system that extracts labeled safety information from FDA Structured Product Labels (SPLs) and maps terms to MedDRA. PVLens integrates automation with expert oversight through a web-based review tool. In validation against 97 drug labels, PVLens achieved an F1 score of 0.882, with high recall (0.983) and moderate precision (0.799). By offering a scalable, more accurate and continuously updated alternative to SIDER, PVLens enhances real-time pharamcovigilance with improved accuracy and contemporaneous insights.
♻ ☆ TSKANMixer: Kolmogorov-Arnold Networks with MLP-Mixer Model for Time Series Forecasting AAAI 2025
Time series forecasting has long been a focus of research across diverse fields, including economics, energy, healthcare, and traffic management. Recent works have introduced innovative architectures for time series models, such as the Time-Series Mixer (TSMixer), which leverages multi-layer perceptrons (MLPs) to enhance prediction accuracy by effectively capturing both spatial and temporal dependencies within the data. In this paper, we investigate the capabilities of the Kolmogorov-Arnold Networks (KANs) for time-series forecasting by modifying TSMixer with a KAN layer (TSKANMixer). Experimental results demonstrate that TSKANMixer tends to improve prediction accuracy over the original TSMixer across multiple datasets, ranking among the top-performing models compared to other time series approaches. Our results show that the KANs are promising alternatives to improve the performance of time series forecasting by replacing or extending traditional MLPs.
comment: 8 pages, 4 figures, 7 tables and accepted at the AI4TS: AI for Time Series Analysis workshop, AAAI 2025
♻ ☆ Denoising VAE as an Explainable Feature Reduction and Diagnostic Pipeline for Autism Based on Resting state fMRI
Autism spectrum disorders (ASDs) are developmental conditions characterized by restricted interests and difficulties in communication. The complexity of ASD has resulted in a deficiency of objective diagnostic biomarkers. Deep learning methods have gained recognition for addressing these challenges in neuroimaging analysis, but finding and interpreting such diagnostic biomarkers are still challenging computationally. Here, we propose a feature reduction pipeline using resting-state fMRI data. We used Craddock atlas and Power atlas to extract functional connectivity data from rs-fMRI, resulting in over 30 thousand features. By using a denoising variational autoencoder, our proposed pipeline further compresses the connectivity features into 5 latent Gaussian distributions, providing is a low-dimensional representation of the data to promote computational efficiency and interpretability. To test the method, we employed the extracted latent representations to classify ASD using traditional classifiers such as SVM on a large multi-site dataset. The 95% confidence interval for the prediction accuracy of SVM is [0.63, 0.76] after site harmonization using the extracted latent distributions. Without using DVAE for dimensionality reduction, the prediction accuracy is 0.70, which falls within the interval. The DVAE successfully encoded the diagnostic information from rs-fMRI data without sacrificing prediction performance. The runtime for training the DVAE and obtaining classification results from its extracted latent features was 7 times shorter compared to training classifiers directly on the raw data. Our findings suggest that the Power atlas provides more effective brain connectivity insights for diagnosing ASD than Craddock atlas. Additionally, we visualized the latent representations to gain insights into the brain networks contributing to the differences between ASD and neurotypical brains.
♻ ☆ A Context-Aware Approach for Enhancing Data Imputation with Pre-trained Language Models
This paper presents a novel approach named \textbf{C}ontextually \textbf{R}elevant \textbf{I}mputation leveraging pre-trained \textbf{L}anguage \textbf{M}odels (\textbf{CRILM}) for handling missing data in tabular datasets. Instead of relying on traditional numerical estimations, CRILM uses pre-trained language models (LMs) to create contextually relevant descriptors for missing values. This method aligns datasets with LMs' strengths, allowing large LMs to generate these descriptors and small LMs to be fine-tuned on the enriched datasets for enhanced downstream task performance. Our evaluations demonstrate CRILM's superior performance and robustness across MCAR, MAR, and challenging MNAR scenarios, with up to a 10\% improvement over the best-performing baselines. By mitigating biases, particularly in MNAR settings, CRILM improves downstream task performance and offers a cost-effective solution for resource-constrained environments.
♻ ☆ An Exponential Separation Between Quantum and Quantum-Inspired Classical Algorithms for Linear Systems
Achieving a provable exponential quantum speedup for an important machine learning task has been a central research goal since the seminal HHL quantum algorithm for solving linear systems and the subsequent quantum recommender systems algorithm by Kerenidis and Prakash. These algorithms were initially believed to be strong candidates for exponential speedups, but a lower bound ruling out similar classical improvements remained absent. In breakthrough work by Tang, it was demonstrated that this lack of progress in classical lower bounds was for good reasons. Concretely, she gave a classical counterpart of the quantum recommender systems algorithm, reducing the quantum advantage to a mere polynomial. Her approach is quite general and was named quantum-inspired classical algorithms. Since then, almost all the initially exponential quantum machine learning speedups have been reduced to polynomial via new quantum-inspired classical algorithms. From the current state-of-affairs, it is unclear whether we can hope for exponential quantum speedups for any natural machine learning task. In this work, we present the first such provable exponential separation between quantum and quantum-inspired classical algorithms for the basic problem of solving a linear system when the input matrix is well-conditioned and has sparse rows and columns.
♻ ☆ Self-Contrastive Forward-Forward Algorithm
Agents that operate autonomously benefit from lifelong learning capabilities. However, compatible training algorithms must comply with the decentralized nature of these systems, which imposes constraints on both the parameter counts and the computational resources. The Forward-Forward (FF) algorithm is one of these. FF relies only on feedforward operations, the same used for inference, for optimizing layer-wise objectives. This purely forward approach eliminates the need for transpose operations required in traditional backpropagation. Despite its potential, FF has failed to reach state-of-the-art performance on most standard benchmark tasks, in part due to unreliable negative data generation methods for unsupervised learning. In this work, we propose the Self-Contrastive Forward-Forward (SCFF) algorithm, a competitive training method aimed at closing this performance gap. Inspired by standard self-supervised contrastive learning for vision tasks, SCFF generates positive and negative inputs applicable across various datasets. The method demonstrates superior performance compared to existing unsupervised local learning algorithms on several benchmark datasets, including MNIST, CIFAR-10, STL-10, and Tiny ImageNet. We extend FF's application to training recurrent neural networks, expanding its utility to sequential data tasks. These findings pave the way for high-accuracy, real-time learning on resource-constrained edge devices.
♻ ☆ Graph Anomaly Detection in Time Series: A Survey
With the recent advances in technology, a wide range of systems continue to collect a large amount of data over time and thus generate time series. Time-Series Anomaly Detection (TSAD) is an important task in various time-series applications such as e-commerce, cybersecurity, vehicle maintenance, and healthcare monitoring. However, this task is very challenging as it requires considering both the intra-variable dependency (relationships within a variable over time) and the inter-variable dependency (relationships between multiple variables) existing in time-series data. Recent graph-based approaches have made impressive progress in tackling the challenges of this field. In this survey, we conduct a comprehensive and up-to-date review of TSAD using graphs, referred to as G-TSAD. First, we explore the significant potential of graph representation for time-series data and and its contributions to facilitating anomaly detection. Then, we review state-of-the-art graph anomaly detection techniques, mostly leveraging deep learning architectures, in the context of time series. For each method, we discuss its strengths, limitations, and the specific applications where it excels. Finally, we address both the technical and application challenges currently facing the field, and suggest potential future directions for advancing research and improving practical outcomes.
comment: 20 pages, 7 figures, 6 tables
♻ ☆ Layer- and Timestep-Adaptive Differentiable Token Compression Ratios for Efficient Diffusion Transformers CVPR 2025
Diffusion Transformers (DiTs) have achieved state-of-the-art (SOTA) image generation quality but suffer from high latency and memory inefficiency, making them difficult to deploy on resource-constrained devices. One major efficiency bottleneck is that existing DiTs apply equal computation across all regions of an image. However, not all image tokens are equally important, and certain localized areas require more computation, such as objects. To address this, we propose DiffCR, a dynamic DiT inference framework with differentiable compression ratios, which automatically learns to dynamically route computation across layers and timesteps for each image token, resulting in efficient DiTs. Specifically, DiffCR integrates three features: (1) A token-level routing scheme where each DiT layer includes a router that is fine-tuned jointly with model weights to predict token importance scores. In this way, unimportant tokens bypass the entire layer's computation; (2) A layer-wise differentiable ratio mechanism where different DiT layers automatically learn varying compression ratios from a zero initialization, resulting in large compression ratios in redundant layers while others remain less compressed or even uncompressed; (3) A timestep-wise differentiable ratio mechanism where each denoising timestep learns its own compression ratio. The resulting pattern shows higher ratios for noisier timesteps and lower ratios as the image becomes clearer. Extensive experiments on text-to-image and inpainting tasks show that DiffCR effectively captures dynamism across token, layer, and timestep axes, achieving superior trade-offs between generation quality and efficiency compared to prior works. The project website is available at https://www.haoranyou.com/diffcr.
comment: Accepted by CVPR 2025
♻ ☆ Stochastic Inference of Plate Bending from Heterogeneous Data: Physics-informed Gaussian Processes via Kirchhoff-Love Theory
Advancements in machine learning and an abundance of structural monitoring data have inspired the integration of mechanical models with probabilistic models to identify a structure's state and quantify the uncertainty of its physical parameters and response. In this paper, we propose an inference methodology for classical Kirchhoff-Love plates via physics-informed Gaussian Processes (GP). A probabilistic model is formulated as a multi-output GP by placing a GP prior on the deflection and deriving the covariance function using the linear differential operators of the plate governing equations. The posteriors of the flexural rigidity, hyperparameters, and plate response are inferred in a Bayesian manner using Markov chain Monte Carlo (MCMC) sampling from noisy measurements. We demonstrate the applicability with two examples: a simply supported plate subjected to a sinusoidal load and a fixed plate subjected to a uniform load. The results illustrate how the proposed methodology can be employed to perform stochastic inference for plate rigidity and physical quantities by integrating measurements from various sensor types and qualities. Potential applications of the presented methodology are in structural health monitoring and uncertainty quantification of plate-like structures.
comment: 25 pages, 11 figures
♻ ☆ GNNMerge: Merging of GNN Models Without Accessing Training Data
Model merging has gained prominence in machine learning as a method to integrate multiple trained models into a single model without accessing the original training data. While existing approaches have demonstrated success in domains such as computer vision and NLP, their application to Graph Neural Networks (GNNs) remains unexplored. These methods often rely on the assumption of shared initialization, which is seldom applicable to GNNs. In this work, we undertake the first benchmarking study of model merging algorithms for GNNs, revealing their limited effectiveness in this context. To address these challenges, we propose GNNMerge, which utilizes a task-agnostic node embedding alignment strategy to merge GNNs. Furthermore, we establish that under a mild relaxation, the proposed optimization objective admits direct analytical solutions for widely used GNN architectures, significantly enhancing its computational efficiency. Empirical evaluations across diverse datasets, tasks, and architectures establish GNNMerge to be up to 24% more accurate than existing methods while delivering over 2 orders of magnitude speed-up compared to training from scratch.
♻ ☆ ScalingNoise: Scaling Inference-Time Search for Generating Infinite Videos
Video diffusion models (VDMs) facilitate the generation of high-quality videos, with current research predominantly concentrated on scaling efforts during training through improvements in data quality, computational resources, and model complexity. However, inference-time scaling has received less attention, with most approaches restricting models to a single generation attempt. Recent studies have uncovered the existence of "golden noises" that can enhance video quality during generation. Building on this, we find that guiding the scaling inference-time search of VDMs to identify better noise candidates not only evaluates the quality of the frames generated in the current step but also preserves the high-level object features by referencing the anchor frame from previous multi-chunks, thereby delivering long-term value. Our analysis reveals that diffusion models inherently possess flexible adjustments of computation by varying denoising steps, and even a one-step denoising approach, when guided by a reward signal, yields significant long-term benefits. Based on the observation, we proposeScalingNoise, a plug-and-play inference-time search strategy that identifies golden initial noises for the diffusion sampling process to improve global content consistency and visual diversity. Specifically, we perform one-step denoising to convert initial noises into a clip and subsequently evaluate its long-term value, leveraging a reward model anchored by previously generated content. Moreover, to preserve diversity, we sample candidates from a tilted noise distribution that up-weights promising noises. In this way, ScalingNoise significantly reduces noise-induced errors, ensuring more coherent and spatiotemporally consistent video generation. Extensive experiments on benchmark datasets demonstrate that the proposed ScalingNoise effectively improves long video generation.
♻ ☆ Demand Estimation with Text and Image Data
We propose a demand estimation method that leverages unstructured text and image data to infer substitution patterns. Using pre-trained deep learning models, we extract embeddings from product images and textual descriptions and incorporate them into a random coefficients logit model. This approach enables researchers to estimate demand even when they lack data on product attributes or when consumers value hard-to-quantify attributes, such as visual design or functional benefits. Using data from a choice experiment, we show that our approach outperforms standard attribute-based models in counterfactual predictions of consumers' second choices. We also apply it across 40 product categories on Amazon and consistently find that text and image data help identify close substitutes within each category.
♻ ☆ Debiased Offline Representation Learning for Fast Online Adaptation in Non-stationary Dynamics
Developing policies that can adjust to non-stationary environments is essential for real-world reinforcement learning applications. However, learning such adaptable policies in offline settings, with only a limited set of pre-collected trajectories, presents significant challenges. A key difficulty arises because the limited offline data makes it hard for the context encoder to differentiate between changes in the environment dynamics and shifts in the behavior policy, often leading to context misassociations. To address this issue, we introduce a novel approach called Debiased Offline Representation for fast online Adaptation (DORA). DORA incorporates an information bottleneck principle that maximizes mutual information between the dynamics encoding and the environmental data, while minimizing mutual information between the dynamics encoding and the actions of the behavior policy. We present a practical implementation of DORA, leveraging tractable bounds of the information bottleneck principle. Our experimental evaluation across six benchmark MuJoCo tasks with variable parameters demonstrates that DORA not only achieves a more precise dynamics encoding but also significantly outperforms existing baselines in terms of performance.
♻ ☆ Deep Cut-informed Graph Embedding and Clustering
Graph clustering aims to divide the graph into different clusters. The recently emerging deep graph clustering approaches are largely built on graph neural networks (GNN). However, GNN is designed for general graph encoding and there is a common issue of representation collapse in existing GNN-based deep graph clustering algorithms. We attribute two main reasons for such issues: (i) the inductive bias of GNN models: GNNs tend to generate similar representations for proximal nodes. Since graphs often contain a non-negligible amount of inter-cluster links, the bias results in error message passing and leads to biased clustering; (ii) the clustering guided loss function: most traditional approaches strive to make all samples closer to pre-learned cluster centers, which causes a degenerate solution assigning all data points to a single label thus make all samples and less discriminative. To address these challenges, we investigate graph clustering from a graph cut perspective and propose an innovative and non-GNN-based Deep Cut-informed Graph embedding and Clustering framework, namely DCGC. This framework includes two modules: (i) cut-informed graph encoding; (ii) self-supervised graph clustering via optimal transport. For the encoding module, we derive a cut-informed graph embedding objective to fuse graph structure and attributes by minimizing their joint normalized cut. For the clustering module, we utilize the optimal transport theory to obtain the clustering assignments, which can balance the guidance of "proximity to the pre-learned cluster center". With the above two tailored designs, DCGC is more suitable for the graph clustering task, which can effectively alleviate the problem of representation collapse and achieve better performance. We conduct extensive experiments to demonstrate that our method is simple but effective compared with benchmarks.
♻ ☆ GNN-Transformer Cooperative Architecture for Trustworthy Graph Contrastive Learning AAAI 2025
Graph contrastive learning (GCL) has become a hot topic in the field of graph representation learning. In contrast to traditional supervised learning relying on a large number of labels, GCL exploits augmentation strategies to generate multiple views and positive/negative pairs, both of which greatly influence the performance. Unfortunately, commonly used random augmentations may disturb the underlying semantics of graphs. Moreover, traditional GNNs, a type of widely employed encoders in GCL, are inevitably confronted with over-smoothing and over-squashing problems. To address these issues, we propose GNN-Transformer Cooperative Architecture for Trustworthy Graph Contrastive Learning (GTCA), which inherits the advantages of both GNN and Transformer, incorporating graph topology to obtain comprehensive graph representations. Theoretical analysis verifies the trustworthiness of the proposed method. Extensive experiments on benchmark datasets demonstrate state-of-the-art empirical performance.
comment: In Proceedings of AAAI 2025
♻ ☆ Latency Minimization for UAV-Enabled Federated Learning: Trajectory Design and Resource Allocation
Federated learning (FL) has become a transformative paradigm for distributed machine learning across wireless networks. However, the performance of FL is often hindered by the unreliable communication links between resource-constrained Internet of Things (IoT) devices and the central server. To overcome this challenge, we propose a novel framework that employs an unmanned aerial vehicle (UAV) as a mobile server to enhance the FL training process. By capitalizing on the UAV's mobility, we establish strong line-of-sight connections with IoT devices, thereby enhancing communication reliability and capacity. To maximize training efficiency, we formulate a latency minimization problem that jointly optimizes bandwidth allocation, computing frequencies, transmit power for both the UAV and IoT devices, and the UAV's flight trajectory. Subsequently, we analyze the required rounds of the IoT devices training and the UAV aggregation for FL convergence. Based on the convergence constraint, we transform the problem into three subproblems and develop an efficient alternating optimization algorithm to solve this problem effectively. Additionally, we provide a thorough analysis of the algorithm's convergence and computational complexity. Extensive numerical results demonstrate that our proposed scheme not only surpasses existing benchmark schemes in reducing latency up to 15.29%, but also achieves training efficiency that nearly matches the ideal scenario.
comment: This manuscript has been submitted to IEEE
♻ ☆ Passenger hazard perception based on EEG signals for highly automated driving vehicles
Enhancing the safety of autonomous vehicles is crucial, especially given recent accidents involving automated systems. As passengers in these vehicles, humans' sensory perception and decision-making can be integrated with autonomous systems to improve safety. This study explores neural mechanisms in passenger-vehicle interactions, leading to the development of a Passenger Cognitive Model (PCM) and the Passenger EEG Decoding Strategy (PEDS). Central to PEDS is a novel Convolutional Recurrent Neural Network (CRNN) that captures spatial and temporal EEG data patterns. The CRNN, combined with stacking algorithms, achieves an accuracy of $85.0\% \pm 3.18\%$. Our findings highlight the predictive power of pre-event EEG data, enhancing the detection of hazardous scenarios and offering a network-driven framework for safer autonomous vehicles.
comment: We have decided to withdraw this submission due to ongoing revisions and further refinements in our research. A revised version may be resubmitted in the future. We appreciate the feedback and interest from the community
♻ ☆ PAPAYA Federated Analytics Stack: Engineering Privacy, Scalability and Practicality
Cross-device Federated Analytics (FA) is a distributed computation paradigm designed to answer analytics queries about and derive insights from data held locally on users' devices. On-device computations combined with other privacy and security measures ensure that only minimal data is transmitted off-device, achieving a high standard of data protection. Despite FA's broad relevance, the applicability of existing FA systems is limited by compromised accuracy; lack of flexibility for data analytics; and an inability to scale effectively. In this paper, we describe our approach to combine privacy, scalability, and practicality to build and deploy a system that overcomes these limitations. Our FA system leverages trusted execution environments (TEEs) and optimizes the use of on-device computing resources to facilitate federated data processing across large fleets of devices, while ensuring robust, defensible, and verifiable privacy safeguards. We focus on federated analytics (statistics and monitoring), in contrast to systems for federated learning (ML workloads), and we flag the key differences.
♻ ☆ Probabilistic neural operators for functional uncertainty quantification
Neural operators aim to approximate the solution operator of a system of differential equations purely from data. They have shown immense success in modeling complex dynamical systems across various domains. However, the occurrence of uncertainties inherent in both model and data has so far rarely been taken into account\textemdash{}a critical limitation in complex, chaotic systems such as weather forecasting. In this paper, we introduce the probabilistic neural operator (PNO), a framework for learning probability distributions over the output function space of neural operators. PNO extends neural operators with generative modeling based on strictly proper scoring rules, integrating uncertainty information directly into the training process. We provide a theoretical justification for the approach and demonstrate improved performance in quantifying uncertainty across different domains and with respect to different baselines. Furthermore, PNO requires minimal adjustment to existing architectures, shows improved performance for most probabilistic prediction tasks, and leads to well-calibrated predictive distributions and adequate uncertainty representations even for long dynamical trajectories. Implementing our approach into large-scale models for physical applications can lead to improvements in corresponding uncertainty quantification and extreme event identification, ultimately leading to a deeper understanding of the prediction of such surrogate models.
♻ ☆ Unlearning during Learning: An Efficient Federated Machine Unlearning Method IJCAI 2024
In recent years, Federated Learning (FL) has garnered significant attention as a distributed machine learning paradigm. To facilitate the implementation of the right to be forgotten, the concept of federated machine unlearning (FMU) has also emerged. However, current FMU approaches often involve additional time-consuming steps and may not offer comprehensive unlearning capabilities, which renders them less practical in real FL scenarios. In this paper, we introduce FedAU, an innovative and efficient FMU framework aimed at overcoming these limitations. Specifically, FedAU incorporates a lightweight auxiliary unlearning module into the learning process and employs a straightforward linear operation to facilitate unlearning. This approach eliminates the requirement for extra time-consuming steps, rendering it well-suited for FL. Furthermore, FedAU exhibits remarkable versatility. It not only enables multiple clients to carry out unlearning tasks concurrently but also supports unlearning at various levels of granularity, including individual data samples, specific classes, and even at the client level. We conducted extensive experiments on MNIST, CIFAR10, and CIFAR100 datasets to evaluate the performance of FedAU. The results demonstrate that FedAU effectively achieves the desired unlearning effect while maintaining model accuracy. Our code is availiable at https://github.com/Liar-Mask/FedAU.
comment: Accepted by IJCAI 2024
♻ ☆ FedMIA: An Effective Membership Inference Attack Exploiting "All for One" Principle in Federated Learning CVPR 2025
Federated Learning (FL) is a promising approach for training machine learning models on decentralized data while preserving privacy. However, privacy risks, particularly Membership Inference Attacks (MIAs), which aim to determine whether a specific data point belongs to a target client's training set, remain a significant concern. Existing methods for implementing MIAs in FL primarily analyze updates from the target client, focusing on metrics such as loss, gradient norm, and gradient difference. However, these methods fail to leverage updates from non-target clients, potentially underutilizing available information. In this paper, we first formulate a one-tailed likelihood-ratio hypothesis test based on the likelihood of updates from non-target clients. Building upon this formulation, we introduce a three-step Membership Inference Attack (MIA) method, called FedMIA, which follows the "all for one"--leveraging updates from all clients across multiple communication rounds to enhance MIA effectiveness. Both theoretical analysis and extensive experimental results demonstrate that FedMIA outperforms existing MIAs in both classification and generative tasks. Additionally, it can be integrated as an extension to existing methods and is robust against various defense strategies, Non-IID data, and different federated structures. Our code is available in https://github.com/Liar-Mask/FedMIA.
comment: 14 pages, 6 figures; Accepted by CVPR 2025
♻ ☆ Pretraining with random noise for uncertainty calibration
Uncertainty calibration is crucial for various machine learning applications, yet it remains challenging. Many models exhibit hallucinations - confident yet inaccurate responses - due to miscalibrated confidence. Here, we show that the common practice of random initialization in deep learning, often considered a standard technique, is an underlying cause of this miscalibration, leading to excessively high confidence in untrained networks. Our method, inspired by developmental neuroscience, addresses this issue by simply pretraining networks with random noise and labels, reducing overconfidence and bringing initial confidence levels closer to chance. This ensures optimal calibration, aligning confidence with accuracy during subsequent data training, without the need for additional pre- or post-processing. Pre-calibrated networks excel at identifying "unknown data," showing low confidence for out-of-distribution inputs, thereby resolving confidence miscalibration.
♻ ☆ On best approximation by multivariate ridge functions with applications to generalized translation networks
We prove sharp upper and lower bounds for the approximation of Sobolev functions by sums of multivariate ridge functions, i.e., functions of the form $\mathbb{R}^d \ni x \mapsto \sum_{k=1}^n h_k(A_k x) \in \mathbb{R}$ with $h_k : \mathbb{R}^\ell \to \mathbb{R}$ and $A_k \in \mathbb{R}^{\ell \times d}$. We show that the order of approximation asymptotically behaves as $n^{-r/(d-\ell)}$, where $r$ is the regularity of the Sobolev functions to be approximated. Our lower bound even holds when approximating $L^\infty$-Sobolev functions of regularity $r$ with error measured in $L^1$, while our upper bound applies to the approximation of $L^p$-Sobolev functions in $L^p$ for any $1 \leq p \leq \infty$. These bounds generalize well-known results about the approximation properties of univariate ridge functions to the multivariate case. Moreover, we use these bounds to obtain sharp asymptotic bounds for the approximation of Sobolev functions using generalized translation networks and complex-valued neural networks.
♻ ☆ Volumetric Surfaces: Representing Fuzzy Geometries with Layered Meshes
High-quality view synthesis relies on volume rendering, splatting, or surface rendering. While surface rendering is typically the fastest, it struggles to accurately model fuzzy geometry like hair. In turn, alpha-blending techniques excel at representing fuzzy materials but require an unbounded number of samples per ray (P1). Further overheads are induced by empty space skipping in volume rendering (P2) and sorting input primitives in splatting (P3). We present a novel representation for real-time view synthesis where the (P1) number of sampling locations is small and bounded, (P2) sampling locations are efficiently found via rasterization, and (P3) rendering is sorting-free. We achieve this by representing objects as semi-transparent multi-layer meshes rendered in a fixed order. First, we model surface layers as signed distance function (SDF) shells with optimal spacing learned during training. Then, we bake them as meshes and fit UV textures. Unlike single-surface methods, our multi-layer representation effectively models fuzzy objects. In contrast to volume and splatting-based methods, our approach enables real-time rendering on low-power laptops and smartphones.
♻ ☆ Video Motion Transfer with Diffusion Transformers CVPR 2025
We propose DiTFlow, a method for transferring the motion of a reference video to a newly synthesized one, designed specifically for Diffusion Transformers (DiT). We first process the reference video with a pre-trained DiT to analyze cross-frame attention maps and extract a patch-wise motion signal called the Attention Motion Flow (AMF). We guide the latent denoising process in an optimization-based, training-free, manner by optimizing latents with our AMF loss to generate videos reproducing the motion of the reference one. We also apply our optimization strategy to transformer positional embeddings, granting us a boost in zero-shot motion transfer capabilities. We evaluate DiTFlow against recently published methods, outperforming all across multiple metrics and human evaluation.
comment: CVPR 2025 - Project page: https://ditflow.github.io/
♻ ☆ Joint Estimation and Prediction of City-wide Delivery Demand: A Large Language Model Empowered Graph-based Learning Approach
The proliferation of e-commerce and urbanization has significantly intensified delivery operations in urban areas, boosting the volume and complexity of delivery demand. Data-driven predictive methods, especially those utilizing machine learning techniques, have emerged to handle these complexities in urban delivery demand management problems. One particularly pressing issue that has yet to be sufficiently addressed is the joint estimation and prediction of city-wide delivery demand, as well as the generalization of the model to new cities. To this end, we formulate this problem as a transferable graph-based spatiotemporal learning task. First, an individual-collective message-passing neural network model is formalized to capture the interaction between demand patterns of associated regions. Second, by exploiting recent advances in large language models (LLMs), we extract general geospatial knowledge encodings from the unstructured locational data using the embedding generated by LLMs. Last, to encourage the cross-city generalization of the model, we integrate the encoding into the demand predictor in a transferable way. Comprehensive empirical evaluation results on two real-world delivery datasets, including eight cities in China and the US, demonstrate that our model significantly outperforms state-of-the-art baselines in accuracy, efficiency, and transferability.
♻ ☆ DR-PETS: Learning-Based Control With Planning in Adversarial Environments
Ensuring robustness against epistemic, possibly adversarial, perturbations is essential for reliable real-world decision-making. While the Probabilistic Ensembles with Trajectory Sampling (PETS) algorithm inherently handles uncertainty via ensemble-based probabilistic models, it lacks guarantees against structured adversarial or worst-case uncertainty distributions. To address this, we propose DR-PETS, a distributionally robust extension of PETS that certifies robustness against adversarial perturbations. We formalize uncertainty via a p-Wasserstein ambiguity set, enabling worst-case-aware planning through a min-max optimization framework. While PETS passively accounts for stochasticity, DR-PETS actively optimizes robustness via a tractable convex approximation integrated into PETS planning loop. Experiments on pendulum stabilization and cart-pole balancing show that DR-PETS certifies robustness against adversarial parameter perturbations, achieving consistent performance in worst-case scenarios where PETS deteriorates.
♻ ☆ Starjob: Dataset for LLM-Driven Job Shop Scheduling
Large Language Models (LLMs) have shown remarkable capabilities across various domains, but their potential for solving combinatorial optimization problems remains largely unexplored. In this paper, we investigate the applicability of LLMs to the Job Shop Scheduling Problem (JSSP), a classic challenge in combinatorial optimization that requires efficient job allocation to machines to minimize makespan. To this end, we introduce Starjob, the first supervised dataset for JSSP, comprising 130k instances specifically designed for training LLMs. Leveraging this dataset, we fine-tune the LLaMA 8B 4-bit quantized model with the LoRA method to develop an end-to-end scheduling approach. Our evaluation on standard benchmarks demonstrates that the proposed LLM-based method not only surpasses traditional Priority Dispatching Rules (PDRs) but also achieves notable improvements over state-of-the-art neural approaches like L2D, with an average improvement of 15.36% on DMU and 7.85% on Taillard benchmarks. These results highlight the untapped potential of LLMs in tackling combinatorial optimization problems, paving the way for future advancements in this area.
comment: arXiv admin note: substantial text overlap with arXiv:2408.06993
♻ ☆ A Logic for Reasoning About Aggregate-Combine Graph Neural Networks
We propose a modal logic in which counting modalities appear in linear inequalities. We show that each formula can be transformed into an equivalent graph neural network (GNN). We also show that a broad class of GNNs can be transformed efficiently into a formula, thus significantly improving upon the literature about the logical expressiveness of GNNs. We also show that the satisfiability problem is PSPACE-complete. These results bring together the promise of using standard logical methods for reasoning about GNNs and their properties, particularly in applications such as GNN querying, equivalence checking, etc. We prove that such natural problems can be solved in polynomial space.
comment: arXiv admin note: text overlap with arXiv:2307.05150
♻ ☆ Improving clustering quality evaluation in noisy Gaussian mixtures
Clustering is a well-established technique in machine learning and data analysis, widely used across various domains. Cluster validity indices, such as the Average Silhouette Width, Calinski-Harabasz, and Davies-Bouldin indices, play a crucial role in assessing clustering quality when external ground truth labels are unavailable. However, these measures can be affected by the feature relevance issue, potentially leading to unreliable evaluations in high-dimensional or noisy data sets. We introduce a theoretically grounded Feature Importance Rescaling (FIR) method that enhances the quality of clustering validation by adjusting feature contributions based on their dispersion. It attenuates noise features, clarifies clustering compactness and separation, and thereby aligns clustering validation more closely with the ground truth. Through extensive experiments on synthetic data sets under different configurations, we demonstrate that FIR consistently improves the correlation between the values of cluster validity indices and the ground truth, particularly in settings with noisy or irrelevant features. The results show that FIR increases the robustness of clustering evaluation, reduces variability in performance across different data sets, and remains effective even when clusters exhibit significant overlap. These findings highlight the potential of FIR as a valuable enhancement of clustering validation, making it a practical tool for unsupervised learning tasks where labelled data is unavailable.
♻ ☆ Automatically Adaptive Conformal Risk Control
Science and technology have a growing need for effective mechanisms that ensure reliable, controlled performance from black-box machine learning algorithms. These performance guarantees should ideally hold conditionally on the input-that is the performance guarantees should hold, at least approximately, no matter what the input. However, beyond stylized discrete groupings such as ethnicity and gender, the right notion of conditioning can be difficult to define. For example, in problems such as image segmentation, we want the uncertainty to reflect the intrinsic difficulty of the test sample, but this may be difficult to capture via a conditioning event. Building on the recent work of Gibbs et al. [2023], we propose a methodology for achieving approximate conditional control of statistical risks-the expected value of loss functions-by adapting to the difficulty of test samples. Our framework goes beyond traditional conditional risk control based on user-provided conditioning events to the algorithmic, data-driven determination of appropriate function classes for conditioning. We apply this framework to various regression and segmentation tasks, enabling finer-grained control over model performance and demonstrating that by continuously monitoring and adjusting these parameters, we can achieve superior precision compared to conventional risk-control methods.
♻ ☆ RGB-Th-Bench: A Dense benchmark for Visual-Thermal Understanding of Vision Language Models
We introduce RGB-Th-Bench, the first benchmark designed to evaluate the ability of Vision-Language Models (VLMs) to comprehend RGB-Thermal image pairs. While VLMs have demonstrated remarkable progress in visual reasoning and multimodal understanding, their evaluation has been predominantly limited to RGB-based benchmarks, leaving a critical gap in assessing their capabilities in infrared vision tasks. Existing visible-infrared datasets are either task-specific or lack high-quality annotations necessary for rigorous model evaluation. To address these limitations, RGB-Th-Bench provides a comprehensive evaluation framework covering 14 distinct skill dimensions, with a total of 1,600+ expert-annotated Yes/No questions. The benchmark employs two accuracy metrics: a standard question-level accuracy and a stricter skill-level accuracy, which evaluates model robustness across multiple questions within each skill dimension. This design ensures a thorough assessment of model performance, including resilience to adversarial and hallucinated responses. We conduct extensive evaluations on 19 state-of-the-art VLMs, revealing significant performance gaps in RGB-Thermal understanding. Our results show that even the strongest models struggle with thermal image comprehension, with performance heavily constrained by their RGB-based capabilities. Additionally, the lack of large-scale application-specific and expert-annotated thermal-caption-pair datasets in pre-training is an important reason of the observed performance gap. RGB-Th-Bench highlights the urgent need for further advancements in multimodal learning to bridge the gap between visible and thermal image understanding. The dataset is available through this link, and the evaluation code will also be made publicly available.
♻ ☆ On learning higher-order cumulants in diffusion models NeurIPS 2024
To analyse how diffusion models learn correlations beyond Gaussian ones, we study the behaviour of higher-order cumulants, or connected n-point functions, under both the forward and backward process. We derive explicit expressions for the moment- and cumulant-generating functionals, in terms of the distribution of the initial data and properties of forward process. It is shown analytically that during the forward process higher-order cumulants are conserved in models without a drift, such as the variance-expanding scheme, and that therefore the endpoint of the forward process maintains nontrivial correlations. We demonstrate that since these correlations are encoded in the score function, higher-order cumulants are learnt in the backward process, also when starting from a normal prior. We confirm our analytical results in an exactly solvable toy model with nonzero cumulants and in scalar lattice field theory.
comment: 21 pages, many figures. Extended version of contribution awarded "best 'physics for AI' paper award" in the NeurIPS 2024 workshop "Machine Learning and the Physical Sciences"; v2: references and minor clarifications added, version to appear in Machine Learning: Science and Technology
♻ ☆ Beyond [cls]: Exploring the true potential of Masked Image Modeling representations
Masked Image Modeling (MIM) has emerged as a promising approach for Self-Supervised Learning (SSL) of visual representations. However, the out-of-the-box performance of MIMs is typically inferior to competing approaches. Most users cannot afford fine-tuning due to the need for large amounts of data, high GPU consumption, and specialized user knowledge. Therefore, the practical use of MIM representations is limited. In this paper we ask what is the reason for the poor out-of-the-box performance of MIMs. Is it due to weaker features produced by MIM models, or is it due to suboptimal usage? Through detailed analysis, we show that attention in MIMs is spread almost uniformly over many patches, leading to ineffective aggregation by the [cls] token. Based on this insight, we propose Selective Aggregation to better capture the rich semantic information retained in patch tokens, which significantly improves the out-of-the-box performance of MIM.
♻ ☆ LongViTU: Instruction Tuning for Long-Form Video Understanding
This paper introduces LongViTU, a large-scale (~121k QA pairs, ~900h videos), automatically generated dataset for long-form video understanding. We propose a systematic approach that organizes videos into a hierarchical tree structure for QA generation and incorporates self-revision mechanisms to ensure high-quality QA pairs. Each QA pair in LongViTU features: 1) long-term context (average certificate length of 4.6 minutes); 2) rich knowledge and condensed reasoning (commonsense, causality, planning, etc.)). We also offer explicit timestamp annotations of relevant events for each QA pair. We have conducted extensive human studies on LongViTU, and the results prove the quality of our dataset. To better evaluate the challenges posed by LongViTU's emphasis on long-term context and condensed reasoning, we manually curate a subset of LongViTU into a benchmark. Evaluations using a state-of-the-art open-source model (LongVU), a proprietary model (Gemini-1.5-Pro), and human annotators yield GPT-4 scores of 49.9, 52.3, and 81.0, respectively, underscoring the substantial difficulty presented by LongViTU questions. Performing supervised fine-tuning (SFT) of LongVU and LLaVA-Video on LongViTU data results in average performance gains of 2.5% and 3.7%, respectively, across a suite of long video understanding benchmarks (EgoSchema, VideoMME-Long, MLVU, LVBench).
♻ ☆ Combining Relevance and Magnitude for Resource-Aware DNN Pruning
Pruning neural networks, i.e., removing some of their parameters whilst retaining their accuracy, is one of the main ways to reduce the latency of a machine learning pipeline, especially in resource- and/or bandwidth-constrained scenarios. In this context, the pruning technique, i.e., how to choose the parameters to remove, is critical to the system performance. In this paper, we propose a novel pruning approach, called FlexRel and predicated upon combining training-time and inference-time information, namely, parameter magnitude and relevance, in order to improve the resulting accuracy whilst saving both computational resources and bandwidth. Our performance evaluation shows that FlexRel is able to achieve higher pruning factors, saving over 35% bandwidth for typical accuracy targets.
♻ ☆ Self-Expansion of Pre-trained Models with Mixture of Adapters for Continual Learning
Continual learning (CL) aims to continually accumulate knowledge from a non-stationary data stream without catastrophic forgetting of learned knowledge, requiring a balance between stability and adaptability. Relying on the generalizable representation in pre-trained models (PTMs), PTM-based CL methods perform effective continual adaptation on downstream tasks by adding learnable adapters or prompts upon the frozen PTMs. However, many existing PTM-based CL methods use restricted adaptation on a fixed set of these modules to avoid forgetting, suffering from limited CL ability. Periodically adding task-specific modules results in linear model growth rate and impaired knowledge reuse. We propose Self-Expansion of pre-trained models with Modularized Adaptation (SEMA), a novel approach to enhance the control of stability-plasticity balance in PTM-based CL. SEMA automatically decides to reuse or add adapter modules on demand in CL, depending on whether significant distribution shift that cannot be handled is detected at different representation levels. We design modular adapter consisting of a functional adapter and a representation descriptor. The representation descriptors are trained as a distribution shift indicator and used to trigger self-expansion signals. For better composing the adapters, an expandable weighting router is learned jointly for mixture of adapter outputs. SEMA enables better knowledge reuse and sub-linear expansion rate. Extensive experiments demonstrate the effectiveness of the proposed self-expansion method, achieving state-of-the-art performance compared to PTM-based CL methods without memory rehearsal. Code is available at https://github.com/huiyiwang01/SEMA-CL.
comment: Code available at https: https://github.com/huiyiwang01/SEMA-CL
♻ ☆ ATM: Improving Model Merging by Alternating Tuning and Merging
Model merging has recently emerged as a cost-efficient paradigm for multi-task learning. Among current approaches, task arithmetic stands out for its simplicity and effectiveness. In this paper, we motivate the effectiveness of task vectors by linking them to multi-task gradients. We show that in a single-epoch scenario, if the optimization is performed via gradient descent, task vectors are after one step mathematically equivalent to the gradients obtained via gradient descent in a multi-task setting, and still approximate these gradients in subsequent epochs. Furthermore, we show that the effectiveness of task vectors is largely driven by the first epoch's gradient. Given this parallel between task vectors and gradients, we propose viewing model merging as a single step in an iterative process that alternates between tuning and merging (ATM). We then propose two ways to utilize ATM. The first is to replace multi-task learning with ATM in scenarios where data sharing is prohibited, such as federated learning. The second is to improve the outcome of any model merging algorithm by applying a few post-hoc iterations of ATM on a small validation dataset, which is commonly available for hyperparameter tuning. Finally, we provide both empirical and theoretical support for the effectiveness of ATM, demonstrating that it minimizes an upper bound on the loss obtained by jointly finetuning all tasks.
comment: Main paper: 9 Pages, 9 figures, 1 table
♻ ☆ Optimizing Large Model Training through Overlapped Activation Recomputation
Large model training often uses recomputation to alleviate memory pressure and pipelines to exploit the parallelism of data, tensors, and devices. However, existing recomputation approaches may incur high overhead when training real-world models, as they are executed on demand in the critical training path. In this paper, we present Lynx, a new recomputation framework to reduce overhead by overlapping recomputation with communication in training pipelines. To reduce the large search space for recomputation strategies, we propose a heuristic-based recomputation scheduling algorithm, which is based on the observation that there are identical structures in large DNN models so that we can apply the same scheduling policy to all such structures. Additionally, we propose a recomputation-aware model partitioning method to balance each stage's execution time for improved training throughput. Our comprehensive evaluation using GPT models with 1.3B-23B parameters shows that Lynx outperforms existing recomputation approaches by up to 1.37x.
comment: 13 pages
♻ ☆ Joint Estimation of Conditional Mean and Covariance for Unbalanced Panels
We develop a nonparametric, kernel-based joint estimator for conditional mean and covariance matrices in large and unbalanced panels. The estimator is supported by rigorous consistency results and finite-sample guarantees, ensuring its reliability for empirical applications. We apply it to an extensive panel of monthly US stock excess returns from 1962 to 2021, using macroeconomic and firm-specific covariates as conditioning variables. The estimator effectively captures time-varying cross-sectional dependencies, demonstrating robust statistical and economic performance. We find that idiosyncratic risk explains, on average, more than 75% of the cross-sectional variance.
♻ ☆ Feedback-driven object detection and iterative model improvement
Automated object detection has become increasingly valuable across diverse applications, yet efficient, high-quality annotation remains a persistent challenge. In this paper, we present the development and evaluation of a platform designed to interactively improve object detection models. The platform allows uploading and annotating images as well as fine-tuning object detection models. Users can then manually review and refine annotations, further creating improved snapshots that are used for automatic object detection on subsequent image uploads - a process we refer to as semi-automatic annotation resulting in a significant gain in annotation efficiency. Whereas iterative refinement of model results to speed up annotation has become common practice, we are the first to quantitatively evaluate its benefits with respect to time, effort, and interaction savings. Our experimental results show clear evidence for a significant time reduction of up to 53% for semi-automatic compared to manual annotation. Importantly, these efficiency gains did not compromise annotation quality, while matching or occasionally even exceeding the accuracy of manual annotations. These findings demonstrate the potential of our lightweight annotation platform for creating high-quality object detection datasets and provide best practices to guide future development of annotation platforms. The platform is open-source, with the frontend and backend repositories available on GitHub. To support the understanding of our labeling process, we have created an explanatory video demonstrating the methodology using microscopy images of E. coli bacteria as an example.
comment: Code: https://github.com/ml-lab-htw/iterative-annotate Video: https://www.youtube.com/watch?v=CM9uhE8NN5E
♻ ☆ Evaluating the effects of Data Sparsity on the Link-level Bicycling Volume Estimation: A Graph Convolutional Neural Network Approach
Accurate bicycling volume estimation is crucial for making informed decisions and planning about future investments in bicycling infrastructure. However, traditional link-level volume estimation models are effective for motorized traffic but face significant challenges when applied to the bicycling context because of sparse data and the intricate nature of bicycling mobility patterns. To the best of our knowledge, we present the first study to utilize a Graph Convolutional Network (GCN) architecture to model link-level bicycling volumes and systematically investigate the impact of varying levels of data sparsity (0%--99%) on model performance, simulating real-world scenarios. We have leveraged Strava Metro data as the primary source of bicycling counts across 15,933 road segments/links in the City of Melbourne, Australia. To evaluate the effectiveness of the GCN model, we benchmark it against traditional machine learning models, such as linear regression, support vector machines, and random forest. Our results show that the GCN model outperforms these traditional models in predicting Annual Average Daily Bicycle (AADB) counts, demonstrating its ability to capture the spatial dependencies inherent in bicycle traffic networks. While GCN remains robust up to 80% sparsity, its performance declines sharply beyond this threshold, highlighting the challenges of extreme data sparsity. These findings underscore the potential of GCNs in enhancing bicycling volume estimation, while also emphasizing the need for further research on methods to improve model resilience under high-sparsity conditions. Our findings offer valuable insights for city planners aiming to improve bicycling infrastructure and promote sustainable transportation.
♻ ☆ Inductive-Associative Meta-learning Pipeline with Human Cognitive Patterns for Unseen Drug-Target Interaction Prediction
Significant differences in protein structures hinder the generalization of existing drug-target interaction (DTI) models, which often rely heavily on pre-learned binding principles or detailed annotations. In contrast, BioBridge designs an Inductive-Associative pipeline inspired by the workflow of scientists who base their accumulated expertise on drawing insights into novel drug-target pairs from weakly related references. BioBridge predicts novel drug-target interactions using limited sequence data, incorporating multi-level encoders with adversarial training to accumulate transferable binding principles. On these principles basis, BioBridge employs a dynamic prototype meta-learning framework to associate insights from weakly related annotations, enabling robust predictions for previously unseen drug-target pairs. Extensive experiments demonstrate that BioBridge surpasses existing models, especially for unseen proteins. Notably, when only homologous protein binding data is available, BioBridge proves effective for virtual screening of the epidermal growth factor receptor and adenosine receptor, underscoring its potential in drug discovery.
♻ ☆ Time and Memory Trade-off of KV-Cache Compression in Tensor Transformer Decoding
The key-value (KV) cache in the tensor version of transformers presents a significant bottleneck during inference. While previous work analyzes the fundamental space complexity barriers in standard attention mechanisms [Haris and Onak, 2025], our work generalizes the space complexity barriers result to tensor attention version. Our theoretical contributions rely on a reduction from communication complexity and deduce the memory lower bound for tensor-structured attention mechanisms when $d = \Omega(\log n)$. Furthermore, we introduce two types of tensor attention cache and present a trade-off between time and memory for two scenarios. Overall, our work provides a theoretical foundation for us to understand the time-memory tradeoff of KV-Cache compression in tensor attention decoding and offers more perspectives in developing more memory-efficient tensor attention Transformer architectures.
♻ ☆ Group Reasoning Emission Estimation Networks
Accurate greenhouse gas (GHG) emission reporting is critical for governments, businesses, and investors. However, adoption remains limited particularly among small and medium enterprises due to high implementation costs, fragmented emission factor databases, and a lack of robust sector classification methods. To address these challenges, we introduce Group Reasoning Emission Estimation Networks (GREEN), an AI-driven carbon accounting framework that standardizes enterprise-level emission estimation, constructs a large-scale benchmark dataset, and leverages a novel reasoning approach with large language models (LLMs). Specifically, we compile textual descriptions for 20,850 companies with validated North American Industry Classification System (NAICS) labels and align these with an economic model of carbon intensity factors. By reframing sector classification as an information retrieval task, we fine-tune Sentence-BERT models using a contrastive learning loss. To overcome the limitations of single-stage models in handling thousands of hierarchical categories, we propose a Group Reasoning method that ensembles LLM classifiers based on the natural NAICS ontology, decomposing the task into multiple sub-classification steps. We theoretically prove that this approach reduces classification uncertainty and computational complexity. Experiments on 1,114 NAICS categories yield state-of-the-art performance (83.68% Top-1, 91.47% Top-10 accuracy), and case studies on 20 companies report a mean absolute percentage error (MAPE) of 45.88%. The project is available at: https://huggingface.co/datasets/Yvnminc/ExioNAICS.
♻ ☆ Flip Learning: Weakly Supervised Erase to Segment Nodules in Breast Ultrasound
Accurate segmentation of nodules in both 2D breast ultrasound (BUS) and 3D automated breast ultrasound (ABUS) is crucial for clinical diagnosis and treatment planning. Therefore, developing an automated system for nodule segmentation can enhance user independence and expedite clinical analysis. Unlike fully-supervised learning, weakly-supervised segmentation (WSS) can streamline the laborious and intricate annotation process. However, current WSS methods face challenges in achieving precise nodule segmentation, as many of them depend on inaccurate activation maps or inefficient pseudo-mask generation algorithms. In this study, we introduce a novel multi-agent reinforcement learning-based WSS framework called Flip Learning, which relies solely on 2D/3D boxes for accurate segmentation. Specifically, multiple agents are employed to erase the target from the box to facilitate classification tag flipping, with the erased region serving as the predicted segmentation mask. The key contributions of this research are as follows: (1) Adoption of a superpixel/supervoxel-based approach to encode the standardized environment, capturing boundary priors and expediting the learning process. (2) Introduction of three meticulously designed rewards, comprising a classification score reward and two intensity distribution rewards, to steer the agents' erasing process precisely, thereby avoiding both under- and over-segmentation. (3) Implementation of a progressive curriculum learning strategy to enable agents to interact with the environment in a progressively challenging manner, thereby enhancing learning efficiency. Extensively validated on the large in-house BUS and ABUS datasets, our Flip Learning method outperforms state-of-the-art WSS methods and foundation models, and achieves comparable performance as fully-supervised learning algorithms.
comment: Accepted by Medical Image Analysis. 24 pages, 13 figures, 20 tabels
♻ ☆ Sparse Alignment Enhanced Latent Diffusion Transformer for Zero-Shot Speech Synthesis
While recent zero-shot text-to-speech (TTS) models have significantly improved speech quality and expressiveness, mainstream systems still suffer from issues related to speech-text alignment modeling: 1) models without explicit speech-text alignment modeling exhibit less robustness, especially for hard sentences in practical applications; 2) predefined alignment-based models suffer from naturalness constraints of forced alignments. This paper introduces \textit{MegaTTS 3}, a TTS system featuring an innovative sparse alignment algorithm that guides the latent diffusion transformer (DiT). Specifically, we provide sparse alignment boundaries to MegaTTS 3 to reduce the difficulty of alignment without limiting the search space, thereby achieving high naturalness. Moreover, we employ a multi-condition classifier-free guidance strategy for accent intensity adjustment and adopt the piecewise rectified flow technique to accelerate the generation process. Experiments demonstrate that MegaTTS 3 achieves state-of-the-art zero-shot TTS speech quality and supports highly flexible control over accent intensity. Notably, our system can generate high-quality one-minute speech with only 8 sampling steps. Audio samples are available at https://sditdemo.github.io/sditdemo/.
♻ ☆ Sparse Feature Circuits: Discovering and Editing Interpretable Causal Graphs in Language Models
We introduce methods for discovering and applying sparse feature circuits. These are causally implicated subnetworks of human-interpretable features for explaining language model behaviors. Circuits identified in prior work consist of polysemantic and difficult-to-interpret units like attention heads or neurons, rendering them unsuitable for many downstream applications. In contrast, sparse feature circuits enable detailed understanding of unanticipated mechanisms. Because they are based on fine-grained units, sparse feature circuits are useful for downstream tasks: We introduce SHIFT, where we improve the generalization of a classifier by ablating features that a human judges to be task-irrelevant. Finally, we demonstrate an entirely unsupervised and scalable interpretability pipeline by discovering thousands of sparse feature circuits for automatically discovered model behaviors.
comment: Code and data at https://github.com/saprmarks/feature-circuits. Demonstration at https://feature-circuits.xyz
♻ ☆ OODFace: Benchmarking Robustness of Face Recognition under Common Corruptions and Appearance Variations
With the rise of deep learning, facial recognition technology has seen extensive research and rapid development. Although facial recognition is considered a mature technology, we find that existing open-source models and commercial algorithms lack robustness in certain complex Out-of-Distribution (OOD) scenarios, raising concerns about the reliability of these systems. In this paper, we introduce OODFace, which explores the OOD challenges faced by facial recognition models from two perspectives: common corruptions and appearance variations. We systematically design 30 OOD scenarios across 9 major categories tailored for facial recognition. By simulating these challenges on public datasets, we establish three robustness benchmarks: LFW-C/V, CFP-FP-C/V, and YTF-C/V. We then conduct extensive experiments on 19 facial recognition models and 3 commercial APIs, along with extended physical experiments on face masks to assess their robustness. Next, we explore potential solutions from two perspectives: defense strategies and Vision-Language Models (VLMs). Based on the results, we draw several key insights, highlighting the vulnerability of facial recognition systems to OOD data and suggesting possible solutions. Additionally, we offer a unified toolkit that includes all corruption and variation types, easily extendable to other datasets. We hope that our benchmarks and findings can provide guidance for future improvements in facial recognition model robustness.
♻ ☆ DefectFill: Realistic Defect Generation with Inpainting Diffusion Model for Visual Inspection CVPR 2025
Developing effective visual inspection models remains challenging due to the scarcity of defect data. While image generation models have been used to synthesize defect images, producing highly realistic defects remains difficult. We propose DefectFill, a novel method for realistic defect generation that requires only a few reference defect images. It leverages a fine-tuned inpainting diffusion model, optimized with our custom loss functions incorporating defect, object, and attention terms. It enables precise capture of detailed, localized defect features and their seamless integration into defect-free objects. Additionally, our Low-Fidelity Selection method further enhances the defect sample quality. Experiments show that DefectFill generates high-quality defect images, enabling visual inspection models to achieve state-of-the-art performance on the MVTec AD dataset.
comment: Accepted to CVPR 2025
♻ ☆ MoReVQA: Exploring Modular Reasoning Models for Video Question Answering CVPR 2024
This paper addresses the task of video question answering (videoQA) via a decomposed multi-stage, modular reasoning framework. Previous modular methods have shown promise with a single planning stage ungrounded in visual content. However, through a simple and effective baseline, we find that such systems can lead to brittle behavior in practice for challenging videoQA settings. Thus, unlike traditional single-stage planning methods, we propose a multi-stage system consisting of an event parser, a grounding stage, and a final reasoning stage in conjunction with an external memory. All stages are training-free, and performed using few-shot prompting of large models, creating interpretable intermediate outputs at each stage. By decomposing the underlying planning and task complexity, our method, MoReVQA, improves over prior work on standard videoQA benchmarks (NExT-QA, iVQA, EgoSchema, ActivityNet-QA) with state-of-the-art results, and extensions to related tasks (grounded videoQA, paragraph captioning).
comment: CVPR 2024; updated NExT-GQA results in Appendix
♻ ☆ iTool: Boosting Tool Use of Large Language Models via Iterative Reinforced Fine-Tuning ACL
Augmenting large language models (LLMs) with external tools is known as a promising approach to enhancing their capabilities, especially for complex tasks. Synthesizing tool-use data through real-world simulations is an effective way to achieve it. Nevertheless, our investigation reveals that (1) training gains significantly decay as synthetic data increases. The model struggles to benefit from more synthetic data due to potential data diversity issues, resulting in poor performance in complex scenarios. Moreover, we find that (2) this challenge primarily manifests as minor discrepancies between the model's output and the ground truth response (termed as deficiency), such as errors in parameter values that require complex reasoning from the context to resolve. To this end, we propose an iterative reinforced fine-tuning strategy designed to alleviate these challenges. This strategy involves: (1) enhancing the diversity of synthetic data through path exploration of Monte Carlo Tree Search. (2) iteratively identifying deficiency-related data, constructing fine-grained preference pairs to pinpoint deficiencies, and then applying preference optimization to optimize these deficiencies. Our experiments show that models trained using our method achieve about 12\% better performance than baseline models, outperforming larger open-source and closed-source models.
comment: under review ACL
♻ ☆ Robust Feature Learning for Multi-Index Models in High Dimensions ICLR
Recently, there have been numerous studies on feature learning with neural networks, specifically on learning single- and multi-index models where the target is a function of a low-dimensional projection of the input. Prior works have shown that in high dimensions, the majority of the compute and data resources are spent on recovering the low-dimensional projection; once this subspace is recovered, the remainder of the target can be learned independently of the ambient dimension. However, implications of feature learning in adversarial settings remain unexplored. In this work, we take the first steps towards understanding adversarially robust feature learning with neural networks. Specifically, we prove that the hidden directions of a multi-index model offer a Bayes optimal low-dimensional projection for robustness against $\ell_2$-bounded adversarial perturbations under the squared loss, assuming that the multi-index coordinates are statistically independent from the rest of the coordinates. Therefore, robust learning can be achieved by first performing standard feature learning, then robustly tuning a linear readout layer on top of the standard representations. In particular, we show that adversarially robust learning is just as easy as standard learning. Specifically, the additional number of samples needed to robustly learn multi-index models when compared to standard learning does not depend on dimensionality.
comment: 41 pages, 1 figure. To appear in the International Conference on Learning Representations (ICLR), 2025
♻ ☆ Towards Controllable Speech Synthesis in the Era of Large Language Models: A Survey
Text-to-speech (TTS), also known as speech synthesis, is a prominent research area that aims to generate natural-sounding human speech from text. Recently, with the increasing industrial demand, TTS technologies have evolved beyond synthesizing human-like speech to enabling controllable speech generation. This includes fine-grained control over various attributes of synthesized speech such as emotion, prosody, timbre, and duration. In addition, advancements in deep learning, such as diffusion and large language models, have significantly enhanced controllable TTS over the past several years. In this work, we conduct a comprehensive survey of controllable TTS, covering approaches ranging from basic control techniques to methods utilizing natural language prompts, aiming to provide a clear understanding of the current state of research. We examine the general controllable TTS pipeline, challenges, model architectures, and control strategies, offering a comprehensive and clear taxonomy of existing methods. Additionally, we provide a detailed summary of datasets and evaluation metrics and shed some light on the applications and future directions of controllable TTS. To the best of our knowledge, this survey paper provides the first comprehensive review of emerging controllable TTS methods, which can serve as a beneficial resource for both academic researchers and industrial practitioners.
comment: A comprehensive survey on controllable TTS, 26 pages, 7 tables, 6 figures, 317 references. Under review
♻ ☆ Learning Multi-Index Models with Neural Networks via Mean-Field Langevin Dynamics ICLR
We study the problem of learning multi-index models in high-dimensions using a two-layer neural network trained with the mean-field Langevin algorithm. Under mild distributional assumptions on the data, we characterize the effective dimension $d_{\mathrm{eff}}$ that controls both sample and computational complexity by utilizing the adaptivity of neural networks to latent low-dimensional structures. When the data exhibit such a structure, $d_{\mathrm{eff}}$ can be significantly smaller than the ambient dimension. We prove that the sample complexity grows almost linearly with $d_{\mathrm{eff}}$, bypassing the limitations of the information and generative exponents that appeared in recent analyses of gradient-based feature learning. On the other hand, the computational complexity may inevitably grow exponentially with $d_{\mathrm{eff}}$ in the worst-case scenario. Motivated by improving computational complexity, we take the first steps towards polynomial time convergence of the mean-field Langevin algorithm by investigating a setting where the weights are constrained to be on a compact manifold with positive Ricci curvature, such as the hypersphere. There, we study assumptions under which polynomial time convergence is achievable, whereas similar assumptions in the Euclidean setting lead to exponential time complexity.
comment: 36 pages, 2 figures. To appear in the International Conference on Learning Representations (ICLR), 2025
GR00T N1: An Open Foundation Model for Generalist Humanoid Robots
General-purpose robots need a versatile body and an intelligent mind. Recent advancements in humanoid robots have shown great promise as a hardware platform for building generalist autonomy in the human world. A robot foundation model, trained on massive and diverse data sources, is essential for enabling the robots to reason about novel situations, robustly handle real-world variability, and rapidly learn new tasks. To this end, we introduce GR00T N1, an open foundation model for humanoid robots. GR00T N1 is a Vision-Language-Action (VLA) model with a dual-system architecture. The vision-language module (System 2) interprets the environment through vision and language instructions. The subsequent diffusion transformer module (System 1) generates fluid motor actions in real time. Both modules are tightly coupled and jointly trained end-to-end. We train GR00T N1 with a heterogeneous mixture of real-robot trajectories, human videos, and synthetically generated datasets. We show that our generalist robot model GR00T N1 outperforms the state-of-the-art imitation learning baselines on standard simulation benchmarks across multiple robot embodiments. Furthermore, we deploy our model on the Fourier GR-1 humanoid robot for language-conditioned bimanual manipulation tasks, achieving strong performance with high data efficiency.
comment: Authors are listed alphabetically. Project leads are Linxi "Jim" Fan and Yuke Zhu. For more information, see https://developer.nvidia.com/isaac/gr00t
♻ ☆ Towards Better Alignment: Training Diffusion Models with Reinforcement Learning Against Sparse Rewards CVPR 2025
Diffusion models have achieved remarkable success in text-to-image generation. However, their practical applications are hindered by the misalignment between generated images and corresponding text prompts. To tackle this issue, reinforcement learning (RL) has been considered for diffusion model fine-tuning. Yet, RL's effectiveness is limited by the challenge of sparse reward, where feedback is only available at the end of the generation process. This makes it difficult to identify which actions during the denoising process contribute positively to the final generated image, potentially leading to ineffective or unnecessary denoising policies. To this end, this paper presents a novel RL-based framework that addresses the sparse reward problem when training diffusion models. Our framework, named $\text{B}^2\text{-DiffuRL}$, employs two strategies: \textbf{B}ackward progressive training and \textbf{B}ranch-based sampling. For one thing, backward progressive training focuses initially on the final timesteps of denoising process and gradually extends the training interval to earlier timesteps, easing the learning difficulty from sparse rewards. For another, we perform branch-based sampling for each training interval. By comparing the samples within the same branch, we can identify how much the policies of the current training interval contribute to the final image, which helps to learn effective policies instead of unnecessary ones. $\text{B}^2\text{-DiffuRL}$ is compatible with existing optimization algorithms. Extensive experiments demonstrate the effectiveness of $\text{B}^2\text{-DiffuRL}$ in improving prompt-image alignment and maintaining diversity in generated images. The code for this work is available.
comment: Accepted to CVPR 2025, add references
♻ ☆ VERA: Explainable Video Anomaly Detection via Verbalized Learning of Vision-Language Models CVPR 2025
The rapid advancement of vision-language models (VLMs) has established a new paradigm in video anomaly detection (VAD): leveraging VLMs to simultaneously detect anomalies and provide comprehendible explanations for the decisions. Existing work in this direction often assumes the complex reasoning required for VAD exceeds the capabilities of pretrained VLMs. Consequently, these approaches either incorporate specialized reasoning modules during inference or rely on instruction tuning datasets through additional training to adapt VLMs for VAD. However, such strategies often incur substantial computational costs or data annotation overhead. To address these challenges in explainable VAD, we introduce a verbalized learning framework named VERA that enables VLMs to perform VAD without model parameter modifications. Specifically, VERA automatically decomposes the complex reasoning required for VAD into reflections on simpler, more focused guiding questions capturing distinct abnormal patterns. It treats these reflective questions as learnable parameters and optimizes them through data-driven verbal interactions between learner and optimizer VLMs, using coarsely labeled training data. During inference, VERA embeds the learned questions into model prompts to guide VLMs in generating segment-level anomaly scores, which are then refined into frame-level scores via the fusion of scene and temporal contexts. Experimental results on challenging benchmarks demonstrate that the learned questions of VERA are highly adaptable, significantly improving both detection performance and explainability of VLMs for VAD.
comment: Accepted in CVPR 2025
♻ ☆ SyncDiff: Synchronized Motion Diffusion for Multi-Body Human-Object Interaction Synthesis
Synthesizing realistic human-object interaction motions is a critical problem in VR/AR and human animation. Unlike the commonly studied scenarios involving a single human or hand interacting with one object, we address a more generic multi-body setting with arbitrary numbers of humans, hands, and objects. This complexity introduces significant challenges in synchronizing motions due to the high correlations and mutual influences among bodies. To address these challenges, we introduce SyncDiff, a novel method for multi-body interaction synthesis using a synchronized motion diffusion strategy. SyncDiff employs a single diffusion model to capture the joint distribution of multi-body motions. To enhance motion fidelity, we propose a frequency-domain motion decomposition scheme. Additionally, we introduce a new set of alignment scores to emphasize the synchronization of different body motions. SyncDiff jointly optimizes both data sample likelihood and alignment likelihood through an explicit synchronization strategy. Extensive experiments across four datasets with various multi-body configurations demonstrate the superiority of SyncDiff over existing state-of-the-art motion synthesis methods.
comment: 26 pages, 10 figures
♻ ☆ An Empirical Study of the Impact of Federated Learning on Machine Learning Model Accuracy
Federated Learning (FL) enables distributed ML model training on private user data at the global scale. Despite the potential of FL demonstrated in many domains, an in-depth view of its impact on model accuracy remains unclear. In this paper, we investigate, systematically, how this learning paradigm can affect the accuracy of state-of-the-art ML models for a variety of ML tasks. We present an empirical study that involves various data types: text, image, audio, and video, and FL configuration knobs: data distribution, FL scale, client sampling, and local and global computations. Our experiments are conducted in a unified FL framework to achieve high fidelity, with substantial human efforts and resource investments. Based on the results, we perform a quantitative analysis of the impact of FL, and highlight challenging scenarios where applying FL degrades the accuracy of the model drastically and identify cases where the impact is negligible. The detailed and extensive findings can benefit practical deployments and future development of FL.
♻ ☆ LSEAttention is All You Need for Time Series Forecasting
Transformer-based architectures have achieved remarkable success in natural language processing and computer vision. However, their performance in multivariate long-term forecasting often falls short compared to simpler linear baselines. Previous research has identified the traditional attention mechanism as a key factor limiting their effectiveness in this domain. To bridge this gap, we introduce LATST, a novel approach designed to mitigate entropy collapse and training instability common challenges in Transformer-based time series forecasting. We rigorously evaluate LATST across multiple real-world multivariate time series datasets, demonstrating its ability to outperform existing state-of-the-art Transformer models. Notably, LATST manages to achieve competitive performance with fewer parameters than some linear models on certain datasets, highlighting its efficiency and effectiveness.
comment: 8 pages with referencing, 1 figure, 5 tables
♻ ☆ Neural Exploratory Landscape Analysis for Meta-Black-Box-Optimization
Recent research in Meta-Black-Box Optimization (MetaBBO) have shown that meta-trained neural networks can effectively guide the design of black-box optimizers, significantly reducing the need for expert tuning and delivering robust performance across complex problem distributions. Despite their success, a paradox remains: MetaBBO still rely on human-crafted Exploratory Landscape Analysis features to inform the meta-level agent about the low-level optimization progress. To address the gap, this paper proposes Neural Exploratory Landscape Analysis (NeurELA), a novel framework that dynamically profiles landscape features through a two-stage, attention-based neural network, executed in an entirely end-to-end fashion. NeurELA is pre-trained over a variety of MetaBBO algorithms using a multi-task neuroevolution strategy. Extensive experiments show that NeurELA achieves consistently superior performance when integrated into different and even unseen MetaBBO tasks and can be efficiently fine-tuned for further performance boost. This advancement marks a pivotal step in making MetaBBO algorithms more autonomous and broadly applicable. The source code of NeurELA can be accessed at https://github.com/GMC-DRL/Neur-ELA.
♻ ☆ QCPINN: Quantum Classical Physics-Informed Neural Networks for Solving PDEs
Physics-informed neural networks (PINNs) have emerged as promising methods for solving partial differential equations (PDEs) by embedding physical laws into neural architectures. However, these classical approaches often require large number of parameters for solving complex problems or achieving reasonable accuracy. We investigate whether quantum-enhanced architectures can achieve comparable performance while significantly reducing model complexity. We propose a quantum-classical physics-informed neural network (QCPINN) combining quantum and classical components to solve PDEs with fewer parameters while maintaining comparable accuracy and training convergence. Our approach systematically evaluates two quantum circuit paradigms (e.g., continuous-variable (CV) and discrete-variable (DV)) implementations with four circuit topologies (e.g., alternate, cascade, cross-mesh, and layered), two embedding schemes (e.g., amplitude and angle) on five benchmark PDEs (e.g., Helmholtz, lid-driven cavity, wave, Klein-Gordon, and convection-diffusion equations). Results demonstrate that QCPINNs achieve comparable accuracy to classical PINNs while requiring approximately 10% trainable parameters across different PDEs, and resulting in a further 40% reduction in relative L2 error for the convection-diffusion equation. DV-based circuits with angle embedding and cascade configurations consistently exhibited enhanced convergence stability across all problem types. Our finding establishes parameter efficiency as a quantifiable quantum advantage in physics-informed machine learning. By significantly reducing model complexity while maintaining solution quality, QCPINNs represent a potential direction for overcoming computational bottlenecks in scientific computing applications where traditional approaches require large parameter spaces.
♻ ☆ Rethinking the Global Knowledge of CLIP in Training-Free Open-Vocabulary Semantic Segmentation
Recent works modify CLIP to perform open-vocabulary semantic segmentation in a training-free manner (TF-OVSS). In vanilla CLIP, patch-wise image representations mainly encode homogeneous image-level properties, which hinders the application of CLIP to the dense prediction task. Previous TF-OVSS works sacrifice globality to enhance the locality of CLIP features, by making each patch mainly attend to itself or its neighboring patches within a narrow local window. With their modifications,the ability of CLIP to aggregate global context information is largely weakened. Differently, in this paper, we rethink the global knowledge encoded by CLIP and propose GCLIP to answer how to extract and utilize beneficial global knowledge of CLIP for TF-OVSS. As the representation of each patch is finally determined by the attention weights and the Value embeddings, we propose to reshape the last-block attention and Value embeddings to aggregate useful global context into final features. Firstly, we aim to equip the last-block attention with image-level properties while not introducing homogeneous attention patterns across patches. To realize the goal, we fuse the attention from the global-token emerging blocks with the Query-Query attention. Secondly, we aim to make Value embeddings of the last-block attention module more semantically correlated. To realize this, we design a novel channel suppression strategy.Extensive experiments on five standard benchmarks demonstrate that our method consistently outperforms previous state-of-the-arts.
comment: Under review
♻ ☆ Vibravox: A Dataset of French Speech Captured with Body-conduction Audio Sensors
Vibravox is a dataset compliant with the General Data Protection Regulation (GDPR) containing audio recordings using five different body-conduction audio sensors: two in-ear microphones, two bone conduction vibration pickups, and a laryngophone. The dataset also includes audio data from an airborne microphone used as a reference. The Vibravox corpus contains 45 hours per sensor of speech samples and physiological sounds recorded by 188 participants under different acoustic conditions imposed by a high order ambisonics 3D spatializer. Annotations about the recording conditions and linguistic transcriptions are also included in the corpus. We conducted a series of experiments on various speech-related tasks, including speech recognition, speech enhancement, and speaker verification. These experiments were carried out using state-of-the-art models to evaluate and compare their performances on signals captured by the different audio sensors offered by the Vibravox dataset, with the aim of gaining a better grasp of their individual characteristics.
comment: 23 pages, 42 figures
Robotics 56
☆ Flying Vines: Design, Modeling, and Control of a Soft Aerial Robotic Arm
Aerial robotic arms aim to enable inspection and environment interaction in otherwise hard-to-reach areas from the air. However, many aerial manipulators feature bulky or heavy robot manipulators mounted to large, high-payload aerial vehicles. Instead, we propose an aerial robotic arm with low mass and a small stowed configuration called a "flying vine". The flying vine consists of a small, maneuverable quadrotor equipped with a soft, growing, inflated beam as the arm. This soft robot arm is underactuated, and positioning of the end effector is achieved by controlling the coupled quadrotor-vine dynamics. In this work, we present the flying vine design and a modeling and control framework for tracking desired end effector trajectories. The dynamic model leverages data-driven modeling methods and introduces bilinear interpolation to account for time-varying dynamic parameters. We use trajectory optimization to plan quadrotor controls that produce desired end effector motions. Experimental results on a physical prototype demonstrate that our framework enables the flying vine to perform high-speed end effector tracking, laying a foundation for performing dynamic maneuvers with soft aerial manipulators.
comment: Submitted to RA-L
☆ Multi-Robot Coordination Under Physical Limitations
Multi-robot coordination is fundamental to various applications, including autonomous exploration, search and rescue, and cooperative transportation. This paper presents an optimal consensus framework for multi-robot systems (MRSs) that ensures efficient rendezvous while minimizing energy consumption and addressing actuator constraints. A critical challenge in real-world deployments is actuator limitations, particularly wheel velocity saturation, which can significantly degrade control performance. To address this issue, we incorporate Pontryagin Minimum Principle (PMP) into the control design, facilitating constrained optimization while ensuring system stability and feasibility. The resulting optimal control policy effectively balances coordination efficiency and energy consumption, even in the presence of actuation constraints. The proposed framework is validated through extensive numerical simulations and real-world experiments conducted using a team of Robotarium mobile robots. The experimental results confirm that our control strategies achieve reliable and efficient coordinated rendezvous while addressing real-world challenges such as communication delays, sensor noise, and packet loss.
☆ Beyond Visuals: Investigating Force Feedback in Extended Reality for Robot Data Collection
This work explores how force feedback affects various aspects of robot data collection within the Extended Reality (XR) setting. Force feedback has been proved to enhance the user experience in Extended Reality (XR) by providing contact-rich information. However, its impact on robot data collection has not received much attention in the robotics community. This paper addresses this shortcoming by conducting an extensive user study on the effects of force feedback during data collection in XR. We extended two XR-based robot control interfaces, Kinesthetic Teaching and Motion Controllers, with haptic feedback features. The user study is conducted using manipulation tasks ranging from simple pick-place to complex peg assemble, requiring precise operations. The evaluations show that force feedback enhances task performance and user experience, particularly in tasks requiring high-precision manipulation. These improvements vary depending on the robot control interface and task complexity. This paper provides new insights into how different factors influence the impact of force feedback.
☆ Toward Dynamic Control of Tendon-Driven Continuum Robots using Clarke Transform
In this paper, we propose a dynamic model and control framework for tendon-driven continuum robots with multiple segments and an arbitrary number of tendons per segment. Our approach leverages the Clarke transform, the Euler-Lagrange formalism, and the piecewise constant curvature assumption to formulate a dynamic model on a two-dimensional manifold embedded in the joint space that inherently satisfies tendon constraints. We present linear controllers that operate directly on this manifold, along with practical methods for preventing negative tendon forces without compromising control fidelity. We validate these approaches in simulation and on a physical prototype with one segment and five tendons, demonstrating accurate dynamic behavior and robust trajectory tracking under real-time conditions.
comment: 8 pages and 8 figures
☆ Immersive and Wearable Thermal Rendering for Augmented Reality
In augmented reality (AR), where digital content is overlaid onto the real world, realistic thermal feedback has been shown to enhance immersion. Yet current thermal feedback devices, heavily influenced by the needs of virtual reality, often hinder physical interactions and are ineffective for immersion in AR. To bridge this gap, we have identified three design considerations relevant for AR thermal feedback: indirect feedback to maintain dexterity, thermal passthrough to preserve real-world temperature perception, and spatiotemporal rendering for dynamic sensations. We then created a unique and innovative thermal feedback device that satisfies these criteria. Human subject experiments assessing perceptual sensitivity, object temperature matching, spatial pattern recognition, and moving thermal stimuli demonstrated the impact of our design, enabling realistic temperature discrimination, virtual object perception, and enhanced immersion. These findings demonstrate that carefully designed thermal feedback systems can bridge the sensory gap between physical and virtual interactions, enhancing AR realism and usability.
☆ Representation Improvement in Latent Space for Search-Based Testing of Autonomous Robotic Systems
Testing autonomous robotic systems, such as self-driving cars and unmanned aerial vehicles, is challenging due to their interaction with highly unpredictable environments. A common practice is to first conduct simulation-based testing, which, despite reducing real-world risks, remains time-consuming and resource-intensive due to the vast space of possible test scenarios. A number of search-based approaches were proposed to generate test scenarios more efficiently. A key aspect of any search-based test generation approach is the choice of representation used during the search process. However, existing methods for improving test scenario representation remain limited. We propose RILaST (Representation Improvement in Latent Space for Search-Based Testing) approach, which enhances test representation by mapping it to the latent space of a variational autoencoder. We evaluate RILaST on two use cases, including autonomous drone and autonomous lane-keeping assist system. The obtained results show that RILaST allows finding between 3 to 4.6 times more failures than baseline approaches, achieving a high level of test diversity.
☆ Robust Flower Cluster Matching Using The Unscented Transform
Monitoring flowers over time is essential for precision robotic pollination in agriculture. To accomplish this, a continuous spatial-temporal observation of plant growth can be done using stationary RGB-D cameras. However, image registration becomes a serious challenge due to changes in the visual appearance of the plant caused by the pollination process and occlusions from growth and camera angles. Plants flower in a manner that produces distinct clusters on branches. This paper presents a method for matching flower clusters using descriptors generated from RGB-D data and considers allowing for spatial uncertainty within the cluster. The proposed approach leverages the Unscented Transform to efficiently estimate plant descriptor uncertainty tolerances, enabling a robust image-registration process despite temporal changes. The Unscented Transform is used to handle the nonlinear transformations by propagating the uncertainty of flower positions to determine the variations in the descriptor domain. A Monte Carlo simulation is used to validate the Unscented Transform results, confirming our method's effectiveness for flower cluster matching. Therefore, it can facilitate improved robotics pollination in dynamic environments.
comment: *CASE2025 Under Review*
☆ Safety integrity framework for automated driving
This paper describes the comprehensive safety framework that underpinned the development, release process, and regulatory approval of BMW's first SAE Level 3 Automated Driving System. The framework combines established qualitative and quantitative methods from the fields of Systems Engineering, Engineering Risk Analysis, Bayesian Data Analysis, Design of Experiments, and Statistical Learning in a novel manner. The approach systematically minimizes the risks associated with hardware and software faults, performance limitations, and insufficient specifications to an acceptable level that achieves a Positive Risk Balance. At the core of the framework is the systematic identification and quantification of uncertainties associated with hazard scenarios and the redundantly designed system based on designed experiments, field data, and expert knowledge. The residual risk of the system is then estimated through Stochastic Simulation and evaluated by Sensitivity Analysis. By integrating these advanced analytical techniques into the V-Model, the framework fulfills, unifies, and complements existing automotive safety standards. It therefore provides a comprehensive, rigorous, and transparent safety assurance process for the development and deployment of Automated Driving Systems.
☆ Combining Machine Learning and Sampling-Based Search for Multi-Goal Motion Planning with Dynamics IJCAI 2025
This paper considers multi-goal motion planning in unstructured, obstacle-rich environments where a robot is required to reach multiple regions while avoiding collisions. The planned motions must also satisfy the differential constraints imposed by the robot dynamics. To find solutions efficiently, this paper leverages machine learning, Traveling Salesman Problem (TSP), and sampling-based motion planning. The approach expands a motion tree by adding collision-free and dynamically-feasible trajectories as branches. A TSP solver is used to compute a tour for each node to determine the order in which to reach the remaining goals by utilizing a cost matrix. An important aspect of the approach is that it leverages machine learning to construct the cost matrix by combining runtime and distance predictions to single-goal motion-planning problems. During the motion-tree expansion, priority is given to nodes associated with low-cost tours. Experiments with a vehicle model operating in obstacle-rich environments demonstrate the computational efficiency and scalability of the approach.
comment: 10 pages, 2025 International Joint Conference on Artificial Intelligence (IJCAI 2025)
☆ GAIA-2: A Controllable Multi-View Generative World Model for Autonomous Driving
Generative models offer a scalable and flexible paradigm for simulating complex environments, yet current approaches fall short in addressing the domain-specific requirements of autonomous driving - such as multi-agent interactions, fine-grained control, and multi-camera consistency. We introduce GAIA-2, Generative AI for Autonomy, a latent diffusion world model that unifies these capabilities within a single generative framework. GAIA-2 supports controllable video generation conditioned on a rich set of structured inputs: ego-vehicle dynamics, agent configurations, environmental factors, and road semantics. It generates high-resolution, spatiotemporally consistent multi-camera videos across geographically diverse driving environments (UK, US, Germany). The model integrates both structured conditioning and external latent embeddings (e.g., from a proprietary driving model) to facilitate flexible and semantically grounded scene synthesis. Through this integration, GAIA-2 enables scalable simulation of both common and rare driving scenarios, advancing the use of generative world models as a core tool in the development of autonomous systems. Videos are available at https://wayve.ai/thinking/gaia-2.
comment: Technical Report
☆ Decremental Dynamics Planning for Robot Navigation IROS 2025
Most, if not all, robot navigation systems employ a decomposed planning framework that includes global and local planning. To trade-off onboard computation and plan quality, current systems have to limit all robot dynamics considerations only within the local planner, while leveraging an extremely simplified robot representation (e.g., a point-mass holonomic model without dynamics) in the global level. However, such an artificial decomposition based on either full or zero consideration of robot dynamics can lead to gaps between the two levels, e.g., a global path based on a holonomic point-mass model may not be realizable by a non-holonomic robot, especially in highly constrained obstacle environments. Motivated by such a limitation, we propose a novel paradigm, Decremental Dynamics Planning that integrates dynamic constraints into the entire planning process, with a focus on high-fidelity dynamics modeling at the beginning and a gradual fidelity reduction as the planning progresses. To validate the effectiveness of this paradigm, we augment three different planners with DDP and show overall improved planning performance. We also develop a new DDP-based navigation system, which achieves first place in the simulation phase of the 2025 BARN Challenge. Both simulated and physical experiments validate DDP's hypothesized benefits.
comment: 7 pages. 2025 International Conference on Intelligent Robots and Systems (IROS 2025)
☆ Exploring the Effect of Robotic Embodiment and Empathetic Tone of LLMs on Empathy Elicitation
This study investigates the elicitation of empathy toward a third party through interaction with social agents. Participants engaged with either a physical robot or a voice-enabled chatbot, both driven by a large language model (LLM) programmed to exhibit either an empathetic tone or remain neutral. The interaction is focused on a fictional character, Katie Banks, who is in a challenging situation and in need of financial donations. The willingness to help Katie, measured by the number of hours participants were willing to volunteer, along with their perceptions of the agent, were assessed for 60 participants. Results indicate that neither robotic embodiment nor empathetic tone significantly influenced participants' willingness to volunteer. While the LLM effectively simulated human empathy, fostering genuine empathetic responses in participants proved challenging.
comment: *Liza Darwesh, Jaspreet Singh, Marin Marian, and Eduard Alexa contributed equally to this work.*
☆ Perspective-Shifted Neuro-Symbolic World Models: A Framework for Socially-Aware Robot Navigation
Navigating in environments alongside humans requires agents to reason under uncertainty and account for the beliefs and intentions of those around them. Under a sequential decision-making framework, egocentric navigation can naturally be represented as a Markov Decision Process (MDP). However, social navigation additionally requires reasoning about the hidden beliefs of others, inherently leading to a Partially Observable Markov Decision Process (POMDP), where agents lack direct access to others' mental states. Inspired by Theory of Mind and Epistemic Planning, we propose (1) a neuro-symbolic model-based reinforcement learning architecture for social navigation, addressing the challenge of belief tracking in partially observable environments; and (2) a perspective-shift operator for belief estimation, leveraging recent work on Influence-based Abstractions (IBA) in structured multi-agent settings.
☆ MoLe-VLA: Dynamic Layer-skipping Vision Language Action Model via Mixture-of-Layers for Efficient Robot Manipulation
Multimodal Large Language Models (MLLMs) excel in understanding complex language and visual data, enabling generalist robotic systems to interpret instructions and perform embodied tasks. Nevertheless, their real-world deployment is hindered by substantial computational and storage demands. Recent insights into the homogeneous patterns in the LLM layer have inspired sparsification techniques to address these challenges, such as early exit and token pruning. However, these methods often neglect the critical role of the final layers that encode the semantic information most relevant to downstream robotic tasks. Aligning with the recent breakthrough of the Shallow Brain Hypothesis (SBH) in neuroscience and the mixture of experts in model sparsification, we conceptualize each LLM layer as an expert and propose a Mixture-of-Layers Vision-Language-Action model (MoLe-VLA, or simply MoLe) architecture for dynamic LLM layer activation. We introduce a Spatial-Temporal Aware Router (STAR) for MoLe to selectively activate only parts of the layers based on the robot's current state, mimicking the brain's distinct signal pathways specialized for cognition and causal reasoning. Additionally, to compensate for the cognitive ability of LLMs lost in MoLe, we devise a Cognition Self-Knowledge Distillation (CogKD) framework. CogKD enhances the understanding of task demands and improves the generation of task-relevant action sequences by leveraging cognitive features. Extensive experiments conducted in both RLBench simulation and real-world environments demonstrate the superiority of MoLe-VLA in both efficiency and performance. Specifically, MoLe-VLA achieves an 8% improvement in the mean success rate across ten tasks while reducing computational costs by up to x5.6 compared to standard LLMs.
☆ CTS-CBS: A New Approach for Multi-Agent Collaborative Task Sequencing and Path Finding
This paper addresses a generalization problem of Multi-Agent Pathfinding (MAPF), called Collaborative Task Sequencing - Multi-Agent Pathfinding (CTS-MAPF), where agents must plan collision-free paths and visit a series of intermediate task locations in a specific order before reaching their final destinations. To address this problem, we propose a new approach, Collaborative Task Sequencing - Conflict-Based Search (CTS-CBS), which conducts a two-level search. In the high level, it generates a search forest, where each tree corresponds to a joint task sequence derived from the jTSP solution. In the low level, CTS-CBS performs constrained single-agent path planning to generate paths for each agent while adhering to high-level constraints. We also provide heoretical guarantees of its completeness and optimality (or sub-optimality with a bounded parameter). To evaluate the performance of CTS-CBS, we create two datasets, CTS-MAPF and MG-MAPF, and conduct comprehensive experiments. The results show that CTS-CBS adaptations for MG-MAPF outperform baseline algorithms in terms of success rate (up to 20 times larger) and runtime (up to 100 times faster), with less than a 10% sacrifice in solution quality. Furthermore, CTS-CBS offers flexibility by allowing users to adjust the sub-optimality bound omega to balance between solution quality and efficiency. Finally, practical robot tests demonstrate the algorithm's applicability in real-world scenarios.
☆ Turning Circle-based Control Barrier Function for Efficient Collision Avoidance of Nonholonomic Vehicles
This paper presents a new control barrier function (CBF) designed to improve the efficiency of collision avoidance for nonholonomic vehicles. Traditional CBFs typically rely on the shortest Euclidean distance to obstacles, overlooking the limited heading change ability of nonholonomic vehicles. This often leads to abrupt maneuvers and excessive speed reductions, which is not desirable and reduces the efficiency of collision avoidance. Our approach addresses these limitations by incorporating the distance to the turning circle, considering the vehicle's limited maneuverability imposed by its nonholonomic constraints. The proposed CBF is integrated with model predictive control (MPC) to generate more efficient trajectories compared to existing methods that rely solely on Euclidean distance-based CBFs. The effectiveness of the proposed method is validated through numerical simulations on unicycle vehicles and experiments with underactuated surface vehicles.
comment: This work has been submitted to an IEEE journal for possible publication
☆ LGR: LLM-Guided Ranking of Frontiers for Object Goal Navigation
Object Goal Navigation (OGN) is a fundamental task for robots and AI, with key applications such as mobile robot image databases (MRID). In particular, mapless OGN is essential in scenarios involving unknown or dynamic environments. This study aims to enhance recent modular mapless OGN systems by leveraging the commonsense reasoning capabilities of large language models (LLMs). Specifically, we address the challenge of determining the visiting order in frontier-based exploration by framing it as a frontier ranking problem. Our approach is grounded in recent findings that, while LLMs cannot determine the absolute value of a frontier, they excel at evaluating the relative value between multiple frontiers viewed within a single image using the view image as context. We dynamically manage the frontier list by adding and removing elements, using an LLM as a ranking model. The ranking results are represented as reciprocal rank vectors, which are ideal for multi-view, multi-query information fusion. We validate the effectiveness of our method through evaluations in Habitat-Sim.
comment: 10 pages, 11 figures, technical report
☆ A Virtual Fencing Framework for Safe and Efficient Collaborative Robotics
Collaborative robots (cobots) increasingly operate alongside humans, demanding robust real-time safeguarding. Current safety standards (e.g., ISO 10218, ANSI/RIA 15.06, ISO/TS 15066) require risk assessments but offer limited guidance for real-time responses. We propose a virtual fencing approach that detects and predicts human motion, ensuring safe cobot operation. Safety and performance tradeoffs are modeled as an optimization problem and solved via sequential quadratic programming. Experimental validation shows that our method minimizes operational pauses while maintaining safety, providing a modular solution for human-robot collaboration.
☆ Synthetic-to-Real Self-supervised Robust Depth Estimation via Learning with Motion and Structure Priors
Self-supervised depth estimation from monocular cameras in diverse outdoor conditions, such as daytime, rain, and nighttime, is challenging due to the difficulty of learning universal representations and the severe lack of labeled real-world adverse data. Previous methods either rely on synthetic inputs and pseudo-depth labels or directly apply daytime strategies to adverse conditions, resulting in suboptimal results. In this paper, we present the first synthetic-to-real robust depth estimation framework, incorporating motion and structure priors to capture real-world knowledge effectively. In the synthetic adaptation, we transfer motion-structure knowledge inside cost volumes for better robust representation, using a frozen daytime model to train a depth estimator in synthetic adverse conditions. In the innovative real adaptation, which targets to fix synthetic-real gaps, models trained earlier identify the weather-insensitive regions with a designed consistency-reweighting strategy to emphasize valid pseudo-labels. We introduce a new regularization by gathering explicit depth distributions to constrain the model when facing real-world data. Experiments show that our method outperforms the state-of-the-art across diverse conditions in multi-frame and single-frame evaluations. We achieve improvements of 7.5% and 4.3% in AbsRel and RMSE on average for nuScenes and Robotcar datasets (daytime, nighttime, rain). In zero-shot evaluation of DrivingStereo (rain, fog), our method generalizes better than the previous ones.
☆ Learning Adaptive Dexterous Grasping from Single Demonstrations
How can robots learn dexterous grasping skills efficiently and apply them adaptively based on user instructions? This work tackles two key challenges: efficient skill acquisition from limited human demonstrations and context-driven skill selection. We introduce AdaDexGrasp, a framework that learns a library of grasping skills from a single human demonstration per skill and selects the most suitable one using a vision-language model (VLM). To improve sample efficiency, we propose a trajectory following reward that guides reinforcement learning (RL) toward states close to a human demonstration while allowing flexibility in exploration. To learn beyond the single demonstration, we employ curriculum learning, progressively increasing object pose variations to enhance robustness. At deployment, a VLM retrieves the appropriate skill based on user instructions, bridging low-level learned skills with high-level intent. We evaluate AdaDexGrasp in both simulation and real-world settings, showing that our approach significantly improves RL efficiency and enables learning human-like grasp strategies across varied object configurations. Finally, we demonstrate zero-shot transfer of our learned policies to a real-world PSYONIC Ability Hand, with a 90% success rate across objects, significantly outperforming the baseline.
☆ Reasoning and Learning a Perceptual Metric for Self-Training of Reflective Objects in Bin-Picking with a Low-cost Camera
Bin-picking of metal objects using low-cost RGB-D cameras often suffers from sparse depth information and reflective surface textures, leading to errors and the need for manual labeling. To reduce human intervention, we propose a two-stage framework consisting of a metric learning stage and a self-training stage. Specifically, to automatically process data captured by a low-cost camera (LC), we introduce a Multi-object Pose Reasoning (MoPR) algorithm that optimizes pose hypotheses under depth, collision, and boundary constraints. To further refine pose candidates, we adopt a Symmetry-aware Lie-group based Bayesian Gaussian Mixture Model (SaL-BGMM), integrated with the Expectation-Maximization (EM) algorithm, for symmetry-aware filtering. Additionally, we propose a Weighted Ranking Information Noise Contrastive Estimation (WR-InfoNCE) loss to enable the LC to learn a perceptual metric from reconstructed data, supporting self-training on untrained or even unseen objects. Experimental results show that our approach outperforms several state-of-the-art methods on both the ROBI dataset and our newly introduced Self-ROBI dataset.
comment: 9 pages, 10 figures
☆ SARGes: Semantically Aligned Reliable Gesture Generation via Intent Chain
Co-speech gesture generation enhances human-computer interaction realism through speech-synchronized gesture synthesis. However, generating semantically meaningful gestures remains a challenging problem. We propose SARGes, a novel framework that leverages large language models (LLMs) to parse speech content and generate reliable semantic gesture labels, which subsequently guide the synthesis of meaningful co-speech gestures.First, we constructed a comprehensive co-speech gesture ethogram and developed an LLM-based intent chain reasoning mechanism that systematically parses and decomposes gesture semantics into structured inference steps following ethogram criteria, effectively guiding LLMs to generate context-aware gesture labels. Subsequently, we constructed an intent chain-annotated text-to-gesture label dataset and trained a lightweight gesture label generation model, which then guides the generation of credible and semantically coherent co-speech gestures. Experimental results demonstrate that SARGes achieves highly semantically-aligned gesture labeling (50.2% accuracy) with efficient single-pass inference (0.4 seconds). The proposed method provides an interpretable intent reasoning pathway for semantic gesture synthesis.
☆ Offline Reinforcement Learning with Discrete Diffusion Skills
Skills have been introduced to offline reinforcement learning (RL) as temporal abstractions to tackle complex, long-horizon tasks, promoting consistent behavior and enabling meaningful exploration. While skills in offline RL are predominantly modeled within a continuous latent space, the potential of discrete skill spaces remains largely underexplored. In this paper, we propose a compact discrete skill space for offline RL tasks supported by state-of-the-art transformer-based encoder and diffusion-based decoder. Coupled with a high-level policy trained via offline RL techniques, our method establishes a hierarchical RL framework where the trained diffusion decoder plays a pivotal role. Empirical evaluations show that the proposed algorithm, Discrete Diffusion Skill (DDS), is a powerful offline RL method. DDS performs competitively on Locomotion and Kitchen tasks and excels on long-horizon tasks, achieving at least a 12 percent improvement on AntMaze-v2 benchmarks compared to existing offline RL approaches. Furthermore, DDS offers improved interpretability, training stability, and online exploration compared to previous skill-based methods.
☆ DRPA-MPPI: Dynamic Repulsive Potential Augmented MPPI for Reactive Navigation in Unstructured Environments
Reactive mobile robot navigation in unstructured environments is challenging when robots encounter unexpected obstacles that invalidate previously planned trajectories. Model predictive path integral control (MPPI) enables reactive planning, but still suffers from limited prediction horizons that lead to local minima traps near obstacles. Current solutions rely on heuristic cost design or scenario-specific pre-training, which often limits their adaptability to new environments. We introduce dynamic repulsive potential augmented MPPI (DRPA-MPPI), which dynamically detects potential entrapments on the predicted trajectories. Upon detecting local minima, DRPA-MPPI automatically switches between standard goal-oriented optimization and a modified cost function that generates repulsive forces away from local minima. Comprehensive testing in simulated obstacle-rich environments confirms DRPA-MPPI's superior navigation performance and safety compared to conventional methods with less computational burden.
comment: 9 pages, 4 figures, Submitted to the 2025 IEEE International Conference on Automation Science and Engineering (CASE)
☆ Bandwidth Allocation for Cloud-Augmented Autonomous Driving
Autonomous vehicle (AV) control systems increasingly rely on ML models for tasks such as perception and planning. Current practice is to run these models on the car's local hardware due to real-time latency constraints and reliability concerns, which limits model size and thus accuracy. Prior work has observed that we could augment current systems by running larger models in the cloud, relying on faster cloud runtimes to offset the cellular network latency. However, prior work does not account for an important practical constraint: limited cellular bandwidth. We show that, for typical bandwidth levels, proposed techniques for cloud-augmented AV models take too long to transfer data, thus mostly falling back to the on-car models and resulting in no accuracy improvement. In this work, we show that realizing cloud-augmented AV models requires intelligent use of this scarce bandwidth, i.e. carefully allocating bandwidth across tasks and providing multiple data compression and model options. We formulate this as a resource allocation problem to maximize car utility, and present our system \sysname which achieves an increase in average model accuracy by up to 15 percentage points on driving scenarios from the Waymo Open Dataset.
comment: 18 pages, 11 figures
☆ Exploring Interference between Concurrent Skin Stretches
Proprioception is essential for coordinating human movements and enhancing the performance of assistive robotic devices. Skin stretch feedback, which closely aligns with natural proprioception mechanisms, presents a promising method for conveying proprioceptive information. To better understand the impact of interference on skin stretch perception, we conducted a user study with 30 participants that evaluated the effect of two simultaneous skin stretches on user perception. We observed that when participants experience simultaneous skin stretch stimuli, a masking effect occurs which deteriorates perception performance in the collocated skin stretch configurations. However, the perceived workload stays the same. These findings show that interference can affect the perception of skin stretch such that multi-channel skin stretch feedback designs should avoid locating modules in close proximity.
☆ Pellet-based 3D Printing of Soft Thermoplastic Elastomeric Membranes for Soft Robotic Applications
Additive Manufacturing (AM) is a promising solution for handling the complexity of fabricating soft robots. However, the AM of hyperelastic materials is still challenging with limited material types. Within this work, pellet-based 3D printing of very soft thermoplastic elastomers (TPEs) was explored. Our results show that TPEs can have similar engineering stress and maximum strain as Ecoflex OO-10. These TPEs were used to 3D-print airtight thin membranes (0.2-1.2 mm), which could inflate up to a stretch of 1320\%. Combining the membrane's large expansion and softness with the 3D printing of hollow structures simplified the design of a bending actuator that can bend 180 degrees and reach a blocked force of 238 times its weight. In addition, by 3D printing TPE pellets and rigid filaments, the soft membrane could grasp objects by enveloping an object or as a sensorized sucker, which relied on the TPE's softness to conform to the object or act as a seal. In addition, the membrane of the sucker was utilized as a tactile sensor to detect an object before adhesion. These results suggest the feasibility of 3D printing soft robots by using soft TPEs and membranes as an interesting class of materials and sensorized actuators, respectively.
A Study of Perceived Safety for Soft Robotics in Caregiving Tasks
In this project, we focus on human-robot interaction in caregiving scenarios like bathing, where physical contact is inevitable and necessary for proper task execution because force must be applied to the skin. Using finite element analysis, we designed a 3D-printed gripper combining positive and negative pressure for secure yet compliant handling. Preliminary tests showed it exerted a lower, more uniform pressure profile than a standard rigid gripper. In a user study, participants' trust in robots significantly increased after they experienced a brief bathing demonstration performed by a robotic arm equipped with the soft gripper. These results suggest that soft robotics can enhance perceived safety and acceptance in intimate caregiving scenarios.
☆ Unified Multimodal Discrete Diffusion
Multimodal generative models that can understand and generate across multiple modalities are dominated by autoregressive (AR) approaches, which process tokens sequentially from left to right, or top to bottom. These models jointly handle images, text, video, and audio for various tasks such as image captioning, question answering, and image generation. In this work, we explore discrete diffusion models as a unified generative formulation in the joint text and image domain, building upon their recent success in text generation. Discrete diffusion models offer several advantages over AR models, including improved control over quality versus diversity of generated samples, the ability to perform joint multimodal inpainting (across both text and image domains), and greater controllability in generation through guidance. Leveraging these benefits, we present the first Unified Multimodal Discrete Diffusion (UniDisc) model which is capable of jointly understanding and generating text and images for a variety of downstream tasks. We compare UniDisc to multimodal AR models, performing a scaling analysis and demonstrating that UniDisc outperforms them in terms of both performance and inference-time compute, enhanced controllability, editability, inpainting, and flexible trade-off between inference time and generation quality. Code and additional visualizations are available at https://unidisc.github.io.
comment: Project Website: https://unidisc.github.io
☆ Robust Deep Reinforcement Learning in Robotics via Adaptive Gradient-Masked Adversarial Attacks
Deep reinforcement learning (DRL) has emerged as a promising approach for robotic control, but its realworld deployment remains challenging due to its vulnerability to environmental perturbations. Existing white-box adversarial attack methods, adapted from supervised learning, fail to effectively target DRL agents as they overlook temporal dynamics and indiscriminately perturb all state dimensions, limiting their impact on long-term rewards. To address these challenges, we propose the Adaptive Gradient-Masked Reinforcement (AGMR) Attack, a white-box attack method that combines DRL with a gradient-based soft masking mechanism to dynamically identify critical state dimensions and optimize adversarial policies. AGMR selectively allocates perturbations to the most impactful state features and incorporates a dynamic adjustment mechanism to balance exploration and exploitation during training. Extensive experiments demonstrate that AGMR outperforms state-of-the-art adversarial attack methods in degrading the performance of the victim agent and enhances the victim agent's robustness through adversarial defense mechanisms.
comment: 9 pages, 6 figures
☆ Anti Robot Speciesism
Humanoid robots are a form of embodied artificial intelligence (AI) that looks and acts more and more like humans. Powered by generative AI and advances in robotics, humanoid robots can speak and interact with humans rather naturally but are still easily recognizable as robots. But how will we treat humanoids when they seem indistinguishable from humans in appearance and mind? We find a tendency (called "anti-robot" speciesism) to deny such robots humanlike capabilities, driven by motivations to accord members of the human species preferential treatment. Six experiments show that robots are denied humanlike attributes, simply because they are not biological beings and because humans want to avoid feelings of cognitive dissonance when utilizing such robots for unsavory tasks. Thus, people do not rationally attribute capabilities to perfectly humanlike robots but deny them capabilities as it suits them.
☆ In vitro 2 In vivo : Bidirectional and High-Precision Generation of In Vitro and In Vivo Neuronal Spike Data
Neurons encode information in a binary manner and process complex signals. However, predicting or generating diverse neural activity patterns remains challenging. In vitro and in vivo studies provide distinct advantages, yet no robust computational framework seamlessly integrates both data types. We address this by applying the Transformer model, widely used in large-scale language models, to neural data. To handle binary data, we introduced Dice loss, enabling accurate cross-domain neural activity generation. Structural analysis revealed how Dice loss enhances learning and identified key brain regions facilitating high-precision data generation. Our findings support the 3Rs principle in animal research, particularly Replacement, and establish a mathematical framework bridging animal experiments and human clinical studies. This work advances data-driven neuroscience and neural activity modeling, paving the way for more ethical and effective experimental methodologies.
comment: 17 pages, 5 figures
☆ TAR: Teacher-Aligned Representations via Contrastive Learning for Quadrupedal Locomotion IROS
Quadrupedal locomotion via Reinforcement Learning (RL) is commonly addressed using the teacher-student paradigm, where a privileged teacher guides a proprioceptive student policy. However, key challenges such as representation misalignment between the privileged teacher and the proprioceptive-only student, covariate shift due to behavioral cloning, and lack of deployable adaptation lead to poor generalization in real-world scenarios. We propose Teacher-Aligned Representations via Contrastive Learning (TAR), a framework that leverages privileged information with self-supervised contrastive learning to bridge this gap. By aligning representations to a privileged teacher in simulation via contrastive objectives, our student policy learns structured latent spaces and exhibits robust generalization to Out-of-Distribution (OOD) scenarios, surpassing the fully privileged "Teacher". Results showed accelerated training by 2x compared to state-of-the-art baselines to achieve peak performance. OOD scenarios showed better generalization by 40 percent on average compared to existing methods. Additionally, TAR transitions seamlessly into learning during deployment without requiring privileged states, setting a new benchmark in sample-efficient, adaptive locomotion and enabling continual fine-tuning in real-world scenarios. Open-source code and videos are available at https://ammousa.github.io/TARLoco/.
comment: This work has been submitted to the IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) 2025 for review
OTTER: A Vision-Language-Action Model with Text-Aware Visual Feature Extraction
Vision-Language-Action (VLA) models aim to predict robotic actions based on visual observations and language instructions. Existing approaches require fine-tuning pre-trained visionlanguage models (VLMs) as visual and language features are independently fed into downstream policies, degrading the pre-trained semantic alignments. We propose OTTER, a novel VLA architecture that leverages these existing alignments through explicit, text-aware visual feature extraction. Instead of processing all visual features, OTTER selectively extracts and passes only task-relevant visual features that are semantically aligned with the language instruction to the policy transformer. This allows OTTER to keep the pre-trained vision-language encoders frozen. Thereby, OTTER preserves and utilizes the rich semantic understanding learned from large-scale pre-training, enabling strong zero-shot generalization capabilities. In simulation and real-world experiments, OTTER significantly outperforms existing VLA models, demonstrating strong zeroshot generalization to novel objects and environments. Video, code, checkpoints, and dataset: https://ottervla.github.io/.
♻ ☆ DexHandDiff: Interaction-aware Diffusion Planning for Adaptive Dexterous Manipulation CVPR 2025
Dexterous manipulation with contact-rich interactions is crucial for advanced robotics. While recent diffusion-based planning approaches show promise for simple manipulation tasks, they often produce unrealistic ghost states (e.g., the object automatically moves without hand contact) or lack adaptability when handling complex sequential interactions. In this work, we introduce DexHandDiff, an interaction-aware diffusion planning framework for adaptive dexterous manipulation. DexHandDiff models joint state-action dynamics through a dual-phase diffusion process which consists of pre-interaction contact alignment and post-contact goal-directed control, enabling goal-adaptive generalizable dexterous manipulation. Additionally, we incorporate dynamics model-based dual guidance and leverage large language models for automated guidance function generation, enhancing generalizability for physical interactions and facilitating diverse goal adaptation through language cues. Experiments on physical interaction tasks such as door opening, pen and block re-orientation, object relocation, and hammer striking demonstrate DexHandDiff's effectiveness on goals outside training distributions, achieving over twice the average success rate (59.2% vs. 29.5%) compared to existing methods. Our framework achieves an average of 70.7% success rate on goal adaptive dexterous tasks, highlighting its robustness and flexibility in contact-rich manipulation.
comment: Accepted by CVPR 2025. Camera ready version. Previous DexDiffuser. Project page: https://dexdiffuser.github.io/
♻ ☆ UAV-Assisted Self-Supervised Terrain Awareness for Off-Road Navigation ICRA 2025
Terrain awareness is an essential milestone to enable truly autonomous off-road navigation. Accurately predicting terrain characteristics allows optimizing a vehicle's path against potential hazards. Recent methods use deep neural networks to predict traversability-related terrain properties in a self-supervised manner, relying on proprioception as a training signal. However, onboard cameras are inherently limited by their point-of-view relative to the ground, suffering from occlusions and vanishing pixel density with distance. This paper introduces a novel approach for self-supervised terrain characterization using an aerial perspective from a hovering drone. We capture terrain-aligned images while sampling the environment with a ground vehicle, effectively training a simple predictor for vibrations, bumpiness, and energy consumption. Our dataset includes 2.8 km of off-road data collected in forest environment, comprising 13 484 ground-based images and 12 935 aerial images. Our findings show that drone imagery improves terrain property prediction by 21.37 % on the whole dataset and 37.35 % in high vegetation, compared to ground robot images. We conduct ablation studies to identify the main causes of these performance improvements. We also demonstrate the real-world applicability of our approach by scouting an unseen area with a drone, planning and executing an optimized path on the ground.
comment: 7 pages, 5 figures, submitted to ICRA 2025
♻ ☆ Contractive Dynamical Imitation Policies for Efficient Out-of-Sample Recovery
Imitation learning is a data-driven approach to learning policies from expert behavior, but it is prone to unreliable outcomes in out-of-sample (OOS) regions. While previous research relying on stable dynamical systems guarantees convergence to a desired state, it often overlooks transient behavior. We propose a framework for learning policies modeled by contractive dynamical systems, ensuring that all policy rollouts converge regardless of perturbations, and in turn, enable efficient OOS recovery. By leveraging recurrent equilibrium networks and coupling layers, the policy structure guarantees contractivity for any parameter choice, which facilitates unconstrained optimization. We also provide theoretical upper bounds for worst-case and expected loss to rigorously establish the reliability of our method in deployment. Empirically, we demonstrate substantial OOS performance improvements for simulated robotic manipulation and navigation tasks.
comment: International Conference on Learning Representations
♻ ☆ The Morphology-Control Trade-Off: Insights into Soft Robotic Efficiency
Soft robotics holds transformative potential for enabling adaptive and adaptable systems in dynamic environments. However, the interplay between morphological and control complexities and their collective impact on task performance remains poorly understood. Therefore, in this study, we investigate these trade-offs across tasks of differing difficulty levels using four well-used morphological complexity metrics and control complexity measured by FLOPs. We investigate how these factors jointly influence task performance by utilizing the evolutionary robot experiments. Results show that optimal performance depends on the alignment between morphology and control: simpler morphologies and lightweight controllers suffice for easier tasks, while harder tasks demand higher complexities in both dimensions. In addition, a clear trade-off between morphological and control complexities that achieve the same task performance can be observed. Moreover, we also propose a sensitivity analysis to expose the task-specific contributions of individual morphological metrics. Our study establishes a framework for investigating the relationships between morphology, control, and task performance, advancing the development of task-specific robotic designs that balance computational efficiency with adaptability. This study contributes to the practical application of soft robotics in real-world scenarios by providing actionable insights.
comment: The paper is planed to be submitted to a journal
♻ ☆ ManiCM: Real-time 3D Diffusion Policy via Consistency Model for Robotic Manipulation
Diffusion models have been verified to be effective in generating complex distributions from natural images to motion trajectories. Recent diffusion-based methods show impressive performance in 3D robotic manipulation tasks, whereas they suffer from severe runtime inefficiency due to multiple denoising steps, especially with high-dimensional observations. To this end, we propose a real-time robotic manipulation model named ManiCM that imposes the consistency constraint to the diffusion process, so that the model can generate robot actions in only one-step inference. Specifically, we formulate a consistent diffusion process in the robot action space conditioned on the point cloud input, where the original action is required to be directly denoised from any point along the ODE trajectory. To model this process, we design a consistency distillation technique to predict the action sample directly instead of predicting the noise within the vision community for fast convergence in the low-dimensional action manifold. We evaluate ManiCM on 31 robotic manipulation tasks from Adroit and Metaworld, and the results demonstrate that our approach accelerates the state-of-the-art method by 10 times in average inference speed while maintaining competitive average success rate.
comment: https://manicm-fast.github.io/
♻ ☆ LaMOuR: Leveraging Language Models for Out-of-Distribution Recovery in Reinforcement Learning
Deep Reinforcement Learning (DRL) has demonstrated strong performance in robotic control but remains susceptible to out-of-distribution (OOD) states, often resulting in unreliable actions and task failure. While previous methods have focused on minimizing or preventing OOD occurrences, they largely neglect recovery once an agent encounters such states. Although the latest research has attempted to address this by guiding agents back to in-distribution states, their reliance on uncertainty estimation hinders scalability in complex environments. To overcome this limitation, we introduce Language Models for Out-of-Distribution Recovery (LaMOuR), which enables recovery learning without relying on uncertainty estimation. LaMOuR generates dense reward codes that guide the agent back to a state where it can successfully perform its original task, leveraging the capabilities of LVLMs in image description, logical reasoning, and code generation. Experimental results show that LaMOuR substantially enhances recovery efficiency across diverse locomotion tasks and even generalizes effectively to complex environments, including humanoid locomotion and mobile manipulation, where existing methods struggle. The code and supplementary materials are available at https://lamour-rl.github.io/.
comment: 14 pages, 17 figures
♻ ☆ Optimizing Robot Programming: Mixed Reality Gripper Control ICRA 2025
Conventional robot programming methods are complex and time-consuming for users. In recent years, alternative approaches such as mixed reality have been explored to address these challenges and optimize robot programming. While the findings of the mixed reality robot programming methods are convincing, most existing methods rely on gesture interaction for robot programming. Since controller-based interactions have proven to be more reliable, this paper examines three controller-based programming methods within a mixed reality scenario: 1) Classical Jogging, where the user positions the robot's end effector using the controller's thumbsticks, 2) Direct Control, where the controller's position and orientation directly corresponds to the end effector's, and 3) Gripper Control, where the controller is enhanced with a 3D-printed gripper attachment to grasp and release objects. A within-subjects study (n = 30) was conducted to compare these methods. The findings indicate that the Gripper Control condition outperforms the others in terms of task completion time, user experience, mental demand, and task performance, while also being the preferred method. Therefore, it demonstrates promising potential as an effective and efficient approach for future robot programming. Video available at https://youtu.be/83kWr8zUFIQ.
comment: Accepted to ICRA 2025
♻ ☆ Polytope Volume Monitoring Problem: Formulation and Solution via Parametric Linear Program Based Control Barrier Function
Motivated by the latest research on feasible space monitoring of multiple control barrier functions (CBFs) as well as polytopic collision avoidance, this paper studies the Polytope Volume Monitoring (PVM) problem, whose goal is to design a control law for inputs of nonlinear systems to prevent the volume of some state-dependent polytope from decreasing to zero. Recent studies have explored the idea of applying Chebyshev ball method in optimization theory to solve the case study of PVM; however, the underlying difficulties caused by nonsmoothness have not been addressed. This paper continues the study on this topic, where our main contribution is to establish the relationship between nonsmooth CBF and parametric optimization theory through directional derivatives for the first time, so as to solve PVM problems more conveniently. In detail, inspired by Chebyshev ball approach, a parametric linear program (PLP) based nonsmooth barrier function candidate is established for PVM, and then, sufficient conditions for it to be a nonsmooth CBF are proposed, based on which a quadratic program (QP) based safety filter with guaranteed feasibility is proposed to address PVM problems. Finally, a numerical simulation example is given to show the efficiency of the proposed safety filter.
comment: A simplified version is submitted to CDC2025
♻ ☆ Socratic Planner: Self-QA-Based Zero-Shot Planning for Embodied Instruction Following ICRA 2025
Embodied Instruction Following (EIF) is the task of executing natural language instructions by navigating and interacting with objects in interactive environments. A key challenge in EIF is compositional task planning, typically addressed through supervised learning or few-shot in-context learning with labeled data. To this end, we introduce the Socratic Planner, a self-QA-based zero-shot planning method that infers an appropriate plan without any further training. The Socratic Planner first facilitates self-questioning and answering by the Large Language Model (LLM), which in turn helps generate a sequence of subgoals. While executing the subgoals, an embodied agent may encounter unexpected situations, such as unforeseen obstacles. The Socratic Planner then adjusts plans based on dense visual feedback through a visually-grounded re-planning mechanism. Experiments demonstrate the effectiveness of the Socratic Planner, outperforming current state-of-the-art planning models on the ALFRED benchmark across all metrics, particularly excelling in long-horizon tasks that demand complex inference. We further demonstrate its real-world applicability through deployment on a physical robot for long-horizon tasks.
comment: 8 pages, 6 figures, published to ICRA 2025
♻ ☆ RoboSpatial: Teaching Spatial Understanding to 2D and 3D Vision-Language Models for Robotics CVPR 2025
Spatial understanding is a crucial capability that enables robots to perceive their surroundings, reason about their environment, and interact with it meaningfully. In modern robotics, these capabilities are increasingly provided by vision-language models. However, these models face significant challenges in spatial reasoning tasks, as their training data are based on general-purpose image datasets that often lack sophisticated spatial understanding. For example, datasets frequently do not capture reference frame comprehension, yet effective spatial reasoning requires understanding whether to reason from ego-, world-, or object-centric perspectives. To address this issue, we introduce RoboSpatial, a large-scale dataset for spatial understanding in robotics. It consists of real indoor and tabletop scenes, captured as 3D scans and egocentric images, and annotated with rich spatial information relevant to robotics. The dataset includes 1M images, 5k 3D scans, and 3M annotated spatial relationships, and the pairing of 2D egocentric images with 3D scans makes it both 2D- and 3D- ready. Our experiments show that models trained with RoboSpatial outperform baselines on downstream tasks such as spatial affordance prediction, spatial relationship prediction, and robot manipulation.
comment: CVPR 2025
♻ ☆ TopV-Nav: Unlocking the Top-View Spatial Reasoning Potential of MLLM for Zero-shot Object Navigation
The Zero-Shot Object Navigation (ZSON) task requires embodied agents to find a previously unseen object by navigating in unfamiliar environments. Such a goal-oriented exploration heavily relies on the ability to perceive, understand, and reason based on the spatial information of the environment. However, current LLM-based approaches convert visual observations to language descriptions and reason in the linguistic space, leading to the loss of spatial information. In this paper, we introduce TopV-Nav, an MLLM-based method that directly reasons on the top-view map with sufficient spatial information. To fully unlock the MLLM's spatial reasoning potential in top-view perspective, we propose the Adaptive Visual Prompt Generation (AVPG) method to adaptively construct semantically-rich top-view map. It enables the agent to directly utilize spatial information contained in the top-view map to conduct thorough reasoning. Besides, we design a Dynamic Map Scaling (DMS) mechanism to dynamically zoom top-view map at preferred scales, enhancing local fine-grained reasoning. Additionally, we devise a Potential Target Driven (PTD) mechanism to predict and to utilize target locations, facilitating global and human-like exploration. Experiments on MP3D and HM3D datasets demonstrate the superiority of our TopV-Nav.
comment: 10 pages
♻ ☆ General-purpose Clothes Manipulation with Semantic Keypoints ICRA 2025
Clothes manipulation is a critical capability for household robots; yet, existing methods are often confined to specific tasks, such as folding or flattening, due to the complex high-dimensional geometry of deformable fabric. This paper presents CLothes mAnipulation with Semantic keyPoints (CLASP) for general-purpose clothes manipulation, which enables the robot to perform diverse manipulation tasks over different types of clothes. The key idea of CLASP is semantic keypoints -- e.g., "right shoulder", "left sleeve", etc. -- a sparse spatial-semantic representation that is salient for both perception and action. Semantic keypoints of clothes can be effectively extracted from depth images and are sufficient to represent a broad range of clothes manipulation policies. CLASP leverages semantic keypoints to bridge LLM-powered task planning and low-level action execution in a two-level hierarchy. Extensive simulation experiments show that CLASP outperforms baseline methods across diverse clothes types in both seen and unseen tasks. Further, experiments with a Kinova dual-arm system on four distinct tasks -- folding, flattening, hanging, and placing -- confirm CLASP's performance on a real robot.
comment: accepted by IEEE International Conference on Robotics and Automation (ICRA 2025)
♻ ☆ Distributed Motion Control of Multiple Mobile Manipulators for Reducing Interaction Wrench in Object Manipulation
In real-world cooperative manipulation of objects, multiple mobile manipulator systems may suffer from disturbances and asynchrony, leading to excessive interaction wrenches and potentially causing object damage or emergency stops. Existing methods often rely on torque control and dynamic models, which are uncommon in many industrial robots and settings. Additionally, dynamic models often neglect joint friction forces and are not accurate. These methods are challenging to implement and validate in physical systems. To address the problems, this paper presents a novel distributed motion control approach aimed at reducing these unnecessary interaction wrenches. The control law is only based on local information and joint velocity control to enhance practical applicability. The communication delays within the distributed architecture are considered. The stability of the control law is rigorously proven by the Lyapunov theorem. In the simulations, the effectiveness is shown, and the impact of communication graph connectivity and communication delays has been studied. A comparison with other methods shows the advantages of the proposed control law in terms of convergence speed and robustness. Finally, the control law has been validated in physical experiments. It does not require dynamic modeling or torque control, and thus is more user-friendly for physical robots.
♻ ☆ FoAM: Foresight-Augmented Multi-Task Imitation Policy for Robotic Manipulation
Multi-task imitation learning (MTIL) has shown significant potential in robotic manipulation by enabling agents to perform various tasks using a single policy. This simplifies the policy deployment and enhances the agent's adaptability across different scenarios. However, key challenges remain, such as maintaining action reliability (e.g., avoiding abnormal action sequences that deviate from nominal task trajectories) and generalizing to unseen tasks with a few expert demonstrations. To address these challenges, we introduce the Foresight-Augmented Manipulation Policy (FoAM), a novel MTIL policy that pioneers the use of multi-modal goal condition as input and introduces a foresight augmentation in addition to the general action reconstruction. FoAM enables the agent to reason about the visual consequences (states) of its actions and learn more expressive embedding that captures nuanced task variations. Extensive experiments on over 100 tasks in simulation and real-world settings demonstrate that FoAM significantly enhances MTIL policy performance, outperforming state-of-the-art baselines by up to 41% in success rate. Meanwhile, we released our simulation suites, including a total of 10 scenarios and over 80 challenging tasks designed for manipulation policy training and evaluation. See the project homepage projFoAM.github.io for project details.
comment: 8 pages, 4 figures
♻ ☆ Vision-based Multi-future Trajectory Prediction: A Survey
Vision-based trajectory prediction is an important task that supports safe and intelligent behaviours in autonomous systems. Many advanced approaches have been proposed over the years with improved spatial and temporal feature extraction. However, human behaviour is naturally diverse and uncertain. Given the past trajectory and surrounding environment information, an agent can have multiple plausible trajectories in the future. To tackle this problem, an essential task named multi-future trajectory prediction (MTP) has recently been studied. This task aims to generate a diverse, acceptable and explainable distribution of future predictions for each agent. In this paper, we present the first survey for MTP with our unique taxonomies and a comprehensive analysis of frameworks, datasets and evaluation metrics. We also compare models on existing MTP datasets and conduct experiments on the ForkingPath dataset. Finally, we discuss multiple future directions that can help researchers develop novel multi-future trajectory prediction systems and other diverse learning tasks similar to MTP.
comment: Accepted by TNNLS 2025
♻ ☆ SLIM: Scalable and Lightweight LiDAR Mapping in Urban Environments
LiDAR point cloud maps are extensively utilized on roads for robot navigation due to their high consistency. However, dense point clouds face challenges of high memory consumption and reduced maintainability for long-term operations. In this study, we introduce SLIM, a scalable and lightweight mapping system for long-term LiDAR mapping in urban environments. The system begins by parameterizing structural point clouds into lines and planes. These lightweight and structural representations meet the requirements of map merging, pose graph optimization, and bundle adjustment, ensuring incremental management and local consistency. For long-term operations, a map-centric nonlinear factor recovery method is designed to sparsify poses while preserving mapping accuracy. We validate the SLIM system with multi-session real-world LiDAR data from classical LiDAR mapping datasets, including KITTI, NCLT, HeLiPR and M2DGR. The experiments demonstrate its capabilities in mapping accuracy, lightweightness, and scalability. Map re-use is also verified through map-based robot localization. Finally, with multi-session LiDAR data, the SLIM system provides a globally consistent map with low memory consumption (~130 KB/km on KITTI).
comment: Accepted for publication in IEEE Transactions on Robotics. Video: https://youtu.be/8HQnYMf_BWI Code: https://github.com/HKUST-Aerial-Robotics/SLIM
♻ ☆ NuRF: Nudging the Particle Filter in Radiance Fields for Robot Visual Localization
Can we localize a robot on a map only using monocular vision? This study presents NuRF, an adaptive and nudged particle filter framework in radiance fields for 6-DoF robot visual localization. NuRF leverages recent advancements in radiance fields and visual place recognition. Conventional visual place recognition meets the challenges of data sparsity and artifact-induced inaccuracies. By utilizing radiance field-generated novel views, NuRF enhances visual localization performance and combines coarse global localization with the fine-grained pose tracking of a particle filter, ensuring continuous and precise localization. Experimentally, our method converges 7 times faster than existing Monte Carlo-based methods and achieves localization accuracy within 1 meter, offering an efficient and resilient solution for indoor visual localization.
comment: Accepted for Publication in IEEE Transactions on Cognitive and Developmental Systems
Inference-Time Policy Steering through Human Interactions ICRA 2025
Generative policies trained with human demonstrations can autonomously accomplish multimodal, long-horizon tasks. However, during inference, humans are often removed from the policy execution loop, limiting the ability to guide a pre-trained policy towards a specific sub-goal or trajectory shape among multiple predictions. Naive human intervention may inadvertently exacerbate distribution shift, leading to constraint violations or execution failures. To better align policy output with human intent without inducing out-of-distribution errors, we propose an Inference-Time Policy Steering (ITPS) framework that leverages human interactions to bias the generative sampling process, rather than fine-tuning the policy on interaction data. We evaluate ITPS across three simulated and real-world benchmarks, testing three forms of human interaction and associated alignment distance metrics. Among six sampling strategies, our proposed stochastic sampling with diffusion policy achieves the best trade-off between alignment and distribution shift. Videos are available at https://yanweiw.github.io/itps/.
comment: ICRA 2025
♻ ☆ Hybrid Physics-ML Modeling for Marine Vehicle Maneuvering Motions in the Presence of Environmental Disturbances
A hybrid physics-machine learning modeling framework is proposed for the surface vehicles' maneuvering motions to address the modeling capability and stability in the presence of environmental disturbances. From a deep learning perspective, the framework is based on a variant version of residual networks with additional feature extraction. Initially, an imperfect physical model is derived and identified to capture the fundamental hydrodynamic characteristics of marine vehicles. This model is then integrated with a feedforward network through a residual block. Additionally, feature extraction from trigonometric transformations is employed in the machine learning component to account for the periodic influence of currents and waves. The proposed method is evaluated using real navigational data from the 'JH7500' unmanned surface vehicle. The results demonstrate the robust generalizability and accurate long-term prediction capabilities of the nonlinear dynamic model in specific environmental conditions. This approach has the potential to be extended and applied to develop a comprehensive high-fidelity simulator.
comment: The content of the manuscript will undergo significant revisions
♻ ☆ Automated Vehicle Driver Monitoring Dataset from Real-World Scenarios
From SAE Level 3 of automation onwards, drivers are allowed to engage in activities that are not directly related to driving during their travel. However, in level 3, a misunderstanding of the capabilities of the system might lead drivers to engage in secondary tasks, which could impair their ability to react to challenging traffic situations. Anticipating driver activity allows for early detection of risky behaviors, to prevent accidents. To be able to predict the driver activity, a Deep Learning network needs to be trained on a dataset. However, the use of datasets based on simulation for training and the migration to real-world data for prediction has proven to be suboptimal. Hence, this paper presents a real-world driver activity dataset, openly accessible on IEEE Dataport, which encompasses various activities that occur in autonomous driving scenarios under various illumination and weather conditions. Results from the training process showed that the dataset provides an excellent benchmark for implementing models for driver activity recognition.
comment: 6 pages
♻ ☆ Multi-Agent Inverse Reinforcement Learning in Real World Unstructured Pedestrian Crowds
Social robot navigation in crowded public spaces such as university campuses, restaurants, grocery stores, and hospitals, is an increasingly important area of research. One of the core strategies for achieving this goal is to understand humans' intent--underlying psychological factors that govern their motion--by learning their reward functions, typically via inverse reinforcement learning (IRL). Despite significant progress in IRL, learning reward functions of multiple agents simultaneously in dense unstructured pedestrian crowds has remained intractable due to the nature of the tightly coupled social interactions that occur in these scenarios \textit{e.g.} passing, intersections, swerving, weaving, etc. In this paper, we present a new multi-agent maximum entropy inverse reinforcement learning algorithm for real world unstructured pedestrian crowds. Key to our approach is a simple, but effective, mathematical trick which we name the so-called tractability-rationality trade-off trick that achieves tractability at the cost of a slight reduction in accuracy. We compare our approach to the classical single-agent MaxEnt IRL as well as state-of-the-art trajectory prediction methods on several datasets including the ETH, UCY, SCAND, JRDB, and a new dataset, called Speedway, collected at a busy intersection on a University campus focusing on dense, complex agent interactions. Our key findings show that, on the dense Speedway dataset, our approach ranks 1st among top 7 baselines with >2X improvement over single-agent IRL, and is competitive with state-of-the-art large transformer-based encoder-decoder models on sparser datasets such as ETH/UCY (ranks 3rd among top 7 baselines).
♻ ☆ SimBEV: A Synthetic Multi-Task Multi-Sensor Driving Data Generation Tool and Dataset
Bird's-eye view (BEV) perception has garnered significant attention in autonomous driving in recent years, in part because BEV representation facilitates multi-modal sensor fusion. BEV representation enables a variety of perception tasks including BEV segmentation, a concise view of the environment useful for planning a vehicle's trajectory. However, this representation is not fully supported by existing datasets, and creation of new datasets for this purpose can be a time-consuming endeavor. To address this challenge, we introduce SimBEV. SimBEV is a randomized synthetic data generation tool that is extensively configurable and scalable, supports a wide array of sensors, incorporates information from multiple sources to capture accurate BEV ground truth, and enables a variety of perception tasks including BEV segmentation and 3D object detection. SimBEV is used to create the SimBEV dataset, a large collection of annotated perception data from diverse driving scenarios. SimBEV and the SimBEV dataset are open and available to the public.
Artificial Intelligence 155
☆ Mobile-MMLU: A Mobile Intelligence Language Understanding Benchmark
Rapid advancements in large language models (LLMs) have increased interest in deploying them on mobile devices for on-device AI applications. Mobile users interact differently with LLMs compared to desktop users, creating unique expectations and data biases. Current benchmark datasets primarily target at server and desktop environments, and there is a notable lack of extensive datasets specifically designed for mobile contexts. Additionally, mobile devices face strict limitations in storage and computing resources, constraining model size and capabilities, thus requiring optimized efficiency and prioritized knowledge. To address these challenges, we introduce Mobile-MMLU, a large-scale benchmark dataset tailored for mobile intelligence. It consists of 16,186 questions across 80 mobile-related fields, designed to evaluate LLM performance in realistic mobile scenarios. A challenging subset, Mobile-MMLU-Pro, provides advanced evaluation similar in size to MMLU-Pro but significantly more difficult than our standard full set. Both benchmarks use multiple-choice, order-invariant questions focused on practical mobile interactions, such as recipe suggestions, travel planning, and essential daily tasks. The dataset emphasizes critical mobile-specific metrics like inference latency, energy consumption, memory usage, and response quality, offering comprehensive insights into model performance under mobile constraints. Moreover, it prioritizes privacy and adaptability, assessing models' ability to perform on-device processing, maintain user privacy, and adapt to personalized usage patterns. Mobile-MMLU family offers a standardized framework for developing and comparing mobile-optimized LLMs, enabling advancements in productivity and decision-making within mobile computing environments. Our code and data are available at: https://github.com/VILA-Lab/Mobile-MMLU.
comment: An order-invariant and mobile-centric benchmark. Code and data are available at: https://github.com/VILA-Lab/Mobile-MMLU
☆ Understanding R1-Zero-Like Training: A Critical Perspective
DeepSeek-R1-Zero has shown that reinforcement learning (RL) at scale can directly enhance the reasoning capabilities of LLMs without supervised fine-tuning. In this work, we critically examine R1-Zero-like training by analyzing its two core components: base models and RL. We investigate a wide range of base models, including DeepSeek-V3-Base, to understand how pretraining characteristics influence RL performance. Our analysis reveals that DeepSeek-V3-Base already exhibit ''Aha moment'', while Qwen2.5 base models demonstrate strong reasoning capabilities even without prompt templates, suggesting potential pretraining biases. Additionally, we identify an optimization bias in Group Relative Policy Optimization (GRPO), which artificially increases response length (especially for incorrect outputs) during training. To address this, we introduce Dr. GRPO, an unbiased optimization method that improves token efficiency while maintaining reasoning performance. Leveraging these insights, we present a minimalist R1-Zero recipe that achieves 43.3% accuracy on AIME 2024 with a 7B base model, establishing a new state-of-the-art. Our code is available at https://github.com/sail-sg/understand-r1-zero.
☆ ADS-Edit: A Multimodal Knowledge Editing Dataset for Autonomous Driving Systems
Recent advancements in Large Multimodal Models (LMMs) have shown promise in Autonomous Driving Systems (ADS). However, their direct application to ADS is hindered by challenges such as misunderstanding of traffic knowledge, complex road conditions, and diverse states of vehicle. To address these challenges, we propose the use of Knowledge Editing, which enables targeted modifications to a model's behavior without the need for full retraining. Meanwhile, we introduce ADS-Edit, a multimodal knowledge editing dataset specifically designed for ADS, which includes various real-world scenarios, multiple data types, and comprehensive evaluation metrics. We conduct comprehensive experiments and derive several interesting conclusions. We hope that our work will contribute to the further advancement of knowledge editing applications in the field of autonomous driving. Code and data are available in https://github.com/zjunlp/EasyEdit.
comment: Work in progress
☆ Reason-RFT: Reinforcement Fine-Tuning for Visual Reasoning
Visual reasoning abilities play a crucial role in understanding complex multimodal data, advancing both domain-specific applications and artificial general intelligence (AGI). Existing methods improve VLM reasoning via Chain-of-Thought (CoT) supervised fine-tuning, using meticulously annotated training data to enhance visual reasoning capabilities. However, this training paradigm may lead to overfitting and cognitive rigidity, restricting the model's ability to transfer visual reasoning skills across domains and limiting its real-world applicability. To address these limitations, we propose Reason-RFT, a novel reinforcement fine-tuning framework that significantly enhances generalization capabilities in visual reasoning tasks. Reason-RFT introduces a two-phase training framework for visual reasoning: (1) Supervised Fine-Tuning (SFT) with curated Chain-of-Thought (CoT) data activates the reasoning potential of Vision-Language Models (VLMs), followed by (2) Group Relative Policy Optimization (GRPO)-based reinforcement learning that generates multiple reasoning-response pairs, significantly enhancing generalization in visual reasoning tasks. To evaluate Reason-RFT's visual reasoning capabilities, we reconstructed a comprehensive dataset spanning visual counting, structure perception, and spatial transformation.cExperimental results demonstrate Reasoning-RFT's three key advantages: (1) Performance Enhancement: achieving state-of-the-art results across multiple tasks, outperforming most mainstream open-source and proprietary models; (2) Generalization Superiority: consistently maintaining robust performance across diverse tasks and domains, outperforming alternative training paradigms; (3) Data Efficiency: excelling in few-shot learning scenarios while surpassing full-dataset SFT baselines.
comment: 35 pages, 22 figures
☆ Optimal Scaling Laws for Efficiency Gains in a Theoretical Transformer-Augmented Sectional MoE Framework
This paper introduces a theoretical framework for a Transformer-augmented, sectional Mixture-of-Experts (MoE) architecture that aims to enhance computational efficiency while preserving model scalability. Unlike conventional MoE models, which route entire token embeddings to selected experts, our approach portions the embedding dimension itself -- assigning segments of each token's representation to dedicated experts. To combat losses in token representation, we utilize a pre-expert transformer layer to recompute attention across tokens and reduce the sequence length dimensionality. We extend our theory by deriving optimal scaling laws that a non-linear relationship between the number of experts and factors such as model dimensionality, sequence length, and system overhead. These formulations yield closed-form and numerically-solvable expressions for identifying the optimal expert count under given architectural and hardware constraints. As a result, our framework not only provides theoretical bounds for computing efficiency with varying frameworks but also guides practical design choices for scaling large models effectively. While empirical validation is pending, we present a comprehensive experimental road map to evaluate the framework's efficiency, scalability, and practicality in future work.
☆ High Quality Diffusion Distillation on a Single GPU with Relative and Absolute Position Matching
We introduce relative and absolute position matching (RAPM), a diffusion distillation method resulting in high quality generation that can be trained efficiently on a single GPU. Recent diffusion distillation research has achieved excellent results for high-resolution text-to-image generation with methods such as phased consistency models (PCM) and improved distribution matching distillation (DMD2). However, these methods generally require many GPUs (e.g.~8-64) and significant batchsizes (e.g.~128-2048) during training, resulting in memory and compute requirements that are beyond the resources of some researchers. RAPM provides effective single-GPU diffusion distillation training with a batchsize of 1. The new method attempts to mimic the sampling trajectories of the teacher model by matching the relative and absolute positions. The design of relative positions is inspired by PCM. Two discriminators are introduced accordingly in RAPM, one for matching relative positions and the other for absolute positions. Experimental results on StableDiffusion (SD) V1.5 and SDXL indicate that RAPM with 4 timesteps produces comparable FID scores as the best method with 1 timestep under very limited computational resources.
☆ Quantum Neural Network Restatement of Markov Jump Process
Despite the many challenges in exploratory data analysis, artificial neural networks have motivated strong interests in scientists and researchers both in theoretical as well as practical applications. Among sources of such popularity of artificial neural networks the ability of modeling non-linear dynamical systems, generalization, and adaptation possibilities should be mentioned. Despite this, there is still significant debate about the role of various underlying stochastic processes in stabilizing a unique structure for data learning and prediction. One of such obstacles to the theoretical and numerical study of machine intelligent systems is the curse of dimensionality and the sampling from high-dimensional probability distributions. In general, this curse prevents efficient description of states, providing a significant complexity barrier for the system to be efficiently described and studied. In this strand of research, direct treatment and description of such abstract notions of learning theory in terms of quantum information be one of the most favorable candidates. Hence, the subject matter of these articles is devoted to problems of design, adaptation and the formulations of computationally hard problems in terms of quantum mechanical systems. In order to characterize the microscopic description of such dynamics in the language of inferential statistics, covariance matrix estimation of d-dimensional Gaussian densities and Bayesian interpretation of eigenvalue problem for dynamical systems is assessed.
☆ Emotion Detection and Music Recommendation System
As artificial intelligence becomes more and more ingrained in daily life, we present a novel system that uses deep learning for music recommendation and emotion-based detection. Through the use of facial recognition and the DeepFace framework, our method analyses human emotions in real-time and then plays music that reflects the mood it has discovered. The system uses a webcam to take pictures, analyses the most common facial expression, and then pulls a playlist from local storage that corresponds to the mood it has detected. An engaging and customised experience is ensured by allowing users to manually change the song selection via a dropdown menu or navigation buttons. By continuously looping over the playlist, the technology guarantees continuity. The objective of our system is to improve emotional well-being through music therapy by offering a responsive and automated music-selection experience.
☆ Graph-Enhanced Model-Free Reinforcement Learning Agents for Efficient Power Grid Topological Control
The increasing complexity of power grid management, driven by the emergence of prosumers and the demand for cleaner energy solutions, has needed innovative approaches to ensure stability and efficiency. This paper presents a novel approach within the model-free framework of reinforcement learning, aimed at optimizing power network operations without prior expert knowledge. We introduce a masked topological action space, enabling agents to explore diverse strategies for cost reduction while maintaining reliable service using the state logic as a guide for choosing proper actions. Through extensive experimentation across 20 different scenarios in a simulated 5-substation environment, we demonstrate that our approach achieves a consistent reduction in power losses, while ensuring grid stability against potential blackouts. The results underscore the effectiveness of combining dynamic observation formalization with opponent-based training, showing a viable way for autonomous management solutions in modern energy systems or even for building a foundational model for this field.
☆ Flip Learning: Weakly Supervised Erase to Segment Nodules in Breast Ultrasound
Accurate segmentation of nodules in both 2D breast ultrasound (BUS) and 3D automated breast ultrasound (ABUS) is crucial for clinical diagnosis and treatment planning. Therefore, developing an automated system for nodule segmentation can enhance user independence and expedite clinical analysis. Unlike fully-supervised learning, weakly-supervised segmentation (WSS) can streamline the laborious and intricate annotation process. However, current WSS methods face challenges in achieving precise nodule segmentation, as many of them depend on inaccurate activation maps or inefficient pseudo-mask generation algorithms. In this study, we introduce a novel multi-agent reinforcement learning-based WSS framework called Flip Learning, which relies solely on 2D/3D boxes for accurate segmentation. Specifically, multiple agents are employed to erase the target from the box to facilitate classification tag flipping, with the erased region serving as the predicted segmentation mask. The key contributions of this research are as follows: (1) Adoption of a superpixel/supervoxel-based approach to encode the standardized environment, capturing boundary priors and expediting the learning process. (2) Introduction of three meticulously designed rewards, comprising a classification score reward and two intensity distribution rewards, to steer the agents' erasing process precisely, thereby avoiding both under- and over-segmentation. (3) Implementation of a progressive curriculum learning strategy to enable agents to interact with the environment in a progressively challenging manner, thereby enhancing learning efficiency. Extensively validated on the large in-house BUS and ABUS datasets, our Flip Learning method outperforms state-of-the-art WSS methods and foundation models, and achieves comparable performance as fully-supervised learning algorithms.
comment: Accepted by Medical Image Analysis. 24 pages, 13 figures, 18 tabels
☆ Inductive Link Prediction on N-ary Relational Facts via Semantic Hypergraph Reasoning KDD
N-ary relational facts represent semantic correlations among more than two entities. While recent studies have developed link prediction (LP) methods to infer missing relations for knowledge graphs (KGs) containing n-ary relational facts, they are generally limited to transductive settings. Fully inductive settings, where predictions are made on previously unseen entities, remain a significant challenge. As existing methods are mainly entity embedding-based, they struggle to capture entity-independent logical rules. To fill in this gap, we propose an n-ary subgraph reasoning framework for fully inductive link prediction (ILP) on n-ary relational facts. This framework reasons over local subgraphs and has a strong inductive inference ability to capture n-ary patterns. Specifically, we introduce a novel graph structure, the n-ary semantic hypergraph, to facilitate subgraph extraction. Moreover, we develop a subgraph aggregating network, NS-HART, to effectively mine complex semantic correlations within subgraphs. Theoretically, we provide a thorough analysis from the score function optimization perspective to shed light on NS-HART's effectiveness for n-ary ILP tasks. Empirically, we conduct extensive experiments on a series of inductive benchmarks, including transfer reasoning (with and without entity features) and pairwise subgraph reasoning. The results highlight the superiority of the n-ary subgraph reasoning framework and the exceptional inductive ability of NS-HART. The source code of this paper has been made publicly available at https://github.com/yin-gz/Nary-Inductive-SubGraph.
comment: To be published in Proceedings of the 31st ACM SIGKDD Conference on Knowledge Discovery and Data Mining V.1 (KDD'25)
☆ Probabilistic Forecasting for Network Resource Analysis in Integrated Terrestrial and Non-Terrestrial Networks
Efficient resource management is critical for Non-Terrestrial Networks (NTNs) to provide consistent, high-quality service in remote and under-served regions. While traditional single-point prediction methods, such as Long-Short Term Memory (LSTM), have been used in terrestrial networks, they often fall short in NTNs due to the complexity of satellite dynamics, signal latency and coverage variability. Probabilistic forecasting, which quantifies the uncertainties of the predictions, is a robust alternative. In this paper, we evaluate the application of probabilistic forecasting techniques, in particular SFF, to NTN resource allocation scenarios. Our results show their effectiveness in predicting bandwidth and capacity requirements in different NTN segments of probabilistic forecasting compared to single-point prediction techniques such as LSTM. The results show the potential of black probabilistic forecasting models to provide accurate and reliable predictions and to quantify their uncertainty, making them indispensable for optimizing NTN resource allocation. At the end of the paper, we also present application scenarios and a standardization roadmap for the use of probabilistic forecasting in integrated Terrestrial Network (TN)-NTN environments.
☆ AccidentSim: Generating Physically Realistic Vehicle Collision Videos from Real-World Accident Reports
Collecting real-world vehicle accident videos for autonomous driving research is challenging due to their rarity and complexity. While existing driving video generation methods may produce visually realistic videos, they often fail to deliver physically realistic simulations because they lack the capability to generate accurate post-collision trajectories. In this paper, we introduce AccidentSim, a novel framework that generates physically realistic vehicle collision videos by extracting and utilizing the physical clues and contextual information available in real-world vehicle accident reports. Specifically, AccidentSim leverages a reliable physical simulator to replicate post-collision vehicle trajectories from the physical and contextual information in the accident reports and to build a vehicle collision trajectory dataset. This dataset is then used to fine-tune a language model, enabling it to respond to user prompts and predict physically consistent post-collision trajectories across various driving scenarios based on user descriptions. Finally, we employ Neural Radiance Fields (NeRF) to render high-quality backgrounds, merging them with the foreground vehicles that exhibit physically realistic trajectories to generate vehicle collision videos. Experimental results demonstrate that the videos produced by AccidentSim excel in both visual and physical authenticity.
☆ TN-Eval: Rubric and Evaluation Protocols for Measuring the Quality of Behavioral Therapy Notes
Behavioral therapy notes are important for both legal compliance and patient care. Unlike progress notes in physical health, quality standards for behavioral therapy notes remain underdeveloped. To address this gap, we collaborated with licensed therapists to design a comprehensive rubric for evaluating therapy notes across key dimensions: completeness, conciseness, and faithfulness. Further, we extend a public dataset of behavioral health conversations with therapist-written notes and LLM-generated notes, and apply our evaluation framework to measure their quality. We find that: (1) A rubric-based manual evaluation protocol offers more reliable and interpretable results than traditional Likert-scale annotations. (2) LLMs can mimic human evaluators in assessing completeness and conciseness but struggle with faithfulness. (3) Therapist-written notes often lack completeness and conciseness, while LLM-generated notes contain hallucination. Surprisingly, in a blind test, therapists prefer and judge LLM-generated notes to be superior to therapist-written notes.
☆ Procedural Knowledge Ontology (PKO)
Processes, workflows and guidelines are core to ensure the correct functioning of industrial companies: for the successful operations of factory lines, machinery or services, often industry operators rely on their past experience and know-how. The effect is that this Procedural Knowledge (PK) remains tacit and, as such, difficult to exploit efficiently and effectively. This paper presents PKO, the Procedural Knowledge Ontology, which enables the explicit modeling of procedures and their executions, by reusing and extending existing ontologies. PKO is built on requirements collected from three heterogeneous industrial use cases and can be exploited by any AI and data-driven tools that rely on a shared and interoperable representation to support the governance of PK throughout its life cycle. We describe its structure and design methodology, and outline its relevance, quality, and impact by discussing applications leveraging PKO for PK elicitation and exploitation.
☆ $β$-GNN: A Robust Ensemble Approach Against Graph Structure Perturbation
Graph Neural Networks (GNNs) are playing an increasingly important role in the efficient operation and security of computing systems, with applications in workload scheduling, anomaly detection, and resource management. However, their vulnerability to network perturbations poses a significant challenge. We propose $\beta$-GNN, a model enhancing GNN robustness without sacrificing clean data performance. $\beta$-GNN uses a weighted ensemble, combining any GNN with a multi-layer perceptron. A learned dynamic weight, $\beta$, modulates the GNN's contribution. This $\beta$ not only weights GNN influence but also indicates data perturbation levels, enabling proactive mitigation. Experimental results on diverse datasets show $\beta$-GNN's superior adversarial accuracy and attack severity quantification. Crucially, $\beta$-GNN avoids perturbation assumptions, preserving clean data structure and performance.
comment: This is the author's version of the paper accepted at EuroMLSys 2025
☆ Collaborative Storytelling and LLM: A Linguistic Analysis of Automatically-Generated Role-Playing Game Sessions
Role-playing games (RPG) are games in which players interact with one another to create narratives. The role of players in the RPG is largely based on the interaction between players and their characters. This emerging form of shared narrative, primarily oral, is receiving increasing attention. In particular, many authors investigated the use of an LLM as an actor in the game. In this paper, we aim to discover to what extent the language of Large Language Models (LLMs) exhibit oral or written features when asked to generate an RPG session without human interference. We will conduct a linguistic analysis of the lexical and syntactic features of the generated texts and compare the results with analyses of conversations, transcripts of human RPG sessions, and books. We found that LLMs exhibit a pattern that is distinct from all other text categories, including oral conversations, human RPG sessions and books. Our analysis has shown how training influences the way LLMs express themselves and provides important indications of the narrative capabilities of these tools.
comment: 17 pages
☆ State-Aware Perturbation Optimization for Robust Deep Reinforcement Learning
Recently, deep reinforcement learning (DRL) has emerged as a promising approach for robotic control. However, the deployment of DRL in real-world robots is hindered by its sensitivity to environmental perturbations. While existing whitebox adversarial attacks rely on local gradient information and apply uniform perturbations across all states to evaluate DRL robustness, they fail to account for temporal dynamics and state-specific vulnerabilities. To combat the above challenge, we first conduct a theoretical analysis of white-box attacks in DRL by establishing the adversarial victim-dynamics Markov decision process (AVD-MDP), to derive the necessary and sufficient conditions for a successful attack. Based on this, we propose a selective state-aware reinforcement adversarial attack method, named STAR, to optimize perturbation stealthiness and state visitation dispersion. STAR first employs a soft mask-based state-targeting mechanism to minimize redundant perturbations, enhancing stealthiness and attack effectiveness. Then, it incorporates an information-theoretic optimization objective to maximize mutual information between perturbations, environmental states, and victim actions, ensuring a dispersed state-visitation distribution that steers the victim agent into vulnerable states for maximum return reduction. Extensive experiments demonstrate that STAR outperforms state-of-the-art benchmarks.
comment: 15 pages, 11 figures
☆ A decision-theoretic approach to dealing with uncertainty in quantum mechanics
We provide a decision-theoretic framework for dealing with uncertainty in quantum mechanics. This uncertainty is two-fold: on the one hand there may be uncertainty about the state the quantum system is in, and on the other hand, as is essential to quantum mechanical uncertainty, even if the quantum state is known, measurements may still produce an uncertain outcome. In our framework, measurements therefore play the role of acts with an uncertain outcome and our simple decision-theoretic postulates ensure that Born's rule is encapsulated in the utility functions associated with such acts. This approach allows us to uncouple (precise) probability theory from quantum mechanics, in the sense that it leaves room for a more general, so-called imprecise probabilities approach. We discuss the mathematical implications of our findings, which allow us to give a decision-theoretic foundation to recent seminal work by Benavoli, Facchini and Zaffalon, and we compare our approach to earlier and different approaches by Deutsch and Wallace.
comment: 52 pages
☆ StableToolBench-MirrorAPI: Modeling Tool Environments as Mirrors of 7,000+ Real-World APIs
The rapid advancement of large language models (LLMs) has spurred significant interest in tool learning, where LLMs are augmented with external tools to tackle complex tasks. However, existing tool environments face challenges in balancing stability, scalability, and realness, particularly for benchmarking purposes. To address this problem, we propose MirrorAPI, a novel framework that trains specialized LLMs to accurately simulate real API responses, effectively acting as "mirrors" to tool environments. Using a comprehensive dataset of request-response pairs from 7,000+ APIs, we employ supervised fine-tuning and chain-of-thought reasoning to enhance simulation fidelity. MirrorAPI achieves superior accuracy and stability compared to state-of-the-art methods, as demonstrated by its performance on the newly constructed MirrorAPI-Bench and its integration into StableToolBench.
☆ GAIA-2: A Controllable Multi-View Generative World Model for Autonomous Driving
Generative models offer a scalable and flexible paradigm for simulating complex environments, yet current approaches fall short in addressing the domain-specific requirements of autonomous driving - such as multi-agent interactions, fine-grained control, and multi-camera consistency. We introduce GAIA-2, Generative AI for Autonomy, a latent diffusion world model that unifies these capabilities within a single generative framework. GAIA-2 supports controllable video generation conditioned on a rich set of structured inputs: ego-vehicle dynamics, agent configurations, environmental factors, and road semantics. It generates high-resolution, spatiotemporally consistent multi-camera videos across geographically diverse driving environments (UK, US, Germany). The model integrates both structured conditioning and external latent embeddings (e.g., from a proprietary driving model) to facilitate flexible and semantically grounded scene synthesis. Through this integration, GAIA-2 enables scalable simulation of both common and rare driving scenarios, advancing the use of generative world models as a core tool in the development of autonomous systems. Videos are available at https://wayve.ai/thinking/gaia-2.
comment: Technical Report
☆ Design and Evaluation of Neural Network-Based Receiver Architectures for Reliable Communication
Neural network-based receivers leverage deep learning to optimize signal detection and decoding, significantly improving bit-error rate (BER) and block-error rate (BLER) in challenging environments. This study evaluates various architectures and compares their BER and BLER performance across different noise levels. Two novel models, the Dual Attention Transformer (DAT) and the Residual Dual Non-Local Attention Network (RDNLA), integrate self-attention and residual learning to enhance signal reconstruction. These models bypass conventional channel estimation and equalization by directly predicting log-likelihood ratios (LLRs) from received signals, with noise variance as an additional input. Simulations show that DAT and RDNLA outperform traditional and other neural receiver models under varying signal-to-noise ratios (SNR), while their computational efficiency supports their feasibility for next-generation communication systems.
comment: Will be submitted to IEEE Conference
☆ Towards Efficient and General-Purpose Few-Shot Misclassification Detection for Vision-Language Models
Reliable prediction by classifiers is crucial for their deployment in high security and dynamically changing situations. However, modern neural networks often exhibit overconfidence for misclassified predictions, highlighting the need for confidence estimation to detect errors. Despite the achievements obtained by existing methods on small-scale datasets, they all require training from scratch and there are no efficient and effective misclassification detection (MisD) methods, hindering practical application towards large-scale and ever-changing datasets. In this paper, we pave the way to exploit vision language model (VLM) leveraging text information to establish an efficient and general-purpose misclassification detection framework. By harnessing the power of VLM, we construct FSMisD, a Few-Shot prompt learning framework for MisD to refrain from training from scratch and therefore improve tuning efficiency. To enhance misclassification detection ability, we use adaptive pseudo sample generation and a novel negative loss to mitigate the issue of overconfidence by pushing category prompts away from pseudo features. We conduct comprehensive experiments with prompt learning methods and validate the generalization ability across various datasets with domain shift. Significant and consistent improvement demonstrates the effectiveness, efficiency and generalizability of our approach.
comment: preprint
☆ Underwater Image Enhancement by Convolutional Spiking Neural Networks
Underwater image enhancement (UIE) is fundamental for marine applications, including autonomous vision-based navigation. Deep learning methods using convolutional neural networks (CNN) and vision transformers advanced UIE performance. Recently, spiking neural networks (SNN) have gained attention for their lightweight design, energy efficiency, and scalability. This paper introduces UIE-SNN, the first SNN-based UIE algorithm to improve visibility of underwater images. UIE-SNN is a 19- layered convolutional spiking encoder-decoder framework with skip connections, directly trained using surrogate gradient-based backpropagation through time (BPTT) strategy. We explore and validate the influence of training datasets on energy reduction, a unique advantage of UIE-SNN architecture, in contrast to the conventional learning-based architectures, where energy consumption is model-dependent. UIE-SNN optimizes the loss function in latent space representation to reconstruct clear underwater images. Our algorithm performs on par with its non-spiking counterpart methods in terms of PSNR and structural similarity index (SSIM) at reduced timesteps ($T=5$) and energy consumption of $85\%$. The algorithm is trained on two publicly available benchmark datasets, UIEB and EUVP, and tested on unseen images from UIEB, EUVP, LSUI, U45, and our custom UIE dataset. The UIE-SNN algorithm achieves PSNR of \(17.7801~dB\) and SSIM of \(0.7454\) on UIEB, and PSNR of \(23.1725~dB\) and SSIM of \(0.7890\) on EUVP. UIE-SNN achieves this algorithmic performance with fewer operators (\(147.49\) GSOPs) and energy (\(0.1327~J\)) compared to its non-spiking counterpart (GFLOPs = \(218.88\) and Energy=\(1.0068~J\)). Compared with existing SOTA UIE methods, UIE-SNN achieves an average of \(6.5\times\) improvement in energy efficiency. The source code is available at \href{https://github.com/vidya-rejul/UIE-SNN.git}{UIE-SNN}.
☆ Contrastive Learning Guided Latent Diffusion Model for Image-to-Image Translation
The diffusion model has demonstrated superior performance in synthesizing diverse and high-quality images for text-guided image translation. However, there remains room for improvement in both the formulation of text prompts and the preservation of reference image content. First, variations in target text prompts can significantly influence the quality of the generated images, and it is often challenging for users to craft an optimal prompt that fully captures the content of the input image. Second, while existing models can introduce desired modifications to specific regions of the reference image, they frequently induce unintended alterations in areas that should remain unchanged. To address these challenges, we propose pix2pix-zeroCon, a zero-shot diffusion-based method that eliminates the need for additional training by leveraging patch-wise contrastive loss. Specifically, we automatically determine the editing direction in the text embedding space based on the reference image and target prompts. Furthermore, to ensure precise content and structural preservation in the edited image, we introduce cross-attention guiding loss and patch-wise contrastive loss between the generated and original image embeddings within a pre-trained diffusion model. Notably, our approach requires no additional training and operates directly on a pre-trained text-to-image diffusion model. Extensive experiments demonstrate that our method surpasses existing models in image-to-image translation, achieving enhanced fidelity and controllability.
comment: 11 pages, 13 figures
☆ A multi-agentic framework for real-time, autonomous freeform metasurface design
Innovation in nanophotonics currently relies on human experts who synergize specialized knowledge in photonics and coding with simulation and optimization algorithms, entailing design cycles that are time-consuming, computationally demanding, and frequently suboptimal. We introduce MetaChat, a multi-agentic design framework that can translate semantically described photonic design goals into high-performance, freeform device layouts in an automated, nearly real-time manner. Multi-step reasoning is enabled by our Agentic Iterative Monologue (AIM) paradigm, which coherently interfaces agents with code-based tools, other specialized agents, and human designers. Design acceleration is facilitated by Feature-wise Linear Modulation-conditioned Maxwell surrogate solvers that support the generalized evaluation of metasurface structures. We use freeform dielectric metasurfaces as a model system and demonstrate with MetaChat the design of multi-objective, multi-wavelength metasurfaces orders of magnitude faster than conventional methods. These concepts present a scientific computing blueprint for utilizing specialist design agents, surrogate solvers, and human interactions to drive multi-physics innovation and discovery.
comment: 32 pages, 5 figures
☆ From Trial to Triumph: Advancing Long Video Understanding via Visual Context Sample Scaling and Self-reward Alignment
Multi-modal Large language models (MLLMs) show remarkable ability in video understanding. Nevertheless, understanding long videos remains challenging as the models can only process a finite number of frames in a single inference, potentially omitting crucial visual information. To address the challenge, we propose generating multiple predictions through visual context sampling, followed by a scoring mechanism to select the final prediction. Specifically, we devise a bin-wise sampling strategy that enables MLLMs to generate diverse answers based on various combinations of keyframes, thereby enriching the visual context. To determine the final prediction from the sampled answers, we employ a self-reward by linearly combining three scores: (1) a frequency score indicating the prevalence of each option, (2) a marginal confidence score reflecting the inter-intra sample certainty of MLLM predictions, and (3) a reasoning score for different question types, including clue-guided answering for global questions and temporal self-refocusing for local questions. The frequency score ensures robustness through majority correctness, the confidence-aligned score reflects prediction certainty, and the typed-reasoning score addresses cases with sparse key visual information using tailored strategies. Experiments show that this approach covers the correct answer for a high percentage of long video questions, on seven datasets show that our method improves the performance of three MLLMs.
☆ Attention Xception UNet (AXUNet): A Novel Combination of CNN and Self-Attention for Brain Tumor Segmentation
Accurate segmentation of glioma brain tumors is crucial for diagnosis and treatment planning. Deep learning techniques offer promising solutions, but optimal model architectures remain under investigation. We used the BraTS 2021 dataset, selecting T1 with contrast enhancement (T1CE), T2, and Fluid-Attenuated Inversion Recovery (FLAIR) sequences for model development. The proposed Attention Xception UNet (AXUNet) architecture integrates an Xception backbone with dot-product self-attention modules, inspired by state-of-the-art (SOTA) large language models such as Google Bard and OpenAI ChatGPT, within a UNet-shaped model. We compared AXUNet with SOTA models. Comparative evaluation on the test set demonstrated improved results over baseline models. Inception-UNet and Xception-UNet achieved mean Dice scores of 90.88 and 93.24, respectively. Attention ResUNet (AResUNet) attained a mean Dice score of 92.80, with the highest score of 84.92 for enhancing tumor (ET) among all models. Attention Gate UNet (AGUNet) yielded a mean Dice score of 90.38. AXUNet outperformed all models with a mean Dice score of 93.73. It demonstrated superior Dice scores across whole tumor (WT) and tumor core (TC) regions, achieving 92.59 for WT, 86.81 for TC, and 84.89 for ET. The integration of the Xception backbone and dot-product self-attention mechanisms in AXUNet showcases enhanced performance in capturing spatial and contextual information. The findings underscore the potential utility of AXUNet in facilitating precise tumor delineation.
☆ Evaluating Facial Expression Recognition Datasets for Deep Learning: A Benchmark Study with Novel Similarity Metrics
This study investigates the key characteristics and suitability of widely used Facial Expression Recognition (FER) datasets for training deep learning models. In the field of affective computing, FER is essential for interpreting human emotions, yet the performance of FER systems is highly contingent on the quality and diversity of the underlying datasets. To address this issue, we compiled and analyzed 24 FER datasets, including those targeting specific age groups such as children, adults, and the elderly, and processed them through a comprehensive normalization pipeline. In addition, we enriched the datasets with automatic annotations for age and gender, enabling a more nuanced evaluation of their demographic properties. To further assess dataset efficacy, we introduce three novel metricsLocal, Global, and Paired Similarity, which quantitatively measure dataset difficulty, generalization capability, and cross-dataset transferability. Benchmark experiments using state-of-the-art neural networks reveal that large-scale, automatically collected datasets (e.g., AffectNet, FER2013) tend to generalize better, despite issues with labeling noise and demographic biases, whereas controlled datasets offer higher annotation quality but limited variability. Our findings provide actionable recommendations for dataset selection and design, advancing the development of more robust, fair, and effective FER systems.
☆ Perspective-Shifted Neuro-Symbolic World Models: A Framework for Socially-Aware Robot Navigation
Navigating in environments alongside humans requires agents to reason under uncertainty and account for the beliefs and intentions of those around them. Under a sequential decision-making framework, egocentric navigation can naturally be represented as a Markov Decision Process (MDP). However, social navigation additionally requires reasoning about the hidden beliefs of others, inherently leading to a Partially Observable Markov Decision Process (POMDP), where agents lack direct access to others' mental states. Inspired by Theory of Mind and Epistemic Planning, we propose (1) a neuro-symbolic model-based reinforcement learning architecture for social navigation, addressing the challenge of belief tracking in partially observable environments; and (2) a perspective-shift operator for belief estimation, leveraging recent work on Influence-based Abstractions (IBA) in structured multi-agent settings.
☆ Including local feature interactions in deep non-negative matrix factorization networks improves performance
The brain uses positive signals as a means of signaling. Forward interactions in the early visual cortex are also positive, realized by excitatory synapses. Only local interactions also include inhibition. Non-negative matrix factorization (NMF) captures the biological constraint of positive long-range interactions and can be implemented with stochastic spikes. While NMF can serve as an abstract formalization of early neural processing in the visual system, the performance of deep convolutional networks with NMF modules does not match that of CNNs of similar size. However, when the local NMF modules are each followed by a module that mixes the NMF's positive activities, the performances on the benchmark data exceed that of vanilla deep convolutional networks of similar size. This setting can be considered a biologically more plausible emulation of the processing in cortical (hyper-)columns with the potential to improve the performance of deep networks.
☆ FastFT: Accelerating Reinforced Feature Transformation via Advanced Exploration Strategies ICDE 2025
Feature Transformation is crucial for classic machine learning that aims to generate feature combinations to enhance the performance of downstream tasks from a data-centric perspective. Current methodologies, such as manual expert-driven processes, iterative-feedback techniques, and exploration-generative tactics, have shown promise in automating such data engineering workflow by minimizing human involvement. However, three challenges remain in those frameworks: (1) It predominantly depends on downstream task performance metrics, as assessment is time-consuming, especially for large datasets. (2) The diversity of feature combinations will hardly be guaranteed after random exploration ends. (3) Rare significant transformations lead to sparse valuable feedback that hinders the learning processes or leads to less effective results. In response to these challenges, we introduce FastFT, an innovative framework that leverages a trio of advanced strategies.We first decouple the feature transformation evaluation from the outcomes of the generated datasets via the performance predictor. To address the issue of reward sparsity, we developed a method to evaluate the novelty of generated transformation sequences. Incorporating this novelty into the reward function accelerates the model's exploration of effective transformations, thereby improving the search productivity. Additionally, we combine novelty and performance to create a prioritized memory buffer, ensuring that essential experiences are effectively revisited during exploration. Our extensive experimental evaluations validate the performance, efficiency, and traceability of our proposed framework, showcasing its superiority in handling complex feature transformation tasks.
comment: 14 pages, Accepted by ICDE 2025
☆ MoLe-VLA: Dynamic Layer-skipping Vision Language Action Model via Mixture-of-Layers for Efficient Robot Manipulation
Multimodal Large Language Models (MLLMs) excel in understanding complex language and visual data, enabling generalist robotic systems to interpret instructions and perform embodied tasks. Nevertheless, their real-world deployment is hindered by substantial computational and storage demands. Recent insights into the homogeneous patterns in the LLM layer have inspired sparsification techniques to address these challenges, such as early exit and token pruning. However, these methods often neglect the critical role of the final layers that encode the semantic information most relevant to downstream robotic tasks. Aligning with the recent breakthrough of the Shallow Brain Hypothesis (SBH) in neuroscience and the mixture of experts in model sparsification, we conceptualize each LLM layer as an expert and propose a Mixture-of-Layers Vision-Language-Action model (MoLe-VLA, or simply MoLe) architecture for dynamic LLM layer activation. We introduce a Spatial-Temporal Aware Router (STAR) for MoLe to selectively activate only parts of the layers based on the robot's current state, mimicking the brain's distinct signal pathways specialized for cognition and causal reasoning. Additionally, to compensate for the cognitive ability of LLMs lost in MoLe, we devise a Cognition Self-Knowledge Distillation (CogKD) framework. CogKD enhances the understanding of task demands and improves the generation of task-relevant action sequences by leveraging cognitive features. Extensive experiments conducted in both RLBench simulation and real-world environments demonstrate the superiority of MoLe-VLA in both efficiency and performance. Specifically, MoLe-VLA achieves an 8% improvement in the mean success rate across ten tasks while reducing computational costs by up to x5.6 compared to standard LLMs.
☆ VideoGEM: Training-free Action Grounding in Videos
Vision-language foundation models have shown impressive capabilities across various zero-shot tasks, including training-free localization and grounding, primarily focusing on localizing objects in images. However, leveraging those capabilities to localize actions and events in videos is challenging, as actions have less physical outline and are usually described by higher-level concepts. In this work, we propose VideoGEM, the first training-free spatial action grounding method based on pretrained image- and video-language backbones. Namely, we adapt the self-self attention formulation of GEM to spatial activity grounding. We observe that high-level semantic concepts, such as actions, usually emerge in the higher layers of the image- and video-language models. We, therefore, propose a layer weighting in the self-attention path to prioritize higher layers. Additionally, we introduce a dynamic weighting method to automatically tune layer weights to capture each layer`s relevance to a specific prompt. Finally, we introduce a prompt decomposition, processing action, verb, and object prompts separately, resulting in a better spatial localization of actions. We evaluate the proposed approach on three image- and video-language backbones, CLIP, OpenCLIP, and ViCLIP, and on four video grounding datasets, V-HICO, DALY, YouCook-Interactions, and GroundingYouTube, showing that the proposed training-free approach is able to outperform current trained state-of-the-art approaches for spatial video grounding.
☆ Wasserstein Distributionally Robust Bayesian Optimization with Continuous Context
We address the challenge of sequential data-driven decision-making under context distributional uncertainty. This problem arises in numerous real-world scenarios where the learner optimizes black-box objective functions in the presence of uncontrollable contextual variables. We consider the setting where the context distribution is uncertain but known to lie within an ambiguity set defined as a ball in the Wasserstein distance. We propose a novel algorithm for Wasserstein Distributionally Robust Bayesian Optimization that can handle continuous context distributions while maintaining computational tractability. Our theoretical analysis combines recent results in self-normalized concentration in Hilbert spaces and finite-sample bounds for distributionally robust optimization to establish sublinear regret bounds that match state-of-the-art results. Through extensive comparisons with existing approaches on both synthetic and real-world problems, we demonstrate the simplicity, effectiveness, and practical applicability of our proposed method.
☆ Iterative Prompting with Persuasion Skills in Jailbreaking Large Language Models
Large language models (LLMs) are designed to align with human values in their responses. This study exploits LLMs with an iterative prompting technique where each prompt is systematically modified and refined across multiple iterations to enhance its effectiveness in jailbreaking attacks progressively. This technique involves analyzing the response patterns of LLMs, including GPT-3.5, GPT-4, LLaMa2, Vicuna, and ChatGLM, allowing us to adjust and optimize prompts to evade the LLMs' ethical and security constraints. Persuasion strategies enhance prompt effectiveness while maintaining consistency with malicious intent. Our results show that the attack success rates (ASR) increase as the attacking prompts become more refined with the highest ASR of 90% for GPT4 and ChatGLM and the lowest ASR of 68% for LLaMa2. Our technique outperforms baseline techniques (PAIR and PAP) in ASR and shows comparable performance with GCG and ArtPrompt.
☆ A Multilingual, Culture-First Approach to Addressing Misgendering in LLM Applications
Misgendering is the act of referring to someone by a gender that does not match their chosen identity. It marginalizes and undermines a person's sense of self, causing significant harm. English-based approaches have clear-cut approaches to avoiding misgendering, such as the use of the pronoun ``they''. However, other languages pose unique challenges due to both grammatical and cultural constructs. In this work we develop methodologies to assess and mitigate misgendering across 42 languages and dialects using a participatory-design approach to design effective and appropriate guardrails across all languages. We test these guardrails in a standard large language model-based application (meeting transcript summarization), where both the data generation and the annotation steps followed a human-in-the-loop approach. We find that the proposed guardrails are very effective in reducing misgendering rates across all languages in the summaries generated, and without incurring loss of quality. Our human-in-the-loop approach demonstrates a method to feasibly scale inclusive and responsible AI-based solutions across multiple languages and cultures.
☆ Context-Aware Weakly Supervised Image Manipulation Localization with SAM Refinement
Malicious image manipulation poses societal risks, increasing the importance of effective image manipulation detection methods. Recent approaches in image manipulation detection have largely been driven by fully supervised approaches, which require labor-intensive pixel-level annotations. Thus, it is essential to explore weakly supervised image manipulation localization methods that only require image-level binary labels for training. However, existing weakly supervised image manipulation methods overlook the importance of edge information for accurate localization, leading to suboptimal localization performance. To address this, we propose a Context-Aware Boundary Localization (CABL) module to aggregate boundary features and learn context-inconsistency for localizing manipulated areas. Furthermore, by leveraging Class Activation Mapping (CAM) and Segment Anything Model (SAM), we introduce the CAM-Guided SAM Refinement (CGSR) module to generate more accurate manipulation localization maps. By integrating two modules, we present a novel weakly supervised framework based on a dual-branch Transformer-CNN architecture. Our method achieves outstanding localization performance across multiple datasets.
☆ CryoSAMU: Enhancing 3D Cryo-EM Density Maps of Protein Structures at Intermediate Resolution with Structure-Aware Multimodal U-Nets
Enhancing cryogenic electron microscopy (cryo-EM) 3D density maps at intermediate resolution (4-8 {\AA}) is crucial in protein structure determination. Recent advances in deep learning have led to the development of automated approaches for enhancing experimental cryo-EM density maps. Yet, these methods are not optimized for intermediate-resolution maps and rely on map density features alone. To address this, we propose CryoSAMU, a novel method designed to enhance 3D cryo-EM density maps of protein structures using structure-aware multimodal U-Nets and trained on curated intermediate-resolution density maps. We comprehensively evaluate CryoSAMU across various metrics and demonstrate its competitive performance compared to state-of-the-art methods. Notably, CryoSAMU achieves significantly faster processing speed, showing promise for future practical applications. Our code is available at https://github.com/chenwei-zhang/CryoSAMU.
comment: 18 pages, 6 main figures, 2 supplementary figures, 3 main tables, 4 supplementary tables
☆ QualiSpeech: A Speech Quality Assessment Dataset with Natural Language Reasoning and Descriptions
This paper explores a novel perspective to speech quality assessment by leveraging natural language descriptions, offering richer, more nuanced insights than traditional numerical scoring methods. Natural language feedback provides instructive recommendations and detailed evaluations, yet existing datasets lack the comprehensive annotations needed for this approach. To bridge this gap, we introduce QualiSpeech, a comprehensive low-level speech quality assessment dataset encompassing 11 key aspects and detailed natural language comments that include reasoning and contextual insights. Additionally, we propose the QualiSpeech Benchmark to evaluate the low-level speech understanding capabilities of auditory large language models (LLMs). Experimental results demonstrate that finetuned auditory LLMs can reliably generate detailed descriptions of noise and distortion, effectively identifying their types and temporal characteristics. The results further highlight the potential for incorporating reasoning to enhance the accuracy and reliability of quality assessments. The dataset will be released at https://huggingface.co/datasets/tsinghua-ee/QualiSpeech.
comment: 23 pages, 16 figures
☆ Model-Based Offline Reinforcement Learning with Adversarial Data Augmentation
Model-based offline Reinforcement Learning (RL) constructs environment models from offline datasets to perform conservative policy optimization. Existing approaches focus on learning state transitions through ensemble models, rollouting conservative estimation to mitigate extrapolation errors. However, the static data makes it challenging to develop a robust policy, and offline agents cannot access the environment to gather new data. To address these challenges, we introduce Model-based Offline Reinforcement learning with AdversariaL data augmentation (MORAL). In MORAL, we replace the fixed horizon rollout by employing adversaria data augmentation to execute alternating sampling with ensemble models to enrich training data. Specifically, this adversarial process dynamically selects ensemble models against policy for biased sampling, mitigating the optimistic estimation of fixed models, thus robustly expanding the training data for policy optimization. Moreover, a differential factor is integrated into the adversarial process for regularization, ensuring error minimization in extrapolations. This data-augmented optimization adapts to diverse offline tasks without rollout horizon tuning, showing remarkable applicability. Extensive experiments on D4RL benchmark demonstrate that MORAL outperforms other model-based offline RL methods in terms of policy learning and sample efficiency.
☆ Faster Parameter-Efficient Tuning with Token Redundancy Reduction CVPR 2025
Parameter-efficient tuning (PET) aims to transfer pre-trained foundation models to downstream tasks by learning a small number of parameters. Compared to traditional fine-tuning, which updates the entire model, PET significantly reduces storage and transfer costs for each task regardless of exponentially increasing pre-trained model capacity. However, most PET methods inherit the inference latency of their large backbone models and often introduce additional computational overhead due to additional modules (e.g. adapters), limiting their practicality for compute-intensive applications. In this paper, we propose Faster Parameter-Efficient Tuning (FPET), a novel approach that enhances inference speed and training efficiency while maintaining high storage efficiency. Specifically, we introduce a plug-and-play token redundancy reduction module delicately designed for PET. This module refines tokens from the self-attention layer using an adapter to learn the accurate similarity between tokens and cuts off the tokens through a fully-differentiable token merging strategy, which uses a straight-through estimator for optimal token reduction. Experimental results prove that our FPET achieves faster inference and higher memory efficiency than the pre-trained backbone while keeping competitive performance on par with state-of-the-art PET methods.
comment: CVPR 2025 Camera-ready
☆ Are We There Yet? Unraveling the State-of-the-Art Graph Network Intrusion Detection Systems
Network Intrusion Detection Systems (NIDS) are vital for ensuring enterprise security. Recently, Graph-based NIDS (GIDS) have attracted considerable attention because of their capability to effectively capture the complex relationships within the graph structures of data communications. Despite their promise, the reproducibility and replicability of these GIDS remain largely unexplored, posing challenges for developing reliable and robust detection systems. This study bridges this gap by designing a systematic approach to evaluate state-of-the-art GIDS, which includes critically assessing, extending, and clarifying the findings of these systems. We further assess the robustness of GIDS under adversarial attacks. Evaluations were conducted on three public datasets as well as a newly collected large-scale enterprise dataset. Our findings reveal significant performance discrepancies, highlighting challenges related to dataset scale, model inputs, and implementation settings. We demonstrate difficulties in reproducing and replicating results, particularly concerning false positive rates and robustness against adversarial attacks. This work provides valuable insights and recommendations for future research, emphasizing the importance of rigorous reproduction and replication studies in developing robust and generalizable GIDS solutions.
☆ sudo rm -rf agentic_security
Large Language Models (LLMs) are increasingly deployed as computer-use agents, autonomously performing tasks within real desktop or web environments. While this evolution greatly expands practical use cases for humans, it also creates serious security exposures. We present SUDO (Screen-based Universal Detox2Tox Offense), a novel attack framework that systematically bypasses refusal trained safeguards in commercial computer-use agents, such as Claude Computer Use. The core mechanism, Detox2Tox, transforms harmful requests (that agents initially reject) into seemingly benign requests via detoxification, secures detailed instructions from advanced vision language models (VLMs), and then reintroduces malicious content via toxification just before execution. Unlike conventional jailbreaks, SUDO iteratively refines its attacks based on a built-in refusal feedback, making it increasingly effective against robust policy filters. In extensive tests spanning 50 real-world tasks and multiple state-of-the-art VLMs, SUDO achieves a stark attack success rate of 24% (with no refinement), and up to 41% (by its iterative refinement) in Claude Computer Use. By revealing these vulnerabilities and demonstrating the ease with which they can be exploited in real-world computing environments, this paper highlights an immediate need for robust, context-aware safeguards. WARNING: This paper includes harmful or offensive model outputs.
☆ LogicQA: Logical Anomaly Detection with Vision Language Model Generated Questions
Anomaly Detection (AD) focuses on detecting samples that differ from the standard pattern, making it a vital tool in process control. Logical anomalies may appear visually normal yet violate predefined constraints on object presence, arrangement, or quantity, depending on reasoning and explainability. We introduce LogicQA, a framework that enhances AD by providing industrial operators with explanations for logical anomalies. LogicQA compiles automatically generated questions into a checklist and collects responses to identify violations of logical constraints. LogicQA is training-free, annotation-free, and operates in a few-shot setting. We achieve state-of-the-art (SOTA) Logical AD performance on public benchmarks, MVTec LOCO AD, with an AUROC of 87.6 percent and an F1-max of 87.0 percent along with the explanations of anomalies. Also, our approach has shown outstanding performance on semiconductor SEM corporate data, further validating its effectiveness in industrial applications.
☆ ESSR: An 8K@30FPS Super-Resolution Accelerator With Edge Selective Network
Deep learning-based super-resolution (SR) is challenging to implement in resource-constrained edge devices for resolutions beyond full HD due to its high computational complexity and memory bandwidth requirements. This paper introduces an 8K@30FPS SR accelerator with edge-selective dynamic input processing. Dynamic processing chooses the appropriate subnets for different patches based on simple input edge criteria, achieving a 50\% MAC reduction with only a 0.1dB PSNR decrease. The quality of reconstruction images is guaranteed and maximized its potential with \textit{resource adaptive model switching} even under resource constraints. In conjunction with hardware-specific refinements, the model size is reduced by 84\% to 51K, but with a decrease of less than 0.6dB PSNR. Additionally, to support dynamic processing with high utilization, this design incorporates a \textit{configurable group of layer mapping} that synergizes with the \textit{structure-friendly fusion block}, resulting in 77\% hardware utilization and up to 79\% reduction in feature SRAM access. The implementation, using the TSMC 28nm process, can achieve 8K@30FPS throughput at 800MHz with a gate count of 2749K, 0.2075W power consumption, and 4797Mpixels/J energy efficiency, exceeding previous work.
☆ LGR: LLM-Guided Ranking of Frontiers for Object Goal Navigation
Object Goal Navigation (OGN) is a fundamental task for robots and AI, with key applications such as mobile robot image databases (MRID). In particular, mapless OGN is essential in scenarios involving unknown or dynamic environments. This study aims to enhance recent modular mapless OGN systems by leveraging the commonsense reasoning capabilities of large language models (LLMs). Specifically, we address the challenge of determining the visiting order in frontier-based exploration by framing it as a frontier ranking problem. Our approach is grounded in recent findings that, while LLMs cannot determine the absolute value of a frontier, they excel at evaluating the relative value between multiple frontiers viewed within a single image using the view image as context. We dynamically manage the frontier list by adding and removing elements, using an LLM as a ranking model. The ranking results are represented as reciprocal rank vectors, which are ideal for multi-view, multi-query information fusion. We validate the effectiveness of our method through evaluations in Habitat-Sim.
comment: 10 pages, 11 figures, technical report
☆ Dynamic Learning and Productivity for Data Analysts: A Bayesian Hidden Markov Model Perspective
Data analysts are essential in organizations, transforming raw data into insights that drive decision-making and strategy. This study explores how analysts' productivity evolves on a collaborative platform, focusing on two key learning activities: writing queries and viewing peer queries. While traditional research often assumes static models, where performance improves steadily with cumulative learning, such models fail to capture the dynamic nature of real-world learning. To address this, we propose a Hidden Markov Model (HMM) that tracks how analysts transition between distinct learning states based on their participation in these activities. Using an industry dataset with 2,001 analysts and 79,797 queries, this study identifies three learning states: novice, intermediate, and advanced. Productivity increases as analysts advance to higher states, reflecting the cumulative benefits of learning. Writing queries benefits analysts across all states, with the largest gains observed for novices. Viewing peer queries supports novices but may hinder analysts in higher states due to cognitive overload or inefficiencies. Transitions between states are also uneven, with progression from intermediate to advanced being particularly challenging. This study advances understanding of into dynamic learning behavior of knowledge worker and offers practical implications for designing systems, optimizing training, enabling personalized learning, and fostering effective knowledge sharing.
comment: 29 pages; a shorter 11-page version is accepted by HCI International (HCII) 2025;
☆ Dynamics of Algorithmic Content Amplification on TikTok
Intelligent algorithms increasingly shape the content we encounter and engage with online. TikTok's For You feed exemplifies extreme algorithm-driven curation, tailoring the stream of video content almost exclusively based on users' explicit and implicit interactions with the platform. Despite growing attention, the dynamics of content amplification on TikTok remain largely unquantified. How quickly, and to what extent, does TikTok's algorithm amplify content aligned with users' interests? To address these questions, we conduct a sock-puppet audit, deploying bots with different interests to engage with TikTok's "For You" feed. Our findings reveal that content aligned with the bots' interests undergoes strong amplification, with rapid reinforcement typically occurring within the first 200 videos watched. While amplification is consistently observed across all interests, its intensity varies by interest, indicating the emergence of topic-specific biases. Time series analyses and Markov models uncover distinct phases of recommendation dynamics, including persistent content reinforcement and a gradual decline in content diversity over time. Although TikTok's algorithm preserves some content diversity, we find a strong negative correlation between amplification and exploration: as the amplification of interest-aligned content increases, engagement with unseen hashtags declines. These findings contribute to discussions on socio-algorithmic feedback loops in the digital age and the trade-offs between personalization and content diversity.
comment: 34 pages
☆ TraNCE: Transformative Non-linear Concept Explainer for CNNs
Convolutional neural networks (CNNs) have succeeded remarkably in various computer vision tasks. However, they are not intrinsically explainable. While the feature-level understanding of CNNs reveals where the models looked, concept-based explainability methods provide insights into what the models saw. However, their assumption of linear reconstructability of image activations fails to capture the intricate relationships within these activations. Their Fidelity-only approach to evaluating global explanations also presents a new concern. For the first time, we address these limitations with the novel Transformative Nonlinear Concept Explainer (TraNCE) for CNNs. Unlike linear reconstruction assumptions made by existing methods, TraNCE captures the intricate relationships within the activations. This study presents three original contributions to the CNN explainability literature: (i) An automatic concept discovery mechanism based on variational autoencoders (VAEs). This transformative concept discovery process enhances the identification of meaningful concepts from image activations. (ii) A visualization module that leverages the Bessel function to create a smooth transition between prototypical image pixels, revealing not only what the CNN saw but also what the CNN avoided, thereby mitigating the challenges of concept duplication as documented in previous works. (iii) A new metric, the Faith score, integrates both Coherence and Fidelity for a comprehensive evaluation of explainer faithfulness and consistency.
☆ Advancements in Natural Language Processing: Exploring Transformer-Based Architectures for Text Understanding
Natural Language Processing (NLP) has witnessed a transformative leap with the advent of transformer-based architectures, which have significantly enhanced the ability of machines to understand and generate human-like text. This paper explores the advancements in transformer models, such as BERT and GPT, focusing on their superior performance in text understanding tasks compared to traditional methods like recurrent neural networks (RNNs). By analyzing statistical properties through visual representations-including probability density functions of text length distributions and feature space classifications-the study highlights the models' proficiency in handling long-range dependencies, adapting to conditional shifts, and extracting features for classification, even with overlapping classes. Drawing on recent 2024 research, including enhancements in multi-hop knowledge graph reasoning and context-aware chat interactions, the paper outlines a methodology involving data preparation, model selection, pretraining, fine-tuning, and evaluation. The results demonstrate state-of-the-art performance on benchmarks like GLUE and SQuAD, with F1 scores exceeding 90%, though challenges such as high computational costs persist. This work underscores the pivotal role of transformers in modern NLP and suggests future directions, including efficiency optimization and multimodal integration, to further advance language-based AI systems.
comment: This paper has been accepted by the 5th International Conference on Artificial Intelligence and Industrial Technology Applications (AIITA 2025)
☆ Learning Adaptive Dexterous Grasping from Single Demonstrations
How can robots learn dexterous grasping skills efficiently and apply them adaptively based on user instructions? This work tackles two key challenges: efficient skill acquisition from limited human demonstrations and context-driven skill selection. We introduce AdaDexGrasp, a framework that learns a library of grasping skills from a single human demonstration per skill and selects the most suitable one using a vision-language model (VLM). To improve sample efficiency, we propose a trajectory following reward that guides reinforcement learning (RL) toward states close to a human demonstration while allowing flexibility in exploration. To learn beyond the single demonstration, we employ curriculum learning, progressively increasing object pose variations to enhance robustness. At deployment, a VLM retrieves the appropriate skill based on user instructions, bridging low-level learned skills with high-level intent. We evaluate AdaDexGrasp in both simulation and real-world settings, showing that our approach significantly improves RL efficiency and enables learning human-like grasp strategies across varied object configurations. Finally, we demonstrate zero-shot transfer of our learned policies to a real-world PSYONIC Ability Hand, with a 90% success rate across objects, significantly outperforming the baseline.
☆ Generalized Phase Pressure Control Enhanced Reinforcement Learning for Traffic Signal Control
Appropriate traffic state representation is crucial for learning traffic signal control policies. However, most of the current traffic state representations are heuristically designed, with insufficient theoretical support. In this paper, we (1) develop a flexible, efficient, and theoretically grounded method, namely generalized phase pressure (G2P) control, which takes only simple lane features into consideration to decide which phase to be actuated; 2) extend the pressure control theory to a general form for multi-homogeneous-lane road networks based on queueing theory; (3) design a new traffic state representation based on the generalized phase state features from G2P control; and 4) develop a reinforcement learning (RL)-based algorithm template named G2P-XLight, and two RL algorithms, G2P-MPLight and G2P-CoLight, by combining the generalized phase state representation with MPLight and CoLight, two well-performed RL methods for learning traffic signal control policies. Extensive experiments conducted on multiple real-world datasets demonstrate that G2P control outperforms the state-of-the-art (SOTA) heuristic method in the transportation field and other recent human-designed heuristic methods; and that the newly proposed G2P-XLight significantly outperforms SOTA learning-based approaches. Our code is available online.
☆ SARGes: Semantically Aligned Reliable Gesture Generation via Intent Chain
Co-speech gesture generation enhances human-computer interaction realism through speech-synchronized gesture synthesis. However, generating semantically meaningful gestures remains a challenging problem. We propose SARGes, a novel framework that leverages large language models (LLMs) to parse speech content and generate reliable semantic gesture labels, which subsequently guide the synthesis of meaningful co-speech gestures.First, we constructed a comprehensive co-speech gesture ethogram and developed an LLM-based intent chain reasoning mechanism that systematically parses and decomposes gesture semantics into structured inference steps following ethogram criteria, effectively guiding LLMs to generate context-aware gesture labels. Subsequently, we constructed an intent chain-annotated text-to-gesture label dataset and trained a lightweight gesture label generation model, which then guides the generation of credible and semantically coherent co-speech gestures. Experimental results demonstrate that SARGes achieves highly semantically-aligned gesture labeling (50.2% accuracy) with efficient single-pass inference (0.4 seconds). The proposed method provides an interpretable intent reasoning pathway for semantic gesture synthesis.
Assessing SAM for Tree Crown Instance Segmentation from Drone Imagery ICLR 2025
The potential of tree planting as a natural climate solution is often undermined by inadequate monitoring of tree planting projects. Current monitoring methods involve measuring trees by hand for each species, requiring extensive cost, time, and labour. Advances in drone remote sensing and computer vision offer great potential for mapping and characterizing trees from aerial imagery, and large pre-trained vision models, such as the Segment Anything Model (SAM), may be a particularly compelling choice given limited labeled data. In this work, we compare SAM methods for the task of automatic tree crown instance segmentation in high resolution drone imagery of young tree plantations. We explore the potential of SAM for this task, and find that methods using SAM out-of-the-box do not outperform a custom Mask R-CNN, even with well-designed prompts, but that there is potential for methods which tune SAM further. We also show that predictions can be improved by adding Digital Surface Model (DSM) information as an input.
comment: ICLR 2025 ML4RS workshop
☆ Leveraging Implicit Sentiments: Enhancing Reliability and Validity in Psychological Trait Evaluation of LLMs
Recent advancements in Large Language Models (LLMs) have led to their increasing integration into human life. With the transition from mere tools to human-like assistants, understanding their psychological aspects-such as emotional tendencies and personalities-becomes essential for ensuring their trustworthiness. However, current psychological evaluations of LLMs, often based on human psychological assessments like the BFI, face significant limitations. The results from these approaches often lack reliability and have limited validity when predicting LLM behavior in real-world scenarios. In this work, we introduce a novel evaluation instrument specifically designed for LLMs, called Core Sentiment Inventory (CSI). CSI is a bilingual tool, covering both English and Chinese, that implicitly evaluates models' sentiment tendencies, providing an insightful psychological portrait of LLM across three dimensions: optimism, pessimism, and neutrality. Through extensive experiments, we demonstrate that: 1) CSI effectively captures nuanced emotional patterns, revealing significant variation in LLMs across languages and contexts; 2) Compared to current approaches, CSI significantly improves reliability, yielding more consistent results; and 3) The correlation between CSI scores and the sentiment of LLM's real-world outputs exceeds 0.85, demonstrating its strong validity in predicting LLM behavior. We make CSI public available via: https://github.com/dependentsign/CSI.
comment: Code available via https://github.com/dependentsign/CSI
☆ Offline Reinforcement Learning with Discrete Diffusion Skills
Skills have been introduced to offline reinforcement learning (RL) as temporal abstractions to tackle complex, long-horizon tasks, promoting consistent behavior and enabling meaningful exploration. While skills in offline RL are predominantly modeled within a continuous latent space, the potential of discrete skill spaces remains largely underexplored. In this paper, we propose a compact discrete skill space for offline RL tasks supported by state-of-the-art transformer-based encoder and diffusion-based decoder. Coupled with a high-level policy trained via offline RL techniques, our method establishes a hierarchical RL framework where the trained diffusion decoder plays a pivotal role. Empirical evaluations show that the proposed algorithm, Discrete Diffusion Skill (DDS), is a powerful offline RL method. DDS performs competitively on Locomotion and Kitchen tasks and excels on long-horizon tasks, achieving at least a 12 percent improvement on AntMaze-v2 benchmarks compared to existing offline RL approaches. Furthermore, DDS offers improved interpretability, training stability, and online exploration compared to previous skill-based methods.
☆ Look Before Leap: Look-Ahead Planning with Uncertainty in Reinforcement Learning
Model-based reinforcement learning (MBRL) has demonstrated superior sample efficiency compared to model-free reinforcement learning (MFRL). However, the presence of inaccurate models can introduce biases during policy learning, resulting in misleading trajectories. The challenge lies in obtaining accurate models due to limited diverse training data, particularly in regions with limited visits (uncertain regions). Existing approaches passively quantify uncertainty after sample generation, failing to actively collect uncertain samples that could enhance state coverage and improve model accuracy. Moreover, MBRL often faces difficulties in making accurate multi-step predictions, thereby impacting overall performance. To address these limitations, we propose a novel framework for uncertainty-aware policy optimization with model-based exploratory planning. In the model-based planning phase, we introduce an uncertainty-aware k-step lookahead planning approach to guide action selection at each step. This process involves a trade-off analysis between model uncertainty and value function approximation error, effectively enhancing policy performance. In the policy optimization phase, we leverage an uncertainty-driven exploratory policy to actively collect diverse training samples, resulting in improved model accuracy and overall performance of the RL agent. Our approach offers flexibility and applicability to tasks with varying state/action spaces and reward structures. We validate its effectiveness through experiments on challenging robotic manipulation tasks and Atari games, surpassing state-of-the-art methods with fewer interactions, thereby leading to significant performance improvements.
☆ Unlocking the Value of Decentralized Data: A Federated Dual Learning Approach for Model Aggregation
Artificial Intelligence (AI) technologies have revolutionized numerous fields, yet their applications often rely on costly and time-consuming data collection processes. Federated Learning (FL) offers a promising alternative by enabling AI models to be trained on decentralized data where data is scattered across clients (distributed nodes). However, existing FL approaches struggle to match the performance of centralized training due to challenges such as heterogeneous data distribution and communication delays, limiting their potential for breakthroughs. We observe that many real-world use cases involve hybrid data regimes, in which a server (center node) has access to some data while a large amount of data is distributed across associated clients. To improve the utilization of decentralized data under this regime, address data heterogeneity issue, and facilitate asynchronous communication between the server and clients, we propose a dual learning approach that leverages centralized data at the server to guide the merging of model updates from clients. Our method accommodates scenarios where server data is out-of-domain relative to decentralized client data, making it applicable to a wide range of use cases. We provide theoretical analysis demonstrating the faster convergence of our method compared to existing methods. Furthermore, experimental results across various scenarios show that our approach significantly outperforms existing technologies, highlighting its potential to unlock the value of large amounts of decentralized data.
☆ Can We Make Code Green? Understanding Trade-Offs in LLMs vs. Human Code Optimizations
The rapid technological evolution has accelerated software development for various domains and use cases, contributing to a growing share of global carbon emissions. While recent large language models (LLMs) claim to assist developers in optimizing code for performance and energy efficiency, their efficacy in real-world scenarios remains under exploration. In this work, we explore the effectiveness of LLMs in reducing the environmental footprint of real-world projects, focusing on software written in Matlab-widely used in both academia and industry for scientific and engineering applications. We analyze energy-focused optimization on 400 scripts across 100 top GitHub repositories. We examine potential 2,176 optimizations recommended by leading LLMs, such as GPT-3, GPT-4, Llama, and Mixtral, and a senior Matlab developer, on energy consumption, memory usage, execution time consumption, and code correctness. The developer serves as a real-world baseline for comparing typical human and LLM-generated optimizations. Mapping these optimizations to 13 high-level themes, we found that LLMs propose a broad spectrum of improvements--beyond energy efficiency--including improving code readability and maintainability, memory management, error handling while the developer overlooked some parallel processing, error handling etc. However, our statistical tests reveal that the energy-focused optimizations unexpectedly negatively impacted memory usage, with no clear benefits regarding execution time or energy consumption. Our qualitative analysis of energy-time trade-offs revealed that some themes, such as vectorization preallocation, were among the common themes shaping these trade-offs. With LLMs becoming ubiquitous in modern software development, our study serves as a call to action: prioritizing the evaluation of common coding practices to identify the green ones.
☆ Synthesizing world models for bilevel planning
Modern reinforcement learning (RL) systems have demonstrated remarkable capabilities in complex environments, such as video games. However, they still fall short of achieving human-like sample efficiency and adaptability when learning new domains. Theory-based reinforcement learning (TBRL) is an algorithmic framework specifically designed to address this gap. Modeled on cognitive theories, TBRL leverages structured, causal world models - "theories" - as forward simulators for use in planning, generalization and exploration. Although current TBRL systems provide compelling explanations of how humans learn to play video games, they face several technical limitations: their theory languages are restrictive, and their planning algorithms are not scalable. To address these challenges, we introduce TheoryCoder, an instantiation of TBRL that exploits hierarchical representations of theories and efficient program synthesis methods for more powerful learning and planning. TheoryCoder equips agents with general-purpose abstractions (e.g., "move to"), which are then grounded in a particular environment by learning a low-level transition model (a Python program synthesized from observations by a large language model). A bilevel planning algorithm can exploit this hierarchical structure to solve large domains. We demonstrate that this approach can be successfully applied to diverse and challenging grid-world games, where approaches based on directly synthesizing a policy perform poorly. Ablation studies demonstrate the benefits of using hierarchical abstractions.
comment: 25 pages
☆ The Art of Tool Interface Design
We present an agentic framework, Thinker, which achieves state of art performance in challenging reasoning tasks for realistic customer service scenarios that involve complex business logic and human interactions via long horizons. On the $\tau$-bench retail dataset, Thinker achieves 82.6\% success rate with GPT-4o (version 2024-06-01) (baseline: 68.3\%), and 81.9\% success rate with Llama-3.1 405B (baseline: 49.6\%), without any fine-tuning. Thinker effectively closes the gap in reasoning capabilities between the base models by introducing proper structure. The key features of the Thinker framework are: (1) State-Machine Augmented Generation (SMAG), which represents business logic as state machines and the LLM uses state machines as tools. (2) Delegation of tasks from the main reasoning loop to LLM-powered tools. (3) Adaptive context management. Our prompting-only solution achieves signficant gains, while still maintaining a standard agentic architecture with a ReAct style reasoning loop. The key is to innovate on the tool interface design, as exemplified by SMAG and the LLM-powered tools.
☆ Can Large Language Models Predict Associations Among Human Attitudes?
Prior work has shown that large language models (LLMs) can predict human attitudes based on other attitudes, but this work has largely focused on predictions from highly similar and interrelated attitudes. In contrast, human attitudes are often strongly associated even across disparate and dissimilar topics. Using a novel dataset of human responses toward diverse attitude statements, we found that a frontier language model (GPT-4o) was able to recreate the pairwise correlations among individual attitudes and to predict individuals' attitudes from one another. Crucially, in an advance over prior work, we tested GPT-4o's ability to predict in the absence of surface-similarity between attitudes, finding that while surface similarity improves prediction accuracy, the model was still highly-capable of generating meaningful social inferences between dissimilar attitudes. Altogether, our findings indicate that LLMs capture crucial aspects of the deeper, latent structure of human belief systems.
☆ Improving User Behavior Prediction: Leveraging Annotator Metadata in Supervised Machine Learning Models SC
Supervised machine-learning models often underperform in predicting user behaviors from conversational text, hindered by poor crowdsourced label quality and low NLP task accuracy. We introduce the Metadata-Sensitive Weighted-Encoding Ensemble Model (MSWEEM), which integrates annotator meta-features like fatigue and speeding. First, our results show MSWEEM outperforms standard ensembles by 14\% on held-out data and 12\% on an alternative dataset. Second, we find that incorporating signals of annotator behavior, such as speed and fatigue, significantly boosts model performance. Third, we find that annotators with higher qualifications, such as Master's, deliver more consistent and faster annotations. Given the increasing uncertainty over annotation quality, our experiments show that understanding annotator patterns is crucial for enhancing model accuracy in user behavior prediction.
comment: Accepted at CSCW 2025
☆ FinAudio: A Benchmark for Audio Large Language Models in Financial Applications
Audio Large Language Models (AudioLLMs) have received widespread attention and have significantly improved performance on audio tasks such as conversation, audio understanding, and automatic speech recognition (ASR). Despite these advancements, there is an absence of a benchmark for assessing AudioLLMs in financial scenarios, where audio data, such as earnings conference calls and CEO speeches, are crucial resources for financial analysis and investment decisions. In this paper, we introduce \textsc{FinAudio}, the first benchmark designed to evaluate the capacity of AudioLLMs in the financial domain. We first define three tasks based on the unique characteristics of the financial domain: 1) ASR for short financial audio, 2) ASR for long financial audio, and 3) summarization of long financial audio. Then, we curate two short and two long audio datasets, respectively, and develop a novel dataset for financial audio summarization, comprising the \textsc{FinAudio} benchmark. Then, we evaluate seven prevalent AudioLLMs on \textsc{FinAudio}. Our evaluation reveals the limitations of existing AudioLLMs in the financial domain and offers insights for improving AudioLLMs. All datasets and codes will be released.
☆ Patients Speak, AI Listens: LLM-based Analysis of Online Reviews Uncovers Key Drivers for Urgent Care Satisfaction
Investigating the public experience of urgent care facilities is essential for promoting community healthcare development. Traditional survey methods often fall short due to limited scope, time, and spatial coverage. Crowdsourcing through online reviews or social media offers a valuable approach to gaining such insights. With recent advancements in large language models (LLMs), extracting nuanced perceptions from reviews has become feasible. This study collects Google Maps reviews across the DMV and Florida areas and conducts prompt engineering with the GPT model to analyze the aspect-based sentiment of urgent care. We first analyze the geospatial patterns of various aspects, including interpersonal factors, operational efficiency, technical quality, finances, and facilities. Next, we determine Census Block Group(CBG)-level characteristics underpinning differences in public perception, including population density, median income, GINI Index, rent-to-income ratio, household below poverty rate, no insurance rate, and unemployment rate. Our results show that interpersonal factors and operational efficiency emerge as the strongest determinants of patient satisfaction in urgent care, while technical quality, finances, and facilities show no significant independent effects when adjusted for in multivariate models. Among socioeconomic and demographic factors, only population density demonstrates a significant but modest association with patient ratings, while the remaining factors exhibit no significant correlations. Overall, this study highlights the potential of crowdsourcing to uncover the key factors that matter to residents and provide valuable insights for stakeholders to improve public satisfaction with urgent care.
☆ Competitive Multi-armed Bandit Games for Resource Sharing
In modern resource-sharing systems, multiple agents access limited resources with unknown stochastic conditions to perform tasks. When multiple agents access the same resource (arm) simultaneously, they compete for successful usage, leading to contention and reduced rewards. This motivates our study of competitive multi-armed bandit (CMAB) games. In this paper, we study a new N-player K-arm competitive MAB game, where non-myopic players (agents) compete with each other to form diverse private estimations of unknown arms over time. Their possible collisions on same arms and time-varying nature of arm rewards make the policy analysis more involved than existing studies for myopic players. We explicitly analyze the threshold-based structures of social optimum and existing selfish policy, showing that the latter causes prolonged convergence time $\Omega(\frac{K}{\eta^2}\ln({\frac{KN}{\delta}}))$, while socially optimal policy with coordinated communication reduces it to $\mathcal{O}(\frac{K}{N\eta^2}\ln{(\frac{K}{\delta})})$. Based on the comparison, we prove that the competition among selfish players for the best arm can result in an infinite price of anarchy (PoA), indicating an arbitrarily large efficiency loss compared to social optimum. We further prove that no informational (non-monetary) mechanism (including Bayesian persuasion) can reduce the infinite PoA, as the strategic misreporting by non-myopic players undermines such approaches. To address this, we propose a Combined Informational and Side-Payment (CISP) mechanism, which provides socially optimal arm recommendations with proper informational and monetary incentives to players according to their time-varying private beliefs. Our CISP mechanism keeps ex-post budget balanced for social planner and ensures truthful reporting from players, achieving the minimum PoA=1 and same convergence time as social optimum.
comment: This paper has been accepted by IEEE TMC
☆ Sociotechnical Effects of Machine Translation
While the previous chapters have shown how machine translation (MT) can be useful, in this chapter we discuss some of the side-effects and risks that are associated, and how they might be mitigated. With the move to neural MT and approaches using Large Language Models (LLMs), there is an associated impact on climate change, as the models built by multinational corporations are massive. They are hugely expensive to train, consume large amounts of electricity, and output huge volumes of kgCO2 to boot. However, smaller models which still perform to a high level of quality can be built with much lower carbon footprints, and tuning pre-trained models saves on the requirement to train from scratch. We also discuss the possible detrimental effects of MT on translators and other users. The topics of copyright and ownership of data are discussed, as well as ethical considerations on data and MT use. Finally, we show how if done properly, using MT in crisis scenarios can save lives, and we provide a method of how this might be done.
☆ TS-Inverse: A Gradient Inversion Attack Tailored for Federated Time Series Forecasting Models
Federated learning (FL) for time series forecasting (TSF) enables clients with privacy-sensitive time series (TS) data to collaboratively learn accurate forecasting models, for example, in energy load prediction. Unfortunately, privacy risks in FL persist, as servers can potentially reconstruct clients' training data through gradient inversion attacks (GIA). Although GIA is demonstrated for image classification tasks, little is known about time series regression tasks. In this paper, we first conduct an extensive empirical study on inverting TS data across 4 TSF models and 4 datasets, identifying the unique challenges of reconstructing both observations and targets of TS data. We then propose TS-Inverse, a novel GIA that improves the inversion of TS data by (i) learning a gradient inversion model that outputs quantile predictions, (ii) a unique loss function that incorporates periodicity and trend regularization, and (iii) regularization according to the quantile predictions. Our evaluations demonstrate a remarkable performance of TS-Inverse, achieving at least a 2x-10x improvement in terms of the sMAPE metric over existing GIA methods on TS data. Code repository: https://github.com/Capsar/ts-inverse
☆ DEMENTIA-PLAN: An Agent-Based Framework for Multi-Knowledge Graph Retrieval-Augmented Generation in Dementia Care AAAI 2025
Mild-stage dementia patients primarily experience two critical symptoms: severe memory loss and emotional instability. To address these challenges, we propose DEMENTIA-PLAN, an innovative retrieval-augmented generation framework that leverages large language models to enhance conversational support. Our model employs a multiple knowledge graph architecture, integrating various dimensional knowledge representations including daily routine graphs and life memory graphs. Through this multi-graph architecture, DEMENTIA-PLAN comprehensively addresses both immediate care needs and facilitates deeper emotional resonance through personal memories, helping stabilize patient mood while providing reliable memory support. Our notable innovation is the self-reflection planning agent, which systematically coordinates knowledge retrieval and semantic integration across multiple knowledge graphs, while scoring retrieved content from daily routine and life memory graphs to dynamically adjust their retrieval weights for optimized response generation. DEMENTIA-PLAN represents a significant advancement in the clinical application of large language models for dementia care, bridging the gap between AI tools and caregivers interventions.
comment: Accepted by AAAI 2025 Workshop on Knowledge Graphs for Personalized Public Health
☆ LATTE-MV: Learning to Anticipate Table Tennis Hits from Monocular Videos CVPR 2025
Physical agility is a necessary skill in competitive table tennis, but by no means sufficient. Champions excel in this fast-paced and highly dynamic environment by anticipating their opponent's intent - buying themselves the necessary time to react. In this work, we take one step towards designing such an anticipatory agent. Previous works have developed systems capable of real-time table tennis gameplay, though they often do not leverage anticipation. Among the works that forecast opponent actions, their approaches are limited by dataset size and variety. Our paper contributes (1) a scalable system for reconstructing monocular video of table tennis matches in 3D and (2) an uncertainty-aware controller that anticipates opponent actions. We demonstrate in simulation that our policy improves the ball return rate against high-speed hits from 49.9% to 59.0% as compared to a baseline non-anticipatory policy.
comment: CVPR 2025
☆ Prototype Guided Backdoor Defense
Deep learning models are susceptible to {\em backdoor attacks} involving malicious attackers perturbing a small subset of training data with a {\em trigger} to causes misclassifications. Various triggers have been used, including semantic triggers that are easily realizable without requiring the attacker to manipulate the image. The emergence of generative AI has eased the generation of varied poisoned samples. Robustness across types of triggers is crucial to effective defense. We propose Prototype Guided Backdoor Defense (PGBD), a robust post-hoc defense that scales across different trigger types, including previously unsolved semantic triggers. PGBD exploits displacements in the geometric spaces of activations to penalize movements toward the trigger. This is done using a novel sanitization loss of a post-hoc fine-tuning step. The geometric approach scales easily to all types of attacks. PGBD achieves better performance across all settings. We also present the first defense against a new semantic attack on celebrity face images. Project page: \hyperlink{https://venkatadithya9.github.io/pgbd.github.io/}{this https URL}.
☆ D4R -- Exploring and Querying Relational Graphs Using Natural Language and Large Language Models -- the Case of Historical Documents
D4R is a digital platform designed to assist non-technical users, particularly historians, in exploring textual documents through advanced graphical tools for text analysis and knowledge extraction. By leveraging a large language model, D4R translates natural language questions into Cypher queries, enabling the retrieval of data from a Neo4J database. A user-friendly graphical interface allows for intuitive interaction, enabling users to navigate and analyse complex relational data extracted from unstructured textual documents. Originally designed to bridge the gap between AI technologies and historical research, D4R's capabilities extend to various other domains. A demonstration video and a live software demo are available.
comment: 8 pages, 7 figures
☆ Assessing Generative Models for Structured Data
Synthetic tabular data generation has emerged as a promising method to address limited data availability and privacy concerns. With the sharp increase in the performance of large language models in recent years, researchers have been interested in applying these models to the generation of tabular data. However, little is known about the quality of the generated tabular data from large language models. The predominant method for assessing the quality of synthetic tabular data is the train-synthetic-test-real approach, where the artificial examples are compared to the original by how well machine learning models, trained separately on the real and synthetic sets, perform in some downstream tasks. This method does not directly measure how closely the distribution of generated data approximates that of the original. This paper introduces rigorous methods for directly assessing synthetic tabular data against real data by looking at inter-column dependencies within the data. We find that large language models (GPT-2), both when queried via few-shot prompting and when fine-tuned, and GAN (CTGAN) models do not produce data with dependencies that mirror the original real data. Results from this study can inform future practice in synthetic data generation to improve data quality.
☆ Robust Federated Learning Against Poisoning Attacks: A GAN-Based Defense Framework
Federated Learning (FL) enables collaborative model training across decentralized devices without sharing raw data, but it remains vulnerable to poisoning attacks that compromise model integrity. Existing defenses often rely on external datasets or predefined heuristics (e.g. number of malicious clients), limiting their effectiveness and scalability. To address these limitations, we propose a privacy-preserving defense framework that leverages a Conditional Generative Adversarial Network (cGAN) to generate synthetic data at the server for authenticating client updates, eliminating the need for external datasets. Our framework is scalable, adaptive, and seamlessly integrates into FL workflows. Extensive experiments on benchmark datasets demonstrate its robust performance against a variety of poisoning attacks, achieving high True Positive Rate (TPR) and True Negative Rate (TNR) of malicious and benign clients, respectively, while maintaining model accuracy. The proposed framework offers a practical and effective solution for securing federated learning systems.
☆ VinaBench: Benchmark for Faithful and Consistent Visual Narratives CVPR 2025
Visual narrative generation transforms textual narratives into sequences of images illustrating the content of the text. However, generating visual narratives that are faithful to the input text and self-consistent across generated images remains an open challenge, due to the lack of knowledge constraints used for planning the stories. In this work, we propose a new benchmark, VinaBench, to address this challenge. Our benchmark annotates the underlying commonsense and discourse constraints in visual narrative samples, offering systematic scaffolds for learning the implicit strategies of visual storytelling. Based on the incorporated narrative constraints, we further propose novel metrics to closely evaluate the consistency of generated narrative images and the alignment of generations with the input textual narrative. Our results across three generative vision models demonstrate that learning with VinaBench's knowledge constraints effectively improves the faithfulness and cohesion of generated visual narratives.
comment: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR 2025)
☆ Unified Multimodal Discrete Diffusion
Multimodal generative models that can understand and generate across multiple modalities are dominated by autoregressive (AR) approaches, which process tokens sequentially from left to right, or top to bottom. These models jointly handle images, text, video, and audio for various tasks such as image captioning, question answering, and image generation. In this work, we explore discrete diffusion models as a unified generative formulation in the joint text and image domain, building upon their recent success in text generation. Discrete diffusion models offer several advantages over AR models, including improved control over quality versus diversity of generated samples, the ability to perform joint multimodal inpainting (across both text and image domains), and greater controllability in generation through guidance. Leveraging these benefits, we present the first Unified Multimodal Discrete Diffusion (UniDisc) model which is capable of jointly understanding and generating text and images for a variety of downstream tasks. We compare UniDisc to multimodal AR models, performing a scaling analysis and demonstrating that UniDisc outperforms them in terms of both performance and inference-time compute, enhanced controllability, editability, inpainting, and flexible trade-off between inference time and generation quality. Code and additional visualizations are available at https://unidisc.github.io.
comment: Project Website: https://unidisc.github.io
☆ The Backfiring Effect of Weak AI Safety Regulation
Recent policy proposals aim to improve the safety of general-purpose AI, but there is little understanding of the efficacy of different regulatory approaches to AI safety. We present a strategic model that explores the interactions between the regulator, the general-purpose AI technology creators, and domain specialists--those who adapt the AI for specific applications. Our analysis examines how different regulatory measures, targeting different parts of the development chain, affect the outcome of the development process. In particular, we assume AI technology is described by two key attributes: safety and performance. The regulator first sets a minimum safety standard that applies to one or both players, with strict penalties for non-compliance. The general-purpose creator then develops the technology, establishing its initial safety and performance levels. Next, domain specialists refine the AI for their specific use cases, and the resulting revenue is distributed between the specialist and generalist through an ex-ante bargaining process. Our analysis of this game reveals two key insights: First, weak safety regulation imposed only on the domain specialists can backfire. While it might seem logical to regulate use cases (as opposed to the general-purpose technology), our analysis shows that weak regulations targeting domain specialists alone can unintentionally reduce safety. This effect persists across a wide range of settings. Second, in sharp contrast to the previous finding, we observe that stronger, well-placed regulation can in fact benefit all players subjected to it. When regulators impose appropriate safety standards on both AI creators and domain specialists, the regulation functions as a commitment mechanism, leading to safety and performance gains, surpassing what is achieved under no regulation or regulating one player only.
comment: 28 pages, 8 figures
☆ Robust Deep Reinforcement Learning in Robotics via Adaptive Gradient-Masked Adversarial Attacks
Deep reinforcement learning (DRL) has emerged as a promising approach for robotic control, but its realworld deployment remains challenging due to its vulnerability to environmental perturbations. Existing white-box adversarial attack methods, adapted from supervised learning, fail to effectively target DRL agents as they overlook temporal dynamics and indiscriminately perturb all state dimensions, limiting their impact on long-term rewards. To address these challenges, we propose the Adaptive Gradient-Masked Reinforcement (AGMR) Attack, a white-box attack method that combines DRL with a gradient-based soft masking mechanism to dynamically identify critical state dimensions and optimize adversarial policies. AGMR selectively allocates perturbations to the most impactful state features and incorporates a dynamic adjustment mechanism to balance exploration and exploitation during training. Extensive experiments demonstrate that AGMR outperforms state-of-the-art adversarial attack methods in degrading the performance of the victim agent and enhances the victim agent's robustness through adversarial defense mechanisms.
comment: 9 pages, 6 figures
☆ Advancing Vulnerability Classification with BERT: A Multi-Objective Learning Model
The rapid increase in cybersecurity vulnerabilities necessitates automated tools for analyzing and classifying vulnerability reports. This paper presents a novel Vulnerability Report Classifier that leverages the BERT (Bidirectional Encoder Representations from Transformers) model to perform multi-label classification of Common Vulnerabilities and Exposures (CVE) reports from the National Vulnerability Database (NVD). The classifier predicts both the severity (Low, Medium, High, Critical) and vulnerability types (e.g., Buffer Overflow, XSS) from textual descriptions. We introduce a custom training pipeline using a combined loss function-Cross-Entropy for severity and Binary Cross-Entropy with Logits for types-integrated into a Hugging Face Trainer subclass. Experiments on recent NVD data demonstrate promising results, with decreasing evaluation loss across epochs. The system is deployed via a REST API and a Streamlit UI, enabling real-time vulnerability analysis. This work contributes a scalable, open-source solution for cybersecurity practitioners to automate vulnerability triage.
comment: 9 Pages
☆ Exploiting Temporal State Space Sharing for Video Semantic Segmentation
Video semantic segmentation (VSS) plays a vital role in understanding the temporal evolution of scenes. Traditional methods often segment videos frame-by-frame or in a short temporal window, leading to limited temporal context, redundant computations, and heavy memory requirements. To this end, we introduce a Temporal Video State Space Sharing (TV3S) architecture to leverage Mamba state space models for temporal feature sharing. Our model features a selective gating mechanism that efficiently propagates relevant information across video frames, eliminating the need for a memory-heavy feature pool. By processing spatial patches independently and incorporating shifted operation, TV3S supports highly parallel computation in both training and inference stages, which reduces the delay in sequential state space processing and improves the scalability for long video sequences. Moreover, TV3S incorporates information from prior frames during inference, achieving long-range temporal coherence and superior adaptability to extended sequences. Evaluations on the VSPW and Cityscapes datasets reveal that our approach outperforms current state-of-the-art methods, establishing a new standard for VSS with consistent results across long video sequences. By achieving a good balance between accuracy and efficiency, TV3S shows a significant advancement in spatiotemporal modeling, paving the way for efficient video analysis. The code is publicly available at https://github.com/Ashesham/TV3S.git.
comment: IEEE/CVF Conference on Computer Vision and Pattern Recognition 2025
☆ Synthetic Video Enhances Physical Fidelity in Video Synthesis
We investigate how to enhance the physical fidelity of video generation models by leveraging synthetic videos derived from computer graphics pipelines. These rendered videos respect real-world physics, such as maintaining 3D consistency, and serve as a valuable resource that can potentially improve video generation models. To harness this potential, we propose a solution that curates and integrates synthetic data while introducing a method to transfer its physical realism to the model, significantly reducing unwanted artifacts. Through experiments on three representative tasks emphasizing physical consistency, we demonstrate its efficacy in enhancing physical fidelity. While our model still lacks a deep understanding of physics, our work offers one of the first empirical demonstrations that synthetic video enhances physical fidelity in video synthesis. Website: https://kevinz8866.github.io/simulation/
♻ ☆ Assessing Consistency and Reproducibility in the Outputs of Large Language Models: Evidence Across Diverse Finance and Accounting Tasks
This study provides the first comprehensive assessment of consistency and reproducibility in Large Language Model (LLM) outputs in finance and accounting research. We evaluate how consistently LLMs produce outputs given identical inputs through extensive experimentation with 50 independent runs across five common tasks: classification, sentiment analysis, summarization, text generation, and prediction. Using three OpenAI models (GPT-3.5-turbo, GPT-4o-mini, and GPT-4o), we generate over 3.4 million outputs from diverse financial source texts and data, covering MD&As, FOMC statements, finance news articles, earnings call transcripts, and financial statements. Our findings reveal substantial but task-dependent consistency, with binary classification and sentiment analysis achieving near-perfect reproducibility, while complex tasks show greater variability. More advanced models do not consistently demonstrate better consistency and reproducibility, with task-specific patterns emerging. LLMs significantly outperform expert human annotators in consistency and maintain high agreement even where human experts significantly disagree. We further find that simple aggregation strategies across 3-5 runs dramatically improve consistency. We also find that aggregation may come with an additional benefit of improved accuracy for sentiment analysis when using newer models. Simulation analysis reveals that despite measurable inconsistency in LLM outputs, downstream statistical inferences remain remarkably robust. These findings address concerns about what we term "G-hacking," the selective reporting of favorable outcomes from multiple Generative AI runs, by demonstrating that such risks are relatively low for finance and accounting tasks.
comment: 97 pages, 20 tables, 15 figures
♻ ☆ Task-Specific Activation Functions for Neuroevolution using Grammatical Evolution
Activation functions play a critical role in the performance and behaviour of neural networks, significantly impacting their ability to learn and generalise. Traditional activation functions, such as ReLU, sigmoid, and tanh, have been widely used with considerable success. However, these functions may not always provide optimal performance for all tasks and datasets. In this paper, we introduce Neuvo GEAF - an innovative approach leveraging grammatical evolution (GE) to automatically evolve novel activation functions tailored to specific neural network architectures and datasets. Experiments conducted on well-known binary classification datasets show statistically significant improvements in F1-score (between 2.4% and 9.4%) over ReLU using identical network architectures. Notably, these performance gains were achieved without increasing the network's parameter count, supporting the trend toward more efficient neural networks that can operate effectively on resource-constrained edge devices. This paper's findings suggest that evolved activation functions can provide significant performance improvements for compact networks while maintaining energy efficiency during both training and inference phases.
comment: 8 pages, 4 figures, IEEE
♻ ☆ Graph-Instructed Neural Networks for Sparse Grid-Based Discontinuity Detectors
In this paper, we present a novel approach for detecting the discontinuity interfaces of a discontinuous function. This approach leverages Graph-Instructed Neural Networks (GINNs) and sparse grids to address discontinuity detection also in domains of dimension larger than 3. GINNs, trained to identify troubled points on sparse grids, exploit graph structures built on the grids to achieve efficient and accurate discontinuity detection performances. We also introduce a recursive algorithm for general sparse grid-based detectors, characterized by convergence properties and easy applicability. Numerical experiments on functions with dimensions n = 2 and n = 4 demonstrate the efficiency and robust generalization properties of GINNs in detecting discontinuity interfaces. Notably, the trained GINNs offer portability and versatility, allowing integration into various algorithms and sharing among users.
♻ ☆ Certified Robustness via Dynamic Margin Maximization and Improved Lipschitz Regularization NeurIPS 2023
To improve the robustness of deep classifiers against adversarial perturbations, many approaches have been proposed, such as designing new architectures with better robustness properties (e.g., Lipschitz-capped networks), or modifying the training process itself (e.g., min-max optimization, constrained learning, or regularization). These approaches, however, might not be effective at increasing the margin in the input (feature) space. As a result, there has been an increasing interest in developing training procedures that can directly manipulate the decision boundary in the input space. In this paper, we build upon recent developments in this category by developing a robust training algorithm whose objective is to increase the margin in the output (logit) space while regularizing the Lipschitz constant of the model along vulnerable directions. We show that these two objectives can directly promote larger margins in the input space. To this end, we develop a scalable method for calculating guaranteed differentiable upper bounds on the Lipschitz constant of neural networks accurately and efficiently. The relative accuracy of the bounds prevents excessive regularization and allows for more direct manipulation of the decision boundary. Furthermore, our Lipschitz bounding algorithm exploits the monotonicity and Lipschitz continuity of the activation layers, and the resulting bounds can be used to design new layers with controllable bounds on their Lipschitz constant. Experiments on the MNIST, CIFAR-10, and Tiny-ImageNet data sets verify that our proposed algorithm obtains competitively improved results compared to the state-of-the-art.
comment: 37th Conference on Neural Information Processing Systems (NeurIPS 2023)
♻ ☆ Semiring Provenance for Lightweight Description Logics
We investigate semiring provenance--a successful framework originally defined in the relational database setting--for description logics. In this context, the ontology axioms are annotated with elements of a commutative semiring and these annotations are propagated to the ontology consequences in a way that reflects how they are derived. We define a provenance semantics for a language that encompasses several lightweight description logics and show its relationships with semantics that have been defined for ontologies annotated with a specific kind of annotation (such as fuzzy degrees). We show that under some restrictions on the semiring, the semantics satisfies desirable properties (such as extending the semiring provenance defined for databases). We then focus on the well-known why-provenance, for which we study the complexity of problems related to the provenance of an assertion or a conjunctive query answer. Finally, we consider two more restricted cases which correspond to the so-called positive Boolean provenance and lineage in the database setting. For these cases, we exhibit relationships with well-known notions related to explanations in description logics and complete our complexity analysis. As a side contribution, we provide conditions on an $\mathcal{ELHI}_\bot$ ontology that guarantee tractable reasoning.
comment: Paper currently under review. 133 pages
♻ ☆ Networking Systems for Video Anomaly Detection: A Tutorial and Survey
The increasing utilization of surveillance cameras in smart cities, coupled with the surge of online video applications, has heightened concerns regarding public security and privacy protection, which propelled automated Video Anomaly Detection (VAD) into a fundamental research task within the Artificial Intelligence (AI) community. With the advancements in deep learning and edge computing, VAD has made significant progress and advances synergized with emerging applications in smart cities and video internet, which has moved beyond the conventional research scope of algorithm engineering to deployable Networking Systems for VAD (NSVAD), a practical hotspot for intersection exploration in the AI, IoVT, and computing fields. In this article, we delineate the foundational assumptions, learning frameworks, and applicable scenarios of various deep learning-driven VAD routes, offering an exhaustive tutorial for novices in NSVAD. In addition, this article elucidates core concepts by reviewing recent advances and typical solutions and aggregating available research resources accessible at https://github.com/fdjingliu/NSVAD. Lastly, this article projects future development trends and discusses how the integration of AI and computing technologies can address existing research challenges and promote open opportunities, serving as an insightful guide for prospective researchers and engineers.
comment: Revised to ACM Computing Surveys, under review, for more information and supplementary material, please see https://github.com/fdjingliu/NSVAD
♻ ☆ Data Augmentation in Earth Observation: A Diffusion Model Approach
High-quality Earth Observation (EO) imagery is essential for accurate analysis and informed decision making across sectors. However, data scarcity caused by atmospheric conditions, seasonal variations, and limited geographical coverage hinders the effective application of Artificial Intelligence (AI) in EO. Traditional data augmentation techniques, which rely on basic parameterized image transformations, often fail to introduce sufficient diversity across key semantic axes. These axes include natural changes such as snow and floods, human impacts like urbanization and roads, and disasters such as wildfires and storms, which limits the accuracy of AI models in EO applications. To address this, we propose a four-stage data augmentation approach that integrates diffusion models to enhance semantic diversity. Our method employs meta-prompts for instruction generation, vision-language models for rich captioning, EO-specific diffusion model fine-tuning, and iterative data augmentation. Extensive experiments using four augmentation techniques demonstrate that our approach consistently outperforms established methods, generating semantically diverse EO images and improving AI model performance.
comment: 25 pages, 12 figures
♻ ☆ Towards Scalable Foundation Model for Multi-modal and Hyperspectral Geospatial Data
Geospatial raster data, such as that collected by satellite-based imaging systems at different times and spectral bands, hold immense potential for enabling a wide range of high-impact applications. This potential stems from the rich information that is spatially and temporally contextualized across multiple channels and sensing modalities. Recent work has adapted existing self-supervised learning approaches for such geospatial data. However, they fall short of scalable model architectures, leading to inflexibility and computational inefficiencies when faced with an increasing number of channels and modalities. To address these limitations, we introduce Low-rank Efficient Spatial-Spectral Vision Transformer with three key innovations: i) the LESS Attention Block that approximates high-dimensional spatial-spectral attention through Kronecker's product of the low-dimensional spatial and spectral attention components; ii) the Continuous Positional-Channel Embedding Layer that preserves both the continuity and physical characteristics of each spatial-spectral patch; and iii) the Perception Field Mask that exploits local spatial dependencies by constraining attention to neighboring patches. To evaluate the proposed innovations, we construct GFM-Bench, which serves as a comprehensive benchmark for such geospatial raster data. We pretrain LESS ViT using a Hyperspectral Masked Autoencoder framework with integrated positional and channel masking strategies. Experimental results demonstrate that our proposed method achieves competitive performance against state-of-the-art multi-modal geospatial foundation models while outperforming them on cross-satellite generalization tasks with higher computational efficiency. The flexibility and extensibility of our framework make it a promising direction for future geospatial data analysis tasks that involve a wide range of modalities and channels.
♻ ☆ COSMOS: Cross-Modality Self-Distillation for Vision Language Pre-training CVPR 2025
Vision-Language Models (VLMs) trained with contrastive loss have achieved significant advancements in various vision and language tasks. However, the global nature of the contrastive loss makes VLMs focus predominantly on foreground objects, neglecting other crucial information in the image, which limits their effectiveness in downstream tasks. To address these challenges, we propose COSMOS: CrOSs-MOdality Self-distillation for vision-language pre-training that integrates a novel text-cropping strategy and cross-attention module into a self-supervised learning framework. We create global and local views of images and texts (i.e., multi-modal augmentations), which are essential for self-distillation in VLMs. We further introduce a cross-attention module, enabling COSMOS to learn comprehensive cross-modal representations optimized via a cross-modality self-distillation loss. COSMOS consistently outperforms previous strong baselines on various zero-shot downstream tasks, including retrieval, classification, and semantic segmentation. Additionally, it surpasses CLIP-based models trained on larger datasets in visual perception and contextual understanding tasks. Code is available at https://github.com/ExplainableML/cosmos.
comment: CVPR 2025
♻ ☆ Context-Aware Semantic Recomposition Mechanism for Large Language Models
Context-aware processing mechanisms have increasingly become a critical area of exploration for improving the semantic and contextual capabilities of language generation models. The Context-Aware Semantic Recomposition Mechanism (CASRM) was introduced as a novel framework designed to address limitations in coherence, contextual adaptability, and error propagation in large-scale text generation tasks. Through the integration of dynamically generated context vectors and attention modulation layers, CASRM enhances the alignment between token-level representations and broader contextual dependencies. Experimental evaluations demonstrated significant improvements in semantic coherence across multiple domains, including technical, conversational, and narrative text. The ability to adapt to unseen domains and ambiguous inputs was evaluated using a diverse set of test scenarios, highlighting the robustness of the proposed mechanism. A detailed computational analysis revealed that while CASRM introduces additional processing overhead, the gains in linguistic precision and contextual relevance outweigh the marginal increase in complexity. The framework also successfully mitigates error propagation in sequential tasks, improving performance in dialogue continuation and multi-step text synthesis. Additional investigations into token-level attention distribution emphasized the dynamic focus shifts enabled through context-aware enhancements. The findings suggest that CASRM offers a scalable and flexible solution for integrating contextual intelligence into existing language model architectures.
comment: arXiv admin note: This paper has been withdrawn by arXiv due to disputed and unverifiable authorship
♻ ☆ Intelligent Code Embedding Framework for High-Precision Ransomware Detection via Multimodal Execution Path Analysis
Modern threat landscapes continue to evolve with increasing sophistication, challenging traditional detection methodologies and necessitating innovative solutions capable of addressing complex adversarial tactics. A novel framework was developed to identify ransomware activity through multimodal execution path analysis, integrating high-dimensional embeddings and dynamic heuristic derivation mechanisms to capture behavioral patterns across diverse attack variants. The approach demonstrated high adaptability, effectively mitigating obfuscation strategies and polymorphic characteristics often employed by ransomware families to evade detection. Comprehensive experimental evaluations revealed significant advancements in precision, recall, and accuracy metrics compared to baseline techniques, particularly under conditions of variable encryption speeds and obfuscated execution flows. The framework achieved scalable and computationally efficient performance, ensuring robust applicability across a range of system configurations, from resource-constrained environments to high-performance infrastructures. Notable findings included reduced false positive rates and enhanced detection latency, even for ransomware families employing sophisticated encryption mechanisms. The modular design allowed seamless integration of additional modalities, enabling extensibility and future-proofing against emerging threat vectors. Quantitative analyses further highlighted the system's energy efficiency, emphasizing its practicality for deployment in environments with stringent operational constraints. The results underline the importance of integrating advanced computational techniques and dynamic adaptability to safeguard digital ecosystems from increasingly complex threats.
comment: arXiv admin note: This paper has been withdrawn by arXiv due to disputed and unverifiable authorship
♻ ☆ MC-LLaVA: Multi-Concept Personalized Vision-Language Model
Current vision-language models (VLMs) show exceptional abilities across diverse tasks, such as visual question answering. To enhance user experience, recent studies investigate VLM personalization to understand user-provided concepts. However, they mainly focus on single-concept personalization, neglecting the existence and interplay of multiple concepts, which limits real-world applicability. This paper proposes the first multi-concept personalization paradigm, MC-LLaVA. Specifically, MC-LLaVA employs a multi-concept instruction tuning strategy, effectively integrating multiple concepts in a single training step. To reduce the costs related to joint training, we propose a personalized textual prompt that uses visual token information to initialize concept tokens. Additionally, we introduce a personalized visual prompt during inference, aggregating location confidence maps for enhanced recognition and grounding capabilities. To advance multi-concept personalization research, we further contribute a high-quality instruction tuning dataset. We carefully collect images with multiple characters and objects from movies and manually generate question-answer samples for multi-concept scenarios, featuring superior diversity. Comprehensive qualitative and quantitative experiments demonstrate that MC-LLaVA can achieve impressive multi-concept personalized responses, paving the way for VLMs to become better user-specific assistants. The code and dataset will be publicly available at https://github.com/arctanxarc/MC-LLaVA.
♻ ☆ Intuitive Axial Augmentation Using Polar-Sine-Based Piecewise Distortion for Medical Slice-Wise Segmentation
Most data-driven models for medical image analysis rely on universal augmentations to improve accuracy. Experimental evidence has confirmed their effectiveness, but the unclear mechanism underlying them poses a barrier to the widespread acceptance and trust in such methods within the medical community. We revisit and acknowledge the unique characteristics of medical images apart from traditional digital images, and consequently, proposed a medical-specific augmentation algorithm that is more elastic and aligns well with radiology scan procedure. The method performs piecewise affine with sinusoidal distorted ray according to radius on polar coordinates, thus simulating uncertain postures of human lying flat on the scanning table. Our method could generate human visceral distribution without affecting the fundamental relative position on axial plane. Two non-adaptive algorithms, namely Meta-based Scan Table Removal and Similarity-Guided Parameter Search, are introduced to bolster robustness of our augmentation method. In contrast to other methodologies, our method is highlighted for its intuitive design and ease of understanding for medical professionals, thereby enhancing its applicability in clinical scenarios. Experiments show our method improves accuracy with two modality across multiple famous segmentation frameworks without requiring more data samples. Our preview code is available in: https://github.com/MGAMZ/PSBPD.
comment: Published at Smart Health
♻ ☆ Black-Box Forgery Attacks on Semantic Watermarks for Diffusion Models CVPR
Integrating watermarking into the generation process of latent diffusion models (LDMs) simplifies detection and attribution of generated content. Semantic watermarks, such as Tree-Rings and Gaussian Shading, represent a novel class of watermarking techniques that are easy to implement and highly robust against various perturbations. However, our work demonstrates a fundamental security vulnerability of semantic watermarks. We show that attackers can leverage unrelated models, even with different latent spaces and architectures (UNet vs DiT), to perform powerful and realistic forgery attacks. Specifically, we design two watermark forgery attacks. The first imprints a targeted watermark into real images by manipulating the latent representation of an arbitrary image in an unrelated LDM to get closer to the latent representation of a watermarked image. We also show that this technique can be used for watermark removal. The second attack generates new images with the target watermark by inverting a watermarked image and re-generating it with an arbitrary prompt. Both attacks just need a single reference image with the target watermark. Overall, our findings question the applicability of semantic watermarks by revealing that attackers can easily forge or remove these watermarks under realistic conditions.
comment: 28 pages, 22 figures, 8 tables, to be published in The IEEE/CVF Conference on Computer Vision and Pattern Recognition 2025 (CVPR)
♻ ☆ Unleashing Vecset Diffusion Model for Fast Shape Generation
3D shape generation has greatly flourished through the development of so-called "native" 3D diffusion, particularly through the Vecset Diffusion Model (VDM). While recent advancements have shown promising results in generating high-resolution 3D shapes, VDM still struggles with high-speed generation. Challenges exist because of difficulties not only in accelerating diffusion sampling but also VAE decoding in VDM, areas under-explored in previous works. To address these challenges, we present FlashVDM, a systematic framework for accelerating both VAE and DiT in VDM. For DiT, FlashVDM enables flexible diffusion sampling with as few as 5 inference steps and comparable quality, which is made possible by stabilizing consistency distillation with our newly introduced Progressive Flow Distillation. For VAE, we introduce a lightning vecset decoder equipped with Adaptive KV Selection, Hierarchical Volume Decoding, and Efficient Network Design. By exploiting the locality of the vecset and the sparsity of shape surface in the volume, our decoder drastically lowers FLOPs, minimizing the overall decoding overhead. We apply FlashVDM to Hunyuan3D-2 to obtain Hunyuan3D-2 Turbo. Through systematic evaluation, we show that our model significantly outperforms existing fast 3D generation methods, achieving comparable performance to the state-of-the-art while reducing inference time by over 45x for reconstruction and 32x for generation. Code and models are available at https://github.com/Tencent/FlashVDM.
comment: Technical report
♻ ☆ The importance of the clustering model to detect new types of intrusion in data traffic
In the current digital age, the volume of data generated by various cyber activities has become enormous and is constantly increasing. The data may contain valuable insights that can be harnessed to improve cyber security measures. However, much of this data is unclassified and qualitative, which poses significant challenges to traditional analysis methods. Clustering facilitates the identification of hidden patterns and structures in data through grouping similar data points, which makes it simpler to identify and address threats. Clustering can be defined as a data mining (DM) approach, which uses similarity calculations for dividing a data set into several categories. Hierarchical, density-based, along with partitioning clustering algorithms are typical. The presented work use K-means algorithm, which is a popular clustering technique. Utilizing K-means algorithm, we worked with two different types of data: first, we gathered data with the use of XG-boost algorithm following completing the aggregation with K-means algorithm. Data was gathered utilizing Kali Linux environment, cicflowmeter traffic, and Putty Software tools with the use of diverse and simple attacks. The concept could assist in identifying new attack types, which are distinct from the known attacks, and labeling them based on the characteristics they will exhibit, as the dynamic nature regarding cyber threats means that new attack types often emerge, for which labeled data might not yet exist. The model counted the attacks and assigned numbers to each one of them. Secondly, We tried the same work on the ready data inside the Kaggle repository called (Intrusion Detection in Internet of Things Network), and the clustering model worked well and detected the number of attacks correctly as shown in the results section.
comment: 18 pages, 4 figures
♻ ☆ PRECTR: A Synergistic Framework for Integrating Personalized Search Relevance Matching and CTR Prediction
The two primary tasks in the search recommendation system are search relevance matching and click-through rate (CTR) prediction -- the former focuses on seeking relevant items for user queries whereas the latter forecasts which item may better match user interest. Prior research typically develops two models to predict the CTR and search relevance separately, then ranking candidate items based on the fusion of the two outputs. However, such a divide-and-conquer paradigm creates the inconsistency between different models. Meanwhile, the search relevance model mainly concentrates on the degree of objective text matching while neglecting personalized differences among different users, leading to restricted model performance. To tackle these issues, we propose a unified Personalized Search RElevance Matching and CTR Prediction Fusion Model(PRECTR). Specifically, based on the conditional probability fusion mechanism, PRECTR integrates the CTR prediction and search relevance matching into one framework to enhance the interaction and consistency of the two modules. However, directly optimizing CTR binary classification loss may bring challenges to the fusion model's convergence and indefinitely promote the exposure of items with high CTR, regardless of their search relevance. Hence, we further introduce two-stage training and semantic consistency regularization to accelerate the model's convergence and restrain the recommendation of irrelevant items. Finally, acknowledging that different users may have varied relevance preferences, we assessed current users' relevance preferences by analyzing past users' preferences for similar queries and tailored incentives for different candidate items accordingly. Extensive experimental results on our production dataset and online A/B testing demonstrate the effectiveness and superiority of our proposed PRECTR method.
♻ ☆ Fully Distributed Fog Load Balancing with Multi-Agent Reinforcement Learning
Real-time Internet of Things (IoT) applications require real-time support to handle the ever-growing demand for computing resources to process IoT workloads. Fog Computing provides high availability of such resources in a distributed manner. However, these resources must be efficiently managed to distribute unpredictable traffic demands among heterogeneous Fog resources. This paper proposes a fully distributed load-balancing solution with Multi-Agent Reinforcement Learning (MARL) that intelligently distributes IoT workloads to optimize the waiting time while providing fair resource utilization in the Fog network. These agents use transfer learning for life-long self-adaptation to dynamic changes in the environment. By leveraging distributed decision-making, MARL agents effectively minimize the waiting time compared to a single centralized agent solution and other baselines, enhancing end-to-end execution delay. Besides performance gain, a fully distributed solution allows for a global-scale implementation where agents can work independently in small collaboration regions, leveraging nearby local resources. Furthermore, we analyze the impact of a realistic frequency to observe the state of the environment, unlike the unrealistic common assumption in the literature of having observations readily available in real-time for every required action. The findings highlight the trade-off between realism and performance using an interval-based Gossip-based multi-casting protocol against assuming real-time observation availability for every generated workload.
comment: Submitted to IEEE TNSM with 14 pages, 11 figures, and 3 tables
♻ ☆ PG-SAM: Prior-Guided SAM with Medical for Multi-organ Segmentation
Segment Anything Model (SAM) demonstrates powerful zero-shot capabilities; however, its accuracy and robustness significantly decrease when applied to medical image segmentation. Existing methods address this issue through modality fusion, integrating textual and image information to provide more detailed priors. In this study, we argue that the granularity of text and the domain gap affect the accuracy of the priors. Furthermore, the discrepancy between high-level abstract semantics and pixel-level boundary details in images can introduce noise into the fusion process. To address this, we propose Prior-Guided SAM (PG-SAM), which employs a fine-grained modality prior aligner to leverage specialized medical knowledge for better modality alignment. The core of our method lies in efficiently addressing the domain gap with fine-grained text from a medical LLM. Meanwhile, it also enhances the priors' quality after modality alignment, ensuring more accurate segmentation. In addition, our decoder enhances the model's expressive capabilities through multi-level feature fusion and iterative mask optimizer operations, supporting unprompted learning. We also propose a unified pipeline that effectively supplies high-quality semantic information to SAM. Extensive experiments on the Synapse dataset demonstrate that the proposed PG-SAM achieves state-of-the-art performance. Our code is released at https://github.com/logan-0623/PG-SAM.
♻ ☆ OASST-ETC Dataset: Alignment Signals from Eye-tracking Analysis of LLM Responses
While Large Language Models (LLMs) have significantly advanced natural language processing, aligning them with human preferences remains an open challenge. Although current alignment methods rely primarily on explicit feedback, eye-tracking (ET) data offers insights into real-time cognitive processing during reading. In this paper, we present OASST-ETC, a novel eye-tracking corpus capturing reading patterns from 24 participants, while evaluating LLM-generated responses from the OASST1 dataset. Our analysis reveals distinct reading patterns between preferred and non-preferred responses, which we compare with synthetic eye-tracking data. Furthermore, we examine the correlation between human reading measures and attention patterns from various transformer-based models, discovering stronger correlations in preferred responses. This work introduces a unique resource for studying human cognitive processing in LLM evaluation and suggests promising directions for incorporating eye-tracking data into alignment methods. The dataset and analysis code are publicly available.
comment: This paper has been accepted to ACM ETRA 2025 and published on PACMHCI
♻ ☆ A Survey on Event-driven 3D Reconstruction: Development under Different Categories
Event cameras have gained increasing attention for 3D reconstruction due to their high temporal resolution, low latency, and high dynamic range. They capture per-pixel brightness changes asynchronously, allowing accurate reconstruction under fast motion and challenging lighting conditions. In this survey, we provide a comprehensive review of event-driven 3D reconstruction methods, including stereo, monocular, and multimodal systems. We further categorize recent developments based on geometric, learning-based, and hybrid approaches. Emerging trends, such as neural radiance fields and 3D Gaussian splatting with event data, are also covered. The related works are structured chronologically to illustrate the innovations and progression within the field. To support future research, we also highlight key research gaps and future research directions in dataset, experiment, evaluation, event representation, etc.
comment: 6 pages, 1 figure, 6 tables, submitted to an anonymous conference under double-blind review
♻ ☆ Bayesian Modeling of Zero-Shot Classifications for Urban Flood Detection
Street scene datasets, collected from Street View or dashboard cameras, offer a promising means of detecting urban objects and incidents like street flooding. However, a major challenge in using these datasets is their lack of reliable labels: there are myriad types of incidents, many types occur rarely, and ground-truth measures of where incidents occur are lacking. Here, we propose BayFlood, a two-stage approach which circumvents this difficulty. First, we perform zero-shot classification of where incidents occur using a pretrained vision-language model (VLM). Second, we fit a spatial Bayesian model on the VLM classifications. The zero-shot approach avoids the need to annotate large training sets, and the Bayesian model provides frequent desiderata in urban settings - principled measures of uncertainty, smoothing across locations, and incorporation of external data like stormwater accumulation zones. We comprehensively validate this two-stage approach, showing that VLMs provide strong zero-shot signal for floods across multiple cities and time periods, the Bayesian model improves out-of-sample prediction relative to baseline methods, and our inferred flood risk correlates with known external predictors of risk. Having validated our approach, we show it can be used to improve urban flood detection: our analysis reveals 113,738 people who are at high risk of flooding overlooked by current methods, identifies demographic biases in existing methods, and suggests locations for new flood sensors. More broadly, our results showcase how Bayesian modeling of zero-shot LM annotations represents a promising paradigm because it avoids the need to collect large labeled datasets and leverages the power of foundation models while providing the expressiveness and uncertainty quantification of Bayesian models.
comment: In review
♻ ☆ Fantastic Copyrighted Beasts and How (Not) to Generate Them
Recent studies show that image and video generation models can be prompted to reproduce copyrighted content from their training data, raising serious legal concerns about copyright infringement. Copyrighted characters (e.g., Mario, Batman) present a significant challenge: at least one lawsuit has already awarded damages based on the generation of such characters. Consequently, commercial services like DALL-E have started deploying interventions. However, little research has systematically examined these problems: (1) Can users easily prompt models to generate copyrighted characters, even if it is unintentional?; (2) How effective are the existing mitigation strategies? To address these questions, we introduce a novel evaluation framework with metrics that assess both the generated image's similarity to copyrighted characters and its consistency with user intent, grounded in a set of popular copyrighted characters from diverse studios and regions. We show that state-of-the-art image and video generation models can still generate characters even if characters' names are not explicitly mentioned, sometimes with only two generic keywords (e.g., prompting with "videogame, plumber" consistently generates Nintendo's Mario character). We also introduce semi-automatic techniques to identify such keywords or descriptions that trigger character generation. Using this framework, we evaluate mitigation strategies, including prompt rewriting and new approaches we propose. Our findings reveal that common methods, such as DALL-E's prompt rewriting, are insufficient alone and require supplementary strategies like negative prompting. Our work provides empirical grounding for discussions on copyright mitigation strategies and offers actionable insights for model deployers implementing these safeguards.
♻ ☆ Towards End-to-End Neuromorphic Voxel-based 3D Object Reconstruction Without Physical Priors ICME
Neuromorphic cameras, also known as event cameras, are asynchronous brightness-change sensors that can capture extremely fast motion without suffering from motion blur, making them particularly promising for 3D reconstruction in extreme environments. However, existing research on 3D reconstruction using monocular neuromorphic cameras is limited, and most of the methods rely on estimating physical priors and employ complex multi-step pipelines. In this work, we propose an end-to-end method for dense voxel 3D reconstruction using neuromorphic cameras that eliminates the need to estimate physical priors. Our method incorporates a novel event representation to enhance edge features, enabling the proposed feature-enhancement model to learn more effectively. Additionally, we introduced Optimal Binarization Threshold Selection Principle as a guideline for future related work, using the optimal reconstruction results achieved with threshold optimization as the benchmark. Our method achieves a 54.6% improvement in reconstruction accuracy compared to the baseline method.
comment: 6 pages, 15 figures, 5 tables, accepted by IEEE International Conference on Multimedia & Expo (ICME) 2025
♻ ☆ MMGDreamer: Mixed-Modality Graph for Geometry-Controllable 3D Indoor Scene Generation AAAI 2025
Controllable 3D scene generation has extensive applications in virtual reality and interior design, where the generated scenes should exhibit high levels of realism and controllability in terms of geometry. Scene graphs provide a suitable data representation that facilitates these applications. However, current graph-based methods for scene generation are constrained to text-based inputs and exhibit insufficient adaptability to flexible user inputs, hindering the ability to precisely control object geometry. To address this issue, we propose MMGDreamer, a dual-branch diffusion model for scene generation that incorporates a novel Mixed-Modality Graph, visual enhancement module, and relation predictor. The mixed-modality graph allows object nodes to integrate textual and visual modalities, with optional relationships between nodes. It enhances adaptability to flexible user inputs and enables meticulous control over the geometry of objects in the generated scenes. The visual enhancement module enriches the visual fidelity of text-only nodes by constructing visual representations using text embeddings. Furthermore, our relation predictor leverages node representations to infer absent relationships between nodes, resulting in more coherent scene layouts. Extensive experimental results demonstrate that MMGDreamer exhibits superior control of object geometry, achieving state-of-the-art scene generation performance. Project page: https://yangzhifeio.github.io/project/MMGDreamer.
comment: Accepted by AAAI 2025 Main Track
♻ ☆ Scaling Laws of Synthetic Data for Language Models
Large language models (LLMs) achieve strong performance across diverse tasks, largely driven by high-quality web data used in pre-training. However, recent studies indicate this data source is rapidly depleting. Synthetic data emerges as a promising alternative, but it remains unclear whether synthetic datasets exhibit predictable scalability comparable to raw pre-training data. In this work, we systematically investigate the scaling laws of synthetic data by introducing SynthLLM, a scalable framework that transforms pre-training corpora into diverse, high-quality synthetic datasets. Our approach achieves this by automatically extracting and recombining high-level concepts across multiple documents using a graph algorithm. Key findings from our extensive mathematical experiments on SynthLLM include: (1) SynthLLM generates synthetic data that reliably adheres to the rectified scaling law across various model sizes; (2) Performance improvements plateau near 300B tokens; and (3) Larger models approach optimal performance with fewer training tokens. For instance, an 8B model peaks at 1T tokens, while a 3B model requires 4T. Moreover, comparisons with existing synthetic data generation and augmentation methods demonstrate that SynthLLM achieves superior performance and scalability. Our findings highlight synthetic data as a scalable and reliable alternative to organic pre-training corpora, offering a viable path toward continued improvement in model performance.
comment: work in progress
♻ ☆ MARVEL-40M+: Multi-Level Visual Elaboration for High-Fidelity Text-to-3D Content Creation
Generating high-fidelity 3D content from text prompts remains a significant challenge in computer vision due to the limited size, diversity, and annotation depth of the existing datasets. To address this, we introduce MARVEL-40M+, an extensive dataset with 40 million text annotations for over 8.9 million 3D assets aggregated from seven major 3D datasets. Our contribution is a novel multi-stage annotation pipeline that integrates open-source pretrained multi-view VLMs and LLMs to automatically produce multi-level descriptions, ranging from detailed (150-200 words) to concise semantic tags (10-20 words). This structure supports both fine-grained 3D reconstruction and rapid prototyping. Furthermore, we incorporate human metadata from source datasets into our annotation pipeline to add domain-specific information in our annotation and reduce VLM hallucinations. Additionally, we develop MARVEL-FX3D, a two-stage text-to-3D pipeline. We fine-tune Stable Diffusion with our annotations and use a pretrained image-to-3D network to generate 3D textured meshes within 15s. Extensive evaluations show that MARVEL-40M+ significantly outperforms existing datasets in annotation quality and linguistic diversity, achieving win rates of 72.41% by GPT-4 and 73.40% by human evaluators. Project page is available at https://sankalpsinha-cmos.github.io/MARVEL/.
♻ ☆ DEIM: DETR with Improved Matching for Fast Convergence CVPR 2025
We introduce DEIM, an innovative and efficient training framework designed to accelerate convergence in real-time object detection with Transformer-based architectures (DETR). To mitigate the sparse supervision inherent in one-to-one (O2O) matching in DETR models, DEIM employs a Dense O2O matching strategy. This approach increases the number of positive samples per image by incorporating additional targets, using standard data augmentation techniques. While Dense O2O matching speeds up convergence, it also introduces numerous low-quality matches that could affect performance. To address this, we propose the Matchability-Aware Loss (MAL), a novel loss function that optimizes matches across various quality levels, enhancing the effectiveness of Dense O2O. Extensive experiments on the COCO dataset validate the efficacy of DEIM. When integrated with RT-DETR and D-FINE, it consistently boosts performance while reducing training time by 50%. Notably, paired with RT-DETRv2, DEIM achieves 53.2% AP in a single day of training on an NVIDIA 4090 GPU. Additionally, DEIM-trained real-time models outperform leading real-time object detectors, with DEIM-D-FINE-L and DEIM-D-FINE-X achieving 54.7% and 56.5% AP at 124 and 78 FPS on an NVIDIA T4 GPU, respectively, without the need for additional data. We believe DEIM sets a new baseline for advancements in real-time object detection. Our code and pre-trained models are available at https://github.com/ShihuaHuang95/DEIM.
comment: CVPR 2025
♻ ☆ FREIDA: A Framework for developing quantitative agent based models based on qualitative expert knowledge: an example of organised crime
Developing ABMs of organized crime networks supports law enforcement strategies but is often limited by scarce quantitative data. This challenge extends to other psychosocial contexts like mental health and social systems. While qualitative data from reports and interviews is more accessible, current ABM methodologies struggle to integrate both data types effectively. To address this, we propose FREIDA, a mixed-methods framework that combines qualitative and quantitative data to develop, train, and validate ABMs in data-sparse contexts. FREIDA's four-phase process includes data acquisition, conceptual modeling, computational implementation, and model assessment. Using Thematic Content Analysis (TCA), Expected System Behaviors (ESBs) are translated into Training Statements (TS) for calibration and Validation Statements (VS) for assessment. Iterative sensitivity analysis and uncertainty quantification refine the model's accuracy. We apply FREIDA to a case study of the Netherlands cocaine network, producing the Criminal Cocaine Replacement Model (CCRM) to simulate kingpin removal dynamics. FREIDA enables robust ABM development with limited data, aiding law enforcement decisions and resource allocation.
comment: 32 pages, 12 figures, 14 tables, Appendix I-IV
♻ ☆ Retro-li: Small-Scale Retrieval Augmented Generation Supporting Noisy Similarity Searches and Domain Shift Generalization
The retrieval augmented generation (RAG) system such as Retro has been shown to improve language modeling capabilities and reduce toxicity and hallucinations by retrieving from a database of non-parametric memory containing trillions of entries. We introduce Retro-li that shows retrieval can also help using a small-scale database, but it demands more accurate and better neighbors when searching in a smaller hence sparser non-parametric memory. This can be met by using a proper semantic similarity search. We further propose adding a regularization to the non-parametric memory for the first time: it significantly reduces perplexity when the neighbor search operations are noisy during inference, and it improves generalization when a domain shift occurs. We also show that Retro-li's non-parametric memory can potentially be implemented on analog in-memory computing hardware, exhibiting O(1) search time while causing noise in retrieving neighbors, with minimal (<1%) performance loss. Our code is available at: https://github.com/IBM/Retrieval-Enhanced-Transformer-Little.
♻ ☆ Two pathways to resolve relational inconsistencies
When individuals encounter observations that violate their expectations, when will they adjust their expectations and when will they maintain them despite these observations? For example, when individuals expect objects of type A to be smaller than objects B, but observe the opposite, when will they adjust their expectation about the relationship between the two objects (to A being larger than B)? Naively, one would predict that the larger the violation, the greater the adaptation. However, experiments reveal that when violations are extreme, individuals are more likely to hold on to their prior expectations rather than adjust them. To address this puzzle, we tested the adaptation of artificial neural networks (ANNs) capable of relational learning and found a similar phenomenon: Standard learning dynamics dictates that small violations would lead to adjustments of expected relations while larger ones would be resolved using a different mechanism -- a change in object representation that bypasses the need for adaptation of the relational expectations. These results suggest that the experimentally-observed stability of prior expectations when facing large expectation violations is a natural consequence of learning dynamics and does not require any additional mechanisms. We conclude by discussing the effect of intermediate adaptation steps on this stability.
♻ ☆ Agentic AI Software Engineer: Programming with Trust
Large Language Models (LLMs) have shown surprising proficiency in generating code snippets, promising to automate large parts of software engineering via artificial intelligence (AI). We argue that successfully deploying AI software engineers requires a level of trust equal to or even greater than the trust established by human-driven software engineering practices. The recent trend toward LLM agents offers a path toward integrating the power of LLMs to create new code with the power of analysis tools to increase trust in the code. This opinion piece comments on whether LLM agents could dominate software engineering workflows in the future and whether the focus of programming will shift from programming at scale to programming with trust.
comment: 5 pages
♻ ☆ DiTCtrl: Exploring Attention Control in Multi-Modal Diffusion Transformer for Tuning-Free Multi-Prompt Longer Video Generation CVPR 2025
Sora-like video generation models have achieved remarkable progress with a Multi-Modal Diffusion Transformer MM-DiT architecture. However, the current video generation models predominantly focus on single-prompt, struggling to generate coherent scenes with multiple sequential prompts that better reflect real-world dynamic scenarios. While some pioneering works have explored multi-prompt video generation, they face significant challenges including strict training data requirements, weak prompt following, and unnatural transitions. To address these problems, we propose DiTCtrl, a training-free multi-prompt video generation method under MM-DiT architectures for the first time. Our key idea is to take the multi-prompt video generation task as temporal video editing with smooth transitions. To achieve this goal, we first analyze MM-DiT's attention mechanism, finding that the 3D full attention behaves similarly to that of the cross/self-attention blocks in the UNet-like diffusion models, enabling mask-guided precise semantic control across different prompts with attention sharing for multi-prompt video generation. Based on our careful design, the video generated by DiTCtrl achieves smooth transitions and consistent object motion given multiple sequential prompts without additional training. Besides, we also present MPVBench, a new benchmark specially designed for multi-prompt video generation to evaluate the performance of multi-prompt generation. Extensive experiments demonstrate that our method achieves state-of-the-art performance without additional training.
comment: CVPR 2025; 21 pages, 23 figures, Project page: https://onevfall.github.io/project_page/ditctrl ; GitHub repository: https://github.com/TencentARC/DiTCtrl
♻ ☆ Three Kinds of AI Ethics
There is an overwhelming abundance of works in AI Ethics. This growth is chaotic because of how sudden it is, its volume, and its multidisciplinary nature. This makes difficult to keep track of debates, and to systematically characterize goals, research questions, methods, and expertise required by AI ethicists. In this article, I show that the relation between AI and ethics can be characterized in at least three ways, which correspond to three well-represented kinds of AI ethics: ethics and AI; ethics in AI; ethics of AI. I elucidate the features of these three kinds of AI Ethics, characterize their research questions, and identify the kind of expertise that each kind needs. I also show how certain criticisms to AI ethics are misplaced, as being done from the point of view of one kind of AI ethics, to another kind with different goals. All in all, this work sheds light on the nature of AI ethics, and sets the groundwork for more informed discussions about the scope, methods, and training of AI ethicists.
comment: 16 pages, two figures
♻ ☆ Oasis: One Image is All You Need for Multimodal Instruction Data Synthesis
The success of multi-modal large language models (MLLMs) has been largely attributed to the large-scale training data. However, the training data of many MLLMs is unavailable due to privacy concerns. The expensive and labor-intensive process of collecting multi-modal data further exacerbates the problem. Is it possible to synthesize multi-modal training data automatically without compromising diversity and quality? In this paper, we propose a new method, Oasis, to synthesize high-quality multi-modal data with only images. Oasis breaks through traditional methods by prompting only images to the MLLMs, thus extending the data diversity by a large margin. Our method features a delicate quality control method which ensures the data quality. We collected over 500k data and conducted incremental experiments on LLaVA-NeXT. Extensive experiments demonstrate that our method can significantly improve the performance of MLLMs. The image-based synthesis also allows us to focus on the specific-domain ability of MLLMs. Code and dataset are publicly available at https://github.com/Letian2003/MM_INF.
♻ ☆ ManiCM: Real-time 3D Diffusion Policy via Consistency Model for Robotic Manipulation
Diffusion models have been verified to be effective in generating complex distributions from natural images to motion trajectories. Recent diffusion-based methods show impressive performance in 3D robotic manipulation tasks, whereas they suffer from severe runtime inefficiency due to multiple denoising steps, especially with high-dimensional observations. To this end, we propose a real-time robotic manipulation model named ManiCM that imposes the consistency constraint to the diffusion process, so that the model can generate robot actions in only one-step inference. Specifically, we formulate a consistent diffusion process in the robot action space conditioned on the point cloud input, where the original action is required to be directly denoised from any point along the ODE trajectory. To model this process, we design a consistency distillation technique to predict the action sample directly instead of predicting the noise within the vision community for fast convergence in the low-dimensional action manifold. We evaluate ManiCM on 31 robotic manipulation tasks from Adroit and Metaworld, and the results demonstrate that our approach accelerates the state-of-the-art method by 10 times in average inference speed while maintaining competitive average success rate.
comment: https://manicm-fast.github.io/
♻ ☆ LaMOuR: Leveraging Language Models for Out-of-Distribution Recovery in Reinforcement Learning
Deep Reinforcement Learning (DRL) has demonstrated strong performance in robotic control but remains susceptible to out-of-distribution (OOD) states, often resulting in unreliable actions and task failure. While previous methods have focused on minimizing or preventing OOD occurrences, they largely neglect recovery once an agent encounters such states. Although the latest research has attempted to address this by guiding agents back to in-distribution states, their reliance on uncertainty estimation hinders scalability in complex environments. To overcome this limitation, we introduce Language Models for Out-of-Distribution Recovery (LaMOuR), which enables recovery learning without relying on uncertainty estimation. LaMOuR generates dense reward codes that guide the agent back to a state where it can successfully perform its original task, leveraging the capabilities of LVLMs in image description, logical reasoning, and code generation. Experimental results show that LaMOuR substantially enhances recovery efficiency across diverse locomotion tasks and even generalizes effectively to complex environments, including humanoid locomotion and mobile manipulation, where existing methods struggle. The code and supplementary materials are available at https://lamour-rl.github.io/.
comment: 14 pages, 17 figures
♻ ☆ Lemur: Log Parsing with Entropy Sampling and Chain-of-Thought Merging
Logs produced by extensive software systems are integral to monitoring system behaviors. Advanced log analysis facilitates the detection, alerting, and diagnosis of system faults. Log parsing, which entails transforming raw log messages into structured templates, constitutes a critical phase in the automation of log analytics. Existing log parsers fail to identify the correct templates due to reliance on human-made rules. Besides, these methods focus on statistical features while ignoring semantic information in log messages. To address these challenges, we introduce a cutting-edge \textbf{L}og parsing framework with \textbf{E}ntropy sampling and chain-of-thought \textbf{M}erging (\model{}). Specifically, to discard the tedious manual rules, we propose a novel sampling method inspired by information entropy, which efficiently clusters typical logs. Furthermore, to enhance the merging of log templates, we design a chain-of-thought method for large language models (LLMs). LLMs exhibit exceptional semantic comprehension and deftly distinguish between parameters and invariant tokens. We have conducted experiments on large-scale public datasets. Extensive evaluation demonstrates that \model{} achieves state-of-the-art performance and impressive efficiency. The Code is available at https://github.com/zwpride/lemur.
♻ ☆ ARFlow: Human Action-Reaction Flow Matching with Physical Guidance
Human action-reaction synthesis, a fundamental challenge in modeling causal human interactions, plays a critical role in applications ranging from virtual reality to social robotics. While diffusion-based models have demonstrated promising performance, they exhibit two key limitations for interaction synthesis: reliance on complex noise-to-reaction generators with intricate conditional mechanisms, and frequent physical violations in generated motions. To address these issues, we propose Action-Reaction Flow Matching (ARFlow), a novel framework that establishes direct action-to-reaction mappings, eliminating the need for complex conditional mechanisms. Our approach introduces two key innovations: an x1-prediction method that directly outputs human motions instead of velocity fields, enabling explicit constraint enforcement; and a training-free, gradient-based physical guidance mechanism that effectively prevents body penetration artifacts during sampling. Extensive experiments on NTU120 and Chi3D datasets demonstrate that ARFlow not only outperforms existing methods in terms of Fr\'echet Inception Distance and motion diversity but also significantly reduces body collisions, as measured by our new Intersection Volume and Intersection Frequency metrics.
comment: Project Page: https://arflow2025.github.io/
♻ ☆ Contextual Similarity Distillation: Ensemble Uncertainties with a Single Model
Uncertainty quantification is a critical aspect of reinforcement learning and deep learning, with numerous applications ranging from efficient exploration and stable offline reinforcement learning to outlier detection in medical diagnostics. The scale of modern neural networks, however, complicates the use of many theoretically well-motivated approaches such as full Bayesian inference. Approximate methods like deep ensembles can provide reliable uncertainty estimates but still remain computationally expensive. In this work, we propose contextual similarity distillation, a novel approach that explicitly estimates the variance of an ensemble of deep neural networks with a single model, without ever learning or evaluating such an ensemble in the first place. Our method builds on the predictable learning dynamics of wide neural networks, governed by the neural tangent kernel, to derive an efficient approximation of the predictive variance of an infinite ensemble. Specifically, we reinterpret the computation of ensemble variance as a supervised regression problem with kernel similarities as regression targets. The resulting model can estimate predictive variance at inference time with a single forward pass, and can make use of unlabeled target-domain data or data augmentations to refine its uncertainty estimates. We empirically validate our method across a variety of out-of-distribution detection benchmarks and sparse-reward reinforcement learning environments. We find that our single-model method performs competitively and sometimes superior to ensemble-based baselines and serves as a reliable signal for efficient exploration. These results, we believe, position contextual similarity distillation as a principled and scalable alternative for uncertainty quantification in reinforcement learning and general deep learning.
♻ ☆ MetaDE: Evolving Differential Evolution by Differential Evolution
As a cornerstone in the Evolutionary Computation (EC) domain, Differential Evolution (DE) is known for its simplicity and effectiveness in handling challenging black-box optimization problems. While the advantages of DE are well-recognized, achieving peak performance heavily depends on its hyperparameters such as the mutation factor, crossover probability, and the selection of specific DE strategies. Traditional approaches to this hyperparameter dilemma have leaned towards parameter tuning or adaptive mechanisms. However, identifying the optimal settings tailored for specific problems remains a persistent challenge. In response, we introduce MetaDE, an approach that evolves DE's intrinsic hyperparameters and strategies using DE itself at a meta-level. A pivotal aspect of MetaDE is a specialized parameterization technique, which endows it with the capability to dynamically modify DE's parameters and strategies throughout the evolutionary process. To augment computational efficiency, MetaDE incorporates a design that leverages parallel processing through a GPU-accelerated computing framework. Within such a framework, DE is not just a solver but also an optimizer for its own configurations, thus streamlining the process of hyperparameter optimization and problem-solving into a cohesive and automated workflow. Extensive evaluations on the CEC2022 benchmark suite demonstrate MetaDE's promising performance. Moreover, when applied to robot control via evolutionary reinforcement learning, MetaDE also demonstrates promising performance. The source code of MetaDE is publicly accessible at: https://github.com/EMI-Group/metade.
comment: Accepted by IEEE TEVC
♻ ☆ Socratic Planner: Self-QA-Based Zero-Shot Planning for Embodied Instruction Following ICRA 2025
Embodied Instruction Following (EIF) is the task of executing natural language instructions by navigating and interacting with objects in interactive environments. A key challenge in EIF is compositional task planning, typically addressed through supervised learning or few-shot in-context learning with labeled data. To this end, we introduce the Socratic Planner, a self-QA-based zero-shot planning method that infers an appropriate plan without any further training. The Socratic Planner first facilitates self-questioning and answering by the Large Language Model (LLM), which in turn helps generate a sequence of subgoals. While executing the subgoals, an embodied agent may encounter unexpected situations, such as unforeseen obstacles. The Socratic Planner then adjusts plans based on dense visual feedback through a visually-grounded re-planning mechanism. Experiments demonstrate the effectiveness of the Socratic Planner, outperforming current state-of-the-art planning models on the ALFRED benchmark across all metrics, particularly excelling in long-horizon tasks that demand complex inference. We further demonstrate its real-world applicability through deployment on a physical robot for long-horizon tasks.
comment: 8 pages, 6 figures, published to ICRA 2025
♻ ☆ BRIGHT: A Realistic and Challenging Benchmark for Reasoning-Intensive Retrieval
Existing retrieval benchmarks primarily consist of information-seeking queries (e.g., aggregated questions from search engines) where keyword or semantic-based retrieval is usually sufficient. However, many complex real-world queries require in-depth reasoning to identify relevant documents that go beyond surface form matching. For example, finding documentation for a coding question requires understanding the logic and syntax of the functions involved. To better benchmark retrieval on such challenging queries, we introduce BRIGHT, the first text retrieval benchmark that requires intensive reasoning to retrieve relevant documents. Our dataset consists of 1,384 real-world queries spanning diverse domains, such as economics, psychology, mathematics, and coding. These queries are drawn from naturally occurring and carefully curated human data. Extensive evaluation reveals that even state-of-the-art retrieval models perform poorly on BRIGHT. The leading model on the MTEB leaderboard (Muennighoff et al., 2023) SFR-Embedding-Mistral (Meng et al., 2024), which achieves a score of 59.0 nDCG@10,1 produces a score of nDCG@10 of 18.3 on BRIGHT. We show that incorporating explicit reasoning about the query improves retrieval performance by up to 12.2 points. Moreover, incorporating retrieved documents from the top-performing retriever boosts question-answering performance. We believe that BRIGHT paves the way for future research on retrieval systems in more realistic and challenging settings.
comment: 51 pages
♻ ☆ RoboSpatial: Teaching Spatial Understanding to 2D and 3D Vision-Language Models for Robotics CVPR 2025
Spatial understanding is a crucial capability that enables robots to perceive their surroundings, reason about their environment, and interact with it meaningfully. In modern robotics, these capabilities are increasingly provided by vision-language models. However, these models face significant challenges in spatial reasoning tasks, as their training data are based on general-purpose image datasets that often lack sophisticated spatial understanding. For example, datasets frequently do not capture reference frame comprehension, yet effective spatial reasoning requires understanding whether to reason from ego-, world-, or object-centric perspectives. To address this issue, we introduce RoboSpatial, a large-scale dataset for spatial understanding in robotics. It consists of real indoor and tabletop scenes, captured as 3D scans and egocentric images, and annotated with rich spatial information relevant to robotics. The dataset includes 1M images, 5k 3D scans, and 3M annotated spatial relationships, and the pairing of 2D egocentric images with 3D scans makes it both 2D- and 3D- ready. Our experiments show that models trained with RoboSpatial outperform baselines on downstream tasks such as spatial affordance prediction, spatial relationship prediction, and robot manipulation.
comment: CVPR 2025
♻ ☆ TopV-Nav: Unlocking the Top-View Spatial Reasoning Potential of MLLM for Zero-shot Object Navigation
The Zero-Shot Object Navigation (ZSON) task requires embodied agents to find a previously unseen object by navigating in unfamiliar environments. Such a goal-oriented exploration heavily relies on the ability to perceive, understand, and reason based on the spatial information of the environment. However, current LLM-based approaches convert visual observations to language descriptions and reason in the linguistic space, leading to the loss of spatial information. In this paper, we introduce TopV-Nav, an MLLM-based method that directly reasons on the top-view map with sufficient spatial information. To fully unlock the MLLM's spatial reasoning potential in top-view perspective, we propose the Adaptive Visual Prompt Generation (AVPG) method to adaptively construct semantically-rich top-view map. It enables the agent to directly utilize spatial information contained in the top-view map to conduct thorough reasoning. Besides, we design a Dynamic Map Scaling (DMS) mechanism to dynamically zoom top-view map at preferred scales, enhancing local fine-grained reasoning. Additionally, we devise a Potential Target Driven (PTD) mechanism to predict and to utilize target locations, facilitating global and human-like exploration. Experiments on MP3D and HM3D datasets demonstrate the superiority of our TopV-Nav.
comment: 10 pages
♻ ☆ Fine-Grained Domain Generalization with Feature Structuralization
Fine-grained domain generalization (FGDG) is a more challenging task than traditional DG tasks due to its small inter-class variations and relatively large intra-class disparities. When domain distribution changes, the vulnerability of subtle features leads to a severe deterioration in model performance. Nevertheless, humans inherently demonstrate the capacity for generalizing to out-of-distribution data, leveraging structured multi-granularity knowledge that emerges from discerning the commonality and specificity within categories. Likewise, we propose a Feature Structuralized Domain Generalization (FSDG) model, wherein features experience structuralization into common, specific, and confounding segments, harmoniously aligned with their relevant semantic concepts, to elevate performance in FGDG. Specifically, feature structuralization (FS) is accomplished through joint optimization of five constraints: a decorrelation function applied to disentangled segments, three constraints ensuring common feature consistency and specific feature distinctiveness, and a prediction calibration term. By imposing these stipulations, FSDG is prompted to disentangle and align features based on multi-granularity knowledge, facilitating robust subtle distinctions among categories. Extensive experimentation on three benchmarks consistently validates the superiority of FSDG over state-of-the-art counterparts, with an average improvement of 6.2% in FGDG performance. Beyond that, the explainability analysis on explicit concept matching intensity between the shared concepts among categories and the model channels, along with experiments on various mainstream model architectures, substantiates the validity of FS.
♻ ☆ Making AI Less "Thirsty": Uncovering and Addressing the Secret Water Footprint of AI Models
The growing carbon footprint of artificial intelligence (AI) has been undergoing public scrutiny. Nonetheless, the equally important water (withdrawal and consumption) footprint of AI has largely remained under the radar. For example, training the GPT-3 language model in Microsoft's state-of-the-art U.S. data centers can directly evaporate 700,000 liters of clean freshwater, but such information has been kept a secret. More critically, the global AI demand is projected to account for 4.2-6.6 billion cubic meters of water withdrawal in 2027, which is more than the total annual water withdrawal of 4-6 Denmark or half of the United Kingdom. This is concerning, as freshwater scarcity has become one of the most pressing challenges. To respond to the global water challenges, AI can, and also must, take social responsibility and lead by example by addressing its own water footprint. In this paper, we provide a principled methodology to estimate the water footprint of AI, and also discuss the unique spatial-temporal diversities of AI's runtime water efficiency. Finally, we highlight the necessity of holistically addressing water footprint along with carbon footprint to enable truly sustainable AI.
comment: Accepted by Communications of the ACM. Source codes available at: https://github.com/Ren-Research/Making-AI-Less-Thirsty
♻ ☆ General-purpose Clothes Manipulation with Semantic Keypoints ICRA 2025
Clothes manipulation is a critical capability for household robots; yet, existing methods are often confined to specific tasks, such as folding or flattening, due to the complex high-dimensional geometry of deformable fabric. This paper presents CLothes mAnipulation with Semantic keyPoints (CLASP) for general-purpose clothes manipulation, which enables the robot to perform diverse manipulation tasks over different types of clothes. The key idea of CLASP is semantic keypoints -- e.g., "right shoulder", "left sleeve", etc. -- a sparse spatial-semantic representation that is salient for both perception and action. Semantic keypoints of clothes can be effectively extracted from depth images and are sufficient to represent a broad range of clothes manipulation policies. CLASP leverages semantic keypoints to bridge LLM-powered task planning and low-level action execution in a two-level hierarchy. Extensive simulation experiments show that CLASP outperforms baseline methods across diverse clothes types in both seen and unseen tasks. Further, experiments with a Kinova dual-arm system on four distinct tasks -- folding, flattening, hanging, and placing -- confirm CLASP's performance on a real robot.
comment: accepted by IEEE International Conference on Robotics and Automation (ICRA 2025)
♻ ☆ Graph-Level Label-Only Membership Inference Attack against Graph Neural Networks
Graph neural networks (GNNs) are widely used for graph-structured data but are vulnerable to membership inference attacks (MIAs) in graph classification tasks, which determine if a graph was part of the training dataset, potentially causing data leakage. Existing MIAs rely on prediction probability vectors, but they become ineffective when only prediction labels are available. We propose a Graph-level Label-Only Membership Inference Attack (GLO-MIA), which is based on the intuition that the target model's predictions on training data are more stable than those on testing data. GLO-MIA generates a set of perturbed graphs for target graph by adding perturbations to its effective features and queries the target model with the perturbed graphs to get their prediction labels, which are then used to calculate robustness score of the target graph. Finally, by comparing the robustness score with a predefined threshold, the membership of the target graph can be inferred correctly with high probability. Our evaluation on three datasets and four GNN models shows that GLO-MIA achieves an attack accuracy of up to 0.825, outperforming baseline work by 8.5% and closely matching the performance of probability-based MIAs, even with only prediction labels.
♻ ☆ DAWN: Dynamic Frame Avatar with Non-autoregressive Diffusion Framework for Talking Head Video Generation
Talking head generation intends to produce vivid and realistic talking head videos from a single portrait and speech audio clip. Although significant progress has been made in diffusion-based talking head generation, almost all methods rely on autoregressive strategies, which suffer from limited context utilization beyond the current generation step, error accumulation, and slower generation speed. To address these challenges, we present DAWN (Dynamic frame Avatar With Non-autoregressive diffusion), a framework that enables all-at-once generation of dynamic-length video sequences. Specifically, it consists of two main components: (1) audio-driven holistic facial dynamics generation in the latent motion space, and (2) audio-driven head pose and blink generation. Extensive experiments demonstrate that our method generates authentic and vivid videos with precise lip motions, and natural pose/blink movements. Additionally, with a high generation speed, DAWN possesses strong extrapolation capabilities, ensuring the stable production of high-quality long videos. These results highlight the considerable promise and potential impact of DAWN in the field of talking head video generation. Furthermore, we hope that DAWN sparks further exploration of non-autoregressive approaches in diffusion models. Our code will be publicly available at https://github.com/Hanbo-Cheng/DAWN-pytorch.
♻ ☆ VesselSAM: Leveraging SAM for Aortic Vessel Segmentation with LoRA and Atrous Attention
Medical image segmentation is crucial for clinical diagnosis and treatment planning, especially when dealing with complex anatomical structures such as vessels. However, accurately segmenting vessels remains challenging due to their small size, intricate edge structures, and susceptibility to artifacts and imaging noise. In this work, we propose VesselSAM, an enhanced version of the Segment Anything Model (SAM), specifically tailored for aortic vessel segmentation. VesselSAM incorporates AtrousLoRA, a novel module integrating Atrous Attention and Low-Rank Adaptation (LoRA), to enhance segmentation performance. Atrous Attention enables the model to capture multi-scale contextual information, preserving both fine-grained local details and broader global context. Additionally, LoRA facilitates efficient fine-tuning of the frozen SAM image encoder, reducing the number of trainable parameters and thereby enhancing computational efficiency. We evaluate VesselSAM using two challenging datasets: the Aortic Vessel Tree (AVT) dataset and the Type-B Aortic Dissection (TBAD) dataset. VesselSAM achieves state-of-the-art performance, attaining DSC scores of 93.50\%, 93.25\%, 93.02\%, and 93.26\% across multi-center datasets. Our results demonstrate that VesselSAM delivers high segmentation accuracy while significantly reducing computational overhead compared to existing large-scale models. This development paves the way for enhanced AI-based aortic vessel segmentation in clinical environments. The code and models will be released at https://github.com/Adnan-CAS/AtrousLora.
comment: Work in progress
♻ ☆ Bonsai: Gradient-free Graph Condensation for Node Classification
Graph condensation has emerged as a promising avenue to enable scalable training of GNNs by compressing the training dataset while preserving essential graph characteristics. Our study uncovers significant shortcomings in current graph condensation techniques. First, the majority of the algorithms paradoxically require training on the full dataset to perform condensation. Second, due to their gradient-emulating approach, these methods require fresh condensation for any change in hyperparameters or GNN architecture, limiting their flexibility and reusability. Finally, they fail to achieve substantial size reduction due to synthesizing fully-connected, edge-weighted graphs. To address these challenges, we present Bonsai, a novel graph condensation method empowered by the observation that \textit{computation trees} form the fundamental processing units of message-passing GNNs. Bonsai condenses datasets by encoding a careful selection of \textit{exemplar} trees that maximize the representation of all computation trees in the training set. This unique approach imparts Bonsai as the first linear-time, model-agnostic graph condensation algorithm for node classification that outperforms existing baselines across $7$ real-world datasets on accuracy, while being $22$ times faster on average. Bonsai is grounded in rigorous mathematical guarantees on the adopted approximation strategies making it robust to GNN architectures, datasets, and parameters.
♻ ☆ Human Motion Instruction Tuning CVPR 2025
This paper presents LLaMo (Large Language and Human Motion Assistant), a multimodal framework for human motion instruction tuning. In contrast to conventional instruction-tuning approaches that convert non-linguistic inputs, such as video or motion sequences, into language tokens, LLaMo retains motion in its native form for instruction tuning. This method preserves motion-specific details that are often diminished in tokenization, thereby improving the model's ability to interpret complex human behaviors. By processing both video and motion data alongside textual inputs, LLaMo enables a flexible, human-centric analysis. Experimental evaluations across high-complexity domains, including human behaviors and professional activities, indicate that LLaMo effectively captures domain-specific knowledge, enhancing comprehension and prediction in motion-intensive scenarios. We hope LLaMo offers a foundation for future multimodal AI systems with broad applications, from sports analytics to behavioral prediction. Our code and models are available on the project website: https://github.com/ILGLJ/LLaMo.
comment: Accepted by CVPR 2025
♻ ☆ Towards Visual Discrimination and Reasoning of Real-World Physical Dynamics: Physics-Grounded Anomaly Detection CVPR25
Humans detect real-world object anomalies by perceiving, interacting, and reasoning based on object-conditioned physical knowledge. The long-term goal of Industrial Anomaly Detection (IAD) is to enable machines to autonomously replicate this skill. However, current IAD algorithms are largely developed and tested on static, semantically simple datasets, which diverge from real-world scenarios where physical understanding and reasoning are essential. To bridge this gap, we introduce the Physics Anomaly Detection (Phys-AD) dataset, the first large-scale, real-world, physics-grounded video dataset for industrial anomaly detection. Collected using a real robot arm and motor, Phys-AD provides a diverse set of dynamic, semantically rich scenarios. The dataset includes more than 6400 videos across 22 real-world object categories, interacting with robot arms and motors, and exhibits 47 types of anomalies. Anomaly detection in Phys-AD requires visual reasoning, combining both physical knowledge and video content to determine object abnormality. We benchmark state-of-the-art anomaly detection methods under three settings: unsupervised AD, weakly-supervised AD, and video-understanding AD, highlighting their limitations in handling physics-grounded anomalies. Additionally, we introduce the Physics Anomaly Explanation (PAEval) metric, designed to assess the ability of visual-language foundation models to not only detect anomalies but also provide accurate explanations for their underlying physical causes. Our project is available at https://guyao2023.github.io/Phys-AD/.
comment: Accepted by CVPR25
♻ ☆ Multi-agent Application System in Office Collaboration Scenarios
This paper introduces a multi-agent application system designed to enhance office collaboration efficiency and work quality. The system integrates artificial intelligence, machine learning, and natural language processing technologies, achieving functionalities such as task allocation, progress monitoring, and information sharing. The agents within the system are capable of providing personalized collaboration support based on team members' needs and incorporate data analysis tools to improve decision-making quality. The paper also proposes an intelligent agent architecture that separates Plan and Solver, and through techniques such as multi-turn query rewriting and business tool retrieval, it enhances the agent's multi-intent and multi-turn dialogue capabilities. Furthermore, the paper details the design of tools and multi-turn dialogue in the context of office collaboration scenarios, and validates the system's effectiveness through experiments and evaluations. Ultimately, the system has demonstrated outstanding performance in real business applications, particularly in query understanding, task planning, and tool calling. Looking forward, the system is expected to play a more significant role in addressing complex interaction issues within dynamic environments and large-scale multi-agent systems.
comment: Technical report
♻ ☆ TransPlace: Transferable Circuit Global Placement via Graph Neural Network KDD 2025
Global placement, a critical step in designing the physical layout of computer chips, is essential to optimize chip performance. Prior global placement methods optimize each circuit design individually from scratch. Their neglect of transferable knowledge limits solution efficiency and chip performance as circuit complexity drastically increases. This study presents TransPlace, a global placement framework that learns to place millions of mixed-size cells in continuous space. TransPlace introduces i) Netlist Graph to efficiently model netlist topology, ii) Cell-flow and relative position encoding to learn SE(2)-invariant representation, iii) a tailored graph neural network architecture for informed parameterization of placement knowledge, and iv) a two-stage strategy for coarse-to-fine placement. Compared to state-of-the-art placement methods, TransPlace-trained on a few high-quality placements-can place unseen circuits with 1.2x speedup while reducing congestion by 30%, timing by 9%, and wirelength by 5%.
comment: Accepted at KDD 2025
♻ ☆ Forest-of-Thought: Scaling Test-Time Compute for Enhancing LLM Reasoning
Large Language Models (LLMs) have demonstrated remarkable abilities across various language tasks, but solving complex reasoning problems remains a significant challenge. While existing methods, such as Chain-of-Thought (CoT) and Tree-of-Thought (ToT), enhance reasoning by decomposing problems or structuring prompts, they typically perform a single pass of reasoning and may fail to revisit flawed paths, compromising accuracy. To address this limitation, we propose a novel reasoning framework called Forest-of-Thought (FoT), which integrates multiple reasoning trees to leverage collective decision-making for solving complex logical problems. FoT employs sparse activation strategies to select the most relevant reasoning paths, improving both efficiency and accuracy. Additionally, we introduce a dynamic self-correction strategy that enables real-time error correction, along with consensus-guided decision-making strategies to optimize both correctness and computational resources. Experimental results demonstrate that the FoT framework, combined with these strategies, significantly enhances the reasoning capabilities of LLMs, enabling them to solve complex tasks with greater precision and efficiency. Code will be available at https://github.com/iamhankai/Forest-of-Thought.
comment: Preprint
Inference-Time Policy Steering through Human Interactions ICRA 2025
Generative policies trained with human demonstrations can autonomously accomplish multimodal, long-horizon tasks. However, during inference, humans are often removed from the policy execution loop, limiting the ability to guide a pre-trained policy towards a specific sub-goal or trajectory shape among multiple predictions. Naive human intervention may inadvertently exacerbate distribution shift, leading to constraint violations or execution failures. To better align policy output with human intent without inducing out-of-distribution errors, we propose an Inference-Time Policy Steering (ITPS) framework that leverages human interactions to bias the generative sampling process, rather than fine-tuning the policy on interaction data. We evaluate ITPS across three simulated and real-world benchmarks, testing three forms of human interaction and associated alignment distance metrics. Among six sampling strategies, our proposed stochastic sampling with diffusion policy achieves the best trade-off between alignment and distribution shift. Videos are available at https://yanweiw.github.io/itps/.
comment: ICRA 2025
♻ ☆ Uni$\textbf{F}^2$ace: Fine-grained Face Understanding and Generation with Unified Multimodal Models
Unified multimodal models (UMMs) have emerged as a powerful paradigm in foundational computer vision research, demonstrating significant potential in both image understanding and generation. However, existing research in the face domain primarily focuses on $\textbf{coarse}$ facial attribute understanding, with limited capacity to handle $\textbf{fine-grained}$ facial attributes and without addressing generation capabilities. To overcome these limitations, we propose Uni$\textbf{F}^2$ace, the first UMM tailored specifically for fine-grained face understanding and generation. In general, we train Uni$\textbf{F}^2$ace on a self-constructed, specialized dataset utilizing two mutually beneficial diffusion techniques and a two-level mixture-of-experts architecture. Specifically, we first build a large-scale facial dataset, Uni$\textbf{F}^2$ace-130K, which contains 130K image-text pairs with one million question-answering pairs that span a wide range of facial attributes. Second, we establish a theoretical connection between discrete diffusion score matching and masked generative models, optimizing both evidence lower bounds simultaneously, which significantly improves the model's ability to synthesize facial details. Finally, we introduce both token-level and sequence-level mixture-of-experts, enabling efficient fine-grained representation learning for both understanding and generation tasks. Extensive experiments on Uni$\textbf{F}^2$ace-130K demonstrate that Uni$\textbf{F}^2$ace outperforms existing UMMs and generative models, achieving superior performance across both understanding and generation tasks.
♻ ☆ Medical-GAT: Cancer Document Classification Leveraging Graph-Based Residual Network for Scenarios with Limited Data
Accurate classification of cancer-related medical abstracts is crucial for healthcare management and research. However, obtaining large, labeled datasets in the medical domain is challenging due to privacy concerns and the complexity of clinical data. This scarcity of annotated data impedes the development of effective machine learning models for cancer document classification. To address this challenge, we present a curated dataset of 1,874 biomedical abstracts, categorized into thyroid cancer, colon cancer, lung cancer, and generic topics. Our research focuses on leveraging this dataset to improve classification performance, particularly in data-scarce scenarios. We introduce a Residual Graph Attention Network (R-GAT) with multiple graph attention layers that capture the semantic information and structural relationships within cancer-related documents. Our R-GAT model is compared with various techniques, including transformer-based models such as Bidirectional Encoder Representations from Transformers (BERT), RoBERTa, and domain-specific models like BioBERT and Bio+ClinicalBERT. We also evaluated deep learning models (CNNs, LSTMs) and traditional machine learning models (Logistic Regression, SVM). Additionally, we explore ensemble approaches that combine deep learning models to enhance classification. Various feature extraction methods are assessed, including Term Frequency-Inverse Document Frequency (TF-IDF) with unigrams and bigrams, Word2Vec, and tokenizers from BERT and RoBERTa. The R-GAT model outperforms other techniques, achieving precision, recall, and F1 scores of 0.99, 0.97, and 0.98 for thyroid cancer; 0.96, 0.94, and 0.95 for colon cancer; 0.96, 0.99, and 0.97 for lung cancer; and 0.95, 0.96, and 0.95 for generic topics.
♻ ☆ CamSAM2: Segment Anything Accurately in Camouflaged Videos
Video camouflaged object segmentation (VCOS), aiming at segmenting camouflaged objects that seamlessly blend into their environment, is a fundamental vision task with various real-world applications. With the release of SAM2, video segmentation has witnessed significant progress. However, SAM2's capability of segmenting camouflaged videos is suboptimal, especially when given simple prompts such as point and box. To address the problem, we propose Camouflaged SAM2 (CamSAM2), which enhances SAM2's ability to handle camouflaged scenes without modifying SAM2's parameters. Specifically, we introduce a decamouflaged token to provide the flexibility of feature adjustment for VCOS. To make full use of fine-grained and high-resolution features from the current frame and previous frames, we propose implicit object-aware fusion (IOF) and explicit object-aware fusion (EOF) modules, respectively. Object prototype generation (OPG) is introduced to abstract and memorize object prototypes with informative details using high-quality features from previous frames. Extensive experiments are conducted to validate the effectiveness of our approach. While CamSAM2 only adds negligible learnable parameters to SAM2, it substantially outperforms SAM2 on three VCOS datasets, especially achieving 12.2 mDice gains with click prompt on MoCA-Mask and 19.6 mDice gains with mask prompt on SUN-SEG-Hard, with Hiera-T as the backbone. The code will be available at https://github.com/zhoustan/CamSAM2.
♻ ☆ HLV-1K: A Large-scale Hour-Long Video Benchmark for Time-Specific Long Video Understanding ICME 2025
Multimodal large language models have become a popular topic in deep visual understanding due to many promising real-world applications. However, hour-long video understanding, spanning over one hour and containing tens of thousands of visual frames, remains under-explored because of 1) challenging long-term video analyses, 2) inefficient large-model approaches, and 3) lack of large-scale benchmark datasets. Among them, in this paper, we focus on building a large-scale hour-long long video benchmark, HLV-1K, designed to evaluate long video understanding models. HLV-1K comprises 1009 hour-long videos with 14,847 high-quality question answering (QA) and multi-choice question asnwering (MCQA) pairs with time-aware query and diverse annotations, covering frame-level, within-event-level, cross-event-level, and long-term reasoning tasks. We evaluate our benchmark using existing state-of-the-art methods and demonstrate its value for testing deep long video understanding capabilities at different levels and for various tasks. This includes promoting future long video understanding tasks at a granular level, such as deep understanding of long live videos, meeting recordings, and movies.
comment: Accepted to ICME 2025
♻ ☆ Threshold Crossings as Tail Events for Catastrophic AI Risk
We analyse circumstances in which bifurcation-driven jumps in AI systems are associated with emergent heavy-tailed outcome distributions. By analysing how a control parameter's random fluctuations near a catastrophic threshold generate extreme outcomes, we demonstrate in what circumstances the probability of a sudden, large-scale, transition aligns closely with the tail probability of the resulting damage distribution. Our results contribute to research in monitoring, mitigation and control of AI systems when seeking to manage potentially catastrophic AI risk.
comment: Under peer review
♻ ☆ A Multimodal Vision Foundation Model for Clinical Dermatology
Diagnosing and treating skin diseases require advanced visual skills across domains and the ability to synthesize information from multiple imaging modalities. While current deep learning models excel at specific tasks like skin cancer diagnosis from dermoscopic images, they struggle to meet the complex, multimodal requirements of clinical practice. Here, we introduce PanDerm, a multimodal dermatology foundation model pretrained through self-supervised learning on over 2 million real-world skin disease images from 11 clinical institutions across 4 imaging modalities. We evaluated PanDerm on 28 diverse benchmarks, including skin cancer screening, risk stratification, differential diagnosis of common and rare skin conditions, lesion segmentation, longitudinal monitoring, and metastasis prediction and prognosis. PanDerm achieved state-of-the-art performance across all evaluated tasks, often outperforming existing models when using only 10% of labeled data. We conducted three reader studies to assess PanDerm's potential clinical utility. PanDerm outperformed clinicians by 10.2% in early-stage melanoma detection through longitudinal analysis, improved clinicians' skin cancer diagnostic accuracy by 11% on dermoscopy images, and enhanced non-dermatologist healthcare providers' differential diagnosis by 16.5% across 128 skin conditions on clinical photographs. These results demonstrate PanDerm's potential to improve patient care across diverse clinical scenarios and serve as a model for developing multimodal foundation models in other medical specialties, potentially accelerating the integration of AI support in healthcare.
comment: 74 pages; Preprint
♻ ☆ Do We Need to Verify Step by Step? Rethinking Process Supervision from a Theoretical Perspective
As large language models have evolved, it has become crucial to distinguish between process supervision and outcome supervision -- two key reinforcement learning approaches to complex reasoning tasks. While process supervision offers intuitive advantages for long-term credit assignment, the precise relationship between these paradigms has remained an open question. Conventional wisdom suggests that outcome supervision is fundamentally more challenging due to the trajectory-level coverage problem, leading to significant investment in collecting fine-grained process supervision data. In this paper, we take steps towards resolving this debate. Our main theorem shows that, under standard data coverage assumptions, reinforcement learning through outcome supervision is no more statistically difficult than through process supervision, up to polynomial factors in horizon. At the core of this result lies the novel Change of Trajectory Measure Lemma -- a technical tool that bridges return-based trajectory measure and step-level distribution shift. Furthermore, for settings with access to a verifier or a rollout capability, we prove that any policy's advantage function can serve as an optimal process reward model, providing a direct connection between outcome and process supervision. These findings suggest that the empirically observed performance gap -- if any -- between outcome and process supervision likely stems from algorithmic limitations rather than inherent statistical difficulties, potentially transforming how we approach data collection and algorithm design for reinforcement learning.
♻ ☆ CLEAR: Cue Learning using Evolution for Accurate Recognition Applied to Sustainability Data Extraction
Large Language Model (LLM) image recognition is a powerful tool for extracting data from images, but accuracy depends on providing sufficient cues in the prompt - requiring a domain expert for specialized tasks. We introduce Cue Learning using Evolution for Accurate Recognition (CLEAR), which uses a combination of LLMs and evolutionary computation to generate and optimize cues such that recognition of specialized features in images is improved. It achieves this by auto-generating a novel domain-specific representation and then using it to optimize suitable textual cues with a genetic algorithm. We apply CLEAR to the real-world task of identifying sustainability data from interior and exterior images of buildings. We investigate the effects of using a variable-length representation compared to fixed-length and show how LLM consistency can be improved by refactoring from categorical to real-valued estimates. We show that CLEAR enables higher accuracy compared to expert human recognition and human-authored prompts in every task with error rates improved by up to two orders of magnitude and an ablation study evincing solution concision.
comment: 9 pages plus 2 pages of supplemental material
♻ ☆ No Black Box Anymore: Demystifying Clinical Predictive Modeling with Temporal-Feature Cross Attention Mechanism
Despite the outstanding performance of deep learning models in clinical prediction tasks, explainability remains a significant challenge. Inspired by transformer architectures, we introduce the Temporal-Feature Cross Attention Mechanism (TFCAM), a novel deep learning framework designed to capture dynamic interactions among clinical features across time, enhancing both predictive accuracy and interpretability. In an experiment with 1,422 patients with Chronic Kidney Disease, predicting progression to End-Stage Renal Disease, TFCAM outperformed LSTM and RETAIN baselines, achieving an AUROC of 0.95 and an F1-score of 0.69. Beyond performance gains, TFCAM provides multi-level explainability by identifying critical temporal periods, ranking feature importance, and quantifying how features influence each other across time before affecting predictions. Our approach addresses the "black box" limitations of deep learning in healthcare, offering clinicians transparent insights into disease progression mechanisms while maintaining state-of-the-art predictive performance.
comment: 10 pages, 3 figures, submitted to AMIA 2025
♻ ☆ Multi-Agent Inverse Reinforcement Learning in Real World Unstructured Pedestrian Crowds
Social robot navigation in crowded public spaces such as university campuses, restaurants, grocery stores, and hospitals, is an increasingly important area of research. One of the core strategies for achieving this goal is to understand humans' intent--underlying psychological factors that govern their motion--by learning their reward functions, typically via inverse reinforcement learning (IRL). Despite significant progress in IRL, learning reward functions of multiple agents simultaneously in dense unstructured pedestrian crowds has remained intractable due to the nature of the tightly coupled social interactions that occur in these scenarios \textit{e.g.} passing, intersections, swerving, weaving, etc. In this paper, we present a new multi-agent maximum entropy inverse reinforcement learning algorithm for real world unstructured pedestrian crowds. Key to our approach is a simple, but effective, mathematical trick which we name the so-called tractability-rationality trade-off trick that achieves tractability at the cost of a slight reduction in accuracy. We compare our approach to the classical single-agent MaxEnt IRL as well as state-of-the-art trajectory prediction methods on several datasets including the ETH, UCY, SCAND, JRDB, and a new dataset, called Speedway, collected at a busy intersection on a University campus focusing on dense, complex agent interactions. Our key findings show that, on the dense Speedway dataset, our approach ranks 1st among top 7 baselines with >2X improvement over single-agent IRL, and is competitive with state-of-the-art large transformer-based encoder-decoder models on sparser datasets such as ETH/UCY (ranks 3rd among top 7 baselines).
♻ ☆ M-LLM Based Video Frame Selection for Efficient Video Understanding
Recent advances in Multi-Modal Large Language Models (M-LLMs) show promising results in video reasoning. Popular Multi-Modal Large Language Model (M-LLM) frameworks usually apply naive uniform sampling to reduce the number of video frames that are fed into an M-LLM, particularly for long context videos. However, it could lose crucial context in certain periods of a video, so that the downstream M-LLM may not have sufficient visual information to answer a question. To attack this pain point, we propose a light-weight M-LLM -based frame selection method that adaptively select frames that are more relevant to users' queries. In order to train the proposed frame selector, we introduce two supervision signals (i) Spatial signal, where single frame importance score by prompting a M-LLM; (ii) Temporal signal, in which multiple frames selection by prompting Large Language Model (LLM) using the captions of all frame candidates. The selected frames are then digested by a frozen downstream video M-LLM for visual reasoning and question answering. Empirical results show that the proposed M-LLM video frame selector improves the performances various downstream video Large Language Model (video-LLM) across medium (ActivityNet, NExT-QA) and long (EgoSchema, LongVideoBench) context video question answering benchmarks.
♻ ☆ Policy Learning with a Language Bottleneck
Modern AI systems such as self-driving cars and game-playing agents achieve superhuman performance, but often lack human-like generalization, interpretability, and inter-operability with human users. Inspired by the rich interactions between language and decision-making in humans, we introduce Policy Learning with a Language Bottleneck (PLLB), a framework enabling AI agents to generate linguistic rules that capture the high-level strategies underlying rewarding behaviors. PLLB alternates between a *rule generation* step guided by language models, and an *update* step where agents learn new policies guided by rules, even when a rule is insufficient to describe an entire complex policy. Across five diverse tasks, including a two-player signaling game, maze navigation, image reconstruction, and robot grasp planning, we show that PLLB agents are not only able to learn more interpretable and generalizable behaviors, but can also share the learned rules with human users, enabling more effective human-AI coordination. We provide source code for our experiments at https://github.com/meghabyte/bottleneck .
comment: 21 pages, 15 figures, updated with robot manipulation task
♻ ☆ JOG3R: Towards 3D-Consistent Video Generators
Emergent capabilities of image generators have led to many impactful zero- or few-shot applications. Inspired by this success, we investigate whether video generators similarly exhibit 3D-awareness. Using structure-from-motion as a 3D-aware task, we test if intermediate features of a video generator - OpenSora in our case - can support camera pose estimation. Surprisingly, at first, we only find a weak correlation between the two tasks. Deeper investigation reveals that although the video generator produces plausible video frames, the frames themselves are not truly 3D-consistent. Instead, we propose to jointly train for the two tasks, using photometric generation and 3D aware errors. Specifically, we find that SoTA video generation and camera pose estimation (i.e.,DUSt3R [79]) networks share common structures, and propose an architecture that unifies the two. The proposed unified model, named \nameMethod, produces camera pose estimates with competitive quality while producing 3D-consistent videos. In summary, we propose the first unified video generator that is 3D-consistent, generates realistic video frames, and can potentially be repurposed for other 3D-aware tasks.
♻ ☆ Modeling Story Expectations to Understand Engagement: A Generative Framework Using LLMs
Understanding when and why consumers engage with stories is crucial for content creators and platforms. While existing theories suggest that audience beliefs of what is going to happen should play an important role in engagement decisions, empirical work has mostly focused on developing techniques to directly extract features from actual content, rather than capturing forward-looking beliefs, due to the lack of a principled way to model such beliefs in unstructured narrative data. To complement existing feature extraction techniques, this paper introduces a novel framework that leverages large language models to model audience forward-looking beliefs about how stories might unfold. Our method generates multiple potential continuations for each story and extracts features related to expectations, uncertainty, and surprise using established content analysis techniques. Applying our method to over 30,000 book chapters, we demonstrate that our framework complements existing feature engineering techniques by amplifying their marginal explanatory power on average by 31%. The results reveal that different types of engagement-continuing to read, commenting, and voting-are driven by distinct combinations of current and anticipated content features. Our framework provides a novel way to study and explore how audience forward-looking beliefs shape their engagement with narrative media, with implications for marketing strategy in content-focused industries.
♻ ☆ EuroBERT: Scaling Multilingual Encoders for European Languages
General-purpose multilingual vector representations, used in retrieval, regression and classification, are traditionally obtained from bidirectional encoder models. Despite their wide applicability, encoders have been recently overshadowed by advances in generative decoder-only models. However, many innovations driving this progress are not inherently tied to decoders. In this paper, we revisit the development of multilingual encoders through the lens of these advances, and introduce EuroBERT, a family of multilingual encoders covering European and widely spoken global languages. Our models outperform existing alternatives across a diverse range of tasks, spanning multilingual capabilities, mathematics, and coding, and natively supporting sequences of up to 8,192 tokens. We also examine the design decisions behind EuroBERT, offering insights into our dataset composition and training pipeline. We publicly release the EuroBERT models, including intermediate training checkpoints, together with our training framework.
comment: 28 pages, 8 figures, 13 tables
Computer Vision and Pattern Recognition 221
☆ Free4D: Tuning-free 4D Scene Generation with Spatial-Temporal Consistency
We present Free4D, a novel tuning-free framework for 4D scene generation from a single image. Existing methods either focus on object-level generation, making scene-level generation infeasible, or rely on large-scale multi-view video datasets for expensive training, with limited generalization ability due to the scarcity of 4D scene data. In contrast, our key insight is to distill pre-trained foundation models for consistent 4D scene representation, which offers promising advantages such as efficiency and generalizability. 1) To achieve this, we first animate the input image using image-to-video diffusion models followed by 4D geometric structure initialization. 2) To turn this coarse structure into spatial-temporal consistent multiview videos, we design an adaptive guidance mechanism with a point-guided denoising strategy for spatial consistency and a novel latent replacement strategy for temporal coherence. 3) To lift these generated observations into consistent 4D representation, we propose a modulation-based refinement to mitigate inconsistencies while fully leveraging the generated information. The resulting 4D representation enables real-time, controllable rendering, marking a significant advancement in single-image-based 4D scene generation.
comment: Project Page: https://free4d.github.io/ , Code: https://github.com/TQTQliu/Free4D
☆ FB-4D: Spatial-Temporal Coherent Dynamic 3D Content Generation with Feature Banks
With the rapid advancements in diffusion models and 3D generation techniques, dynamic 3D content generation has become a crucial research area. However, achieving high-fidelity 4D (dynamic 3D) generation with strong spatial-temporal consistency remains a challenging task. Inspired by recent findings that pretrained diffusion features capture rich correspondences, we propose FB-4D, a novel 4D generation framework that integrates a Feature Bank mechanism to enhance both spatial and temporal consistency in generated frames. In FB-4D, we store features extracted from previous frames and fuse them into the process of generating subsequent frames, ensuring consistent characteristics across both time and multiple views. To ensure a compact representation, the Feature Bank is updated by a proposed dynamic merging mechanism. Leveraging this Feature Bank, we demonstrate for the first time that generating additional reference sequences through multiple autoregressive iterations can continuously improve generation performance. Experimental results show that FB-4D significantly outperforms existing methods in terms of rendering quality, spatial-temporal consistency, and robustness. It surpasses all multi-view generation tuning-free approaches by a large margin and achieves performance on par with training-based methods.
comment: Project page:https://fb-4d.c7w.tech/
☆ Zero-Shot Audio-Visual Editing via Cross-Modal Delta Denoising
In this paper, we introduce zero-shot audio-video editing, a novel task that requires transforming original audio-visual content to align with a specified textual prompt without additional model training. To evaluate this task, we curate a benchmark dataset, AvED-Bench, designed explicitly for zero-shot audio-video editing. AvED-Bench includes 110 videos, each with a 10-second duration, spanning 11 categories from VGGSound. It offers diverse prompts and scenarios that require precise alignment between auditory and visual elements, enabling robust evaluation. We identify limitations in existing zero-shot audio and video editing methods, particularly in synchronization and coherence between modalities, which often result in inconsistent outcomes. To address these challenges, we propose AvED, a zero-shot cross-modal delta denoising framework that leverages audio-video interactions to achieve synchronized and coherent edits. AvED demonstrates superior results on both AvED-Bench and the recent OAVE dataset to validate its generalization capabilities. Results are available at https://genjib.github.io/project_page/AVED/index.html
comment: Project page: https://genjib.github.io/project_page/AVED/index.html
☆ BASKET: A Large-Scale Video Dataset for Fine-Grained Skill Estimation
We present BASKET, a large-scale basketball video dataset for fine-grained skill estimation. BASKET contains 4,477 hours of video capturing 32,232 basketball players from all over the world. Compared to prior skill estimation datasets, our dataset includes a massive number of skilled participants with unprecedented diversity in terms of gender, age, skill level, geographical location, etc. BASKET includes 20 fine-grained basketball skills, challenging modern video recognition models to capture the intricate nuances of player skill through in-depth video analysis. Given a long highlight video (8-10 minutes) of a particular player, the model needs to predict the skill level (e.g., excellent, good, average, fair, poor) for each of the 20 basketball skills. Our empirical analysis reveals that the current state-of-the-art video models struggle with this task, significantly lagging behind the human baseline. We believe that BASKET could be a useful resource for developing new video models with advanced long-range, fine-grained recognition capabilities. In addition, we hope that our dataset will be useful for domain-specific applications such as fair basketball scouting, personalized player development, and many others. Dataset and code are available at https://github.com/yulupan00/BASKET.
☆ Feature4X: Bridging Any Monocular Video to 4D Agentic AI with Versatile Gaussian Feature Fields
Recent advancements in 2D and multimodal models have achieved remarkable success by leveraging large-scale training on extensive datasets. However, extending these achievements to enable free-form interactions and high-level semantic operations with complex 3D/4D scenes remains challenging. This difficulty stems from the limited availability of large-scale, annotated 3D/4D or multi-view datasets, which are crucial for generalizable vision and language tasks such as open-vocabulary and prompt-based segmentation, language-guided editing, and visual question answering (VQA). In this paper, we introduce Feature4X, a universal framework designed to extend any functionality from 2D vision foundation model into the 4D realm, using only monocular video input, which is widely available from user-generated content. The "X" in Feature4X represents its versatility, enabling any task through adaptable, model-conditioned 4D feature field distillation. At the core of our framework is a dynamic optimization strategy that unifies multiple model capabilities into a single representation. Additionally, to the best of our knowledge, Feature4X is the first method to distill and lift the features of video foundation models (e.g. SAM2, InternVideo2) into an explicit 4D feature field using Gaussian Splatting. Our experiments showcase novel view segment anything, geometric and appearance scene editing, and free-form VQA across all time steps, empowered by LLMs in feedback loops. These advancements broaden the scope of agentic AI applications by providing a foundation for scalable, contextually and spatiotemporally aware systems capable of immersive dynamic 4D scene interaction.
☆ Disentangled Source-Free Personalization for Facial Expression Recognition with Neutral Target Data
Facial Expression Recognition (FER) from videos is a crucial task in various application areas, such as human-computer interaction and health monitoring (e.g., pain, depression, fatigue, and stress). Beyond the challenges of recognizing subtle emotional or health states, the effectiveness of deep FER models is often hindered by the considerable variability of expressions among subjects. Source-free domain adaptation (SFDA) methods are employed to adapt a pre-trained source model using only unlabeled target domain data, thereby avoiding data privacy and storage issues. Typically, SFDA methods adapt to a target domain dataset corresponding to an entire population and assume it includes data from all recognition classes. However, collecting such comprehensive target data can be difficult or even impossible for FER in healthcare applications. In many real-world scenarios, it may be feasible to collect a short neutral control video (displaying only neutral expressions) for target subjects before deployment. These videos can be used to adapt a model to better handle the variability of expressions among subjects. This paper introduces the Disentangled Source-Free Domain Adaptation (DSFDA) method to address the SFDA challenge posed by missing target expression data. DSFDA leverages data from a neutral target control video for end-to-end generation and adaptation of target data with missing non-neutral data. Our method learns to disentangle features related to expressions and identity while generating the missing non-neutral target data, thereby enhancing model accuracy. Additionally, our self-supervision strategy improves model adaptation by reconstructing target images that maintain the same identity and source expression.
☆ ADS-Edit: A Multimodal Knowledge Editing Dataset for Autonomous Driving Systems
Recent advancements in Large Multimodal Models (LMMs) have shown promise in Autonomous Driving Systems (ADS). However, their direct application to ADS is hindered by challenges such as misunderstanding of traffic knowledge, complex road conditions, and diverse states of vehicle. To address these challenges, we propose the use of Knowledge Editing, which enables targeted modifications to a model's behavior without the need for full retraining. Meanwhile, we introduce ADS-Edit, a multimodal knowledge editing dataset specifically designed for ADS, which includes various real-world scenarios, multiple data types, and comprehensive evaluation metrics. We conduct comprehensive experiments and derive several interesting conclusions. We hope that our work will contribute to the further advancement of knowledge editing applications in the field of autonomous driving. Code and data are available in https://github.com/zjunlp/EasyEdit.
comment: Work in progress
☆ Reason-RFT: Reinforcement Fine-Tuning for Visual Reasoning
Visual reasoning abilities play a crucial role in understanding complex multimodal data, advancing both domain-specific applications and artificial general intelligence (AGI). Existing methods improve VLM reasoning via Chain-of-Thought (CoT) supervised fine-tuning, using meticulously annotated training data to enhance visual reasoning capabilities. However, this training paradigm may lead to overfitting and cognitive rigidity, restricting the model's ability to transfer visual reasoning skills across domains and limiting its real-world applicability. To address these limitations, we propose Reason-RFT, a novel reinforcement fine-tuning framework that significantly enhances generalization capabilities in visual reasoning tasks. Reason-RFT introduces a two-phase training framework for visual reasoning: (1) Supervised Fine-Tuning (SFT) with curated Chain-of-Thought (CoT) data activates the reasoning potential of Vision-Language Models (VLMs), followed by (2) Group Relative Policy Optimization (GRPO)-based reinforcement learning that generates multiple reasoning-response pairs, significantly enhancing generalization in visual reasoning tasks. To evaluate Reason-RFT's visual reasoning capabilities, we reconstructed a comprehensive dataset spanning visual counting, structure perception, and spatial transformation.cExperimental results demonstrate Reasoning-RFT's three key advantages: (1) Performance Enhancement: achieving state-of-the-art results across multiple tasks, outperforming most mainstream open-source and proprietary models; (2) Generalization Superiority: consistently maintaining robust performance across diverse tasks and domains, outperforming alternative training paradigms; (3) Data Efficiency: excelling in few-shot learning scenarios while surpassing full-dataset SFT baselines.
comment: 35 pages, 22 figures
☆ UniSTD: Towards Unified Spatio-Temporal Learning across Diverse Disciplines CVPR 2025
Traditional spatiotemporal models generally rely on task-specific architectures, which limit their generalizability and scalability across diverse tasks due to domain-specific design requirements. In this paper, we introduce \textbf{UniSTD}, a unified Transformer-based framework for spatiotemporal modeling, which is inspired by advances in recent foundation models with the two-stage pretraining-then-adaption paradigm. Specifically, our work demonstrates that task-agnostic pretraining on 2D vision and vision-text datasets can build a generalizable model foundation for spatiotemporal learning, followed by specialized joint training on spatiotemporal datasets to enhance task-specific adaptability. To improve the learning capabilities across domains, our framework employs a rank-adaptive mixture-of-expert adaptation by using fractional interpolation to relax the discrete variables so that can be optimized in the continuous space. Additionally, we introduce a temporal module to incorporate temporal dynamics explicitly. We evaluate our approach on a large-scale dataset covering 10 tasks across 4 disciplines, demonstrating that a unified spatiotemporal model can achieve scalable, cross-task learning and support up to 10 tasks simultaneously within one model while reducing training costs in multi-domain applications. Code will be available at https://github.com/1hunters/UniSTD.
comment: Accepted to CVPR 2025
☆ PhysGen3D: Crafting a Miniature Interactive World from a Single Image CVPR 2025
Envisioning physically plausible outcomes from a single image requires a deep understanding of the world's dynamics. To address this, we introduce PhysGen3D, a novel framework that transforms a single image into an amodal, camera-centric, interactive 3D scene. By combining advanced image-based geometric and semantic understanding with physics-based simulation, PhysGen3D creates an interactive 3D world from a static image, enabling us to "imagine" and simulate future scenarios based on user input. At its core, PhysGen3D estimates 3D shapes, poses, physical and lighting properties of objects, thereby capturing essential physical attributes that drive realistic object interactions. This framework allows users to specify precise initial conditions, such as object speed or material properties, for enhanced control over generated video outcomes. We evaluate PhysGen3D's performance against closed-source state-of-the-art (SOTA) image-to-video models, including Pika, Kling, and Gen-3, showing PhysGen3D's capacity to generate videos with realistic physics while offering greater flexibility and fine-grained control. Our results show that PhysGen3D achieves a unique balance of photorealism, physical plausibility, and user-driven interactivity, opening new possibilities for generating dynamic, physics-grounded video from an image.
comment: CVPR 2025, Project page: https://by-luckk.github.io/PhysGen3D
☆ MATHGLANCE: Multimodal Large Language Models Do Not Know Where to Look in Mathematical Diagrams
Diagrams serve as a fundamental form of visual language, representing complex concepts and their inter-relationships through structured symbols, shapes, and spatial arrangements. Unlike natural images, their inherently symbolic and abstract nature poses significant challenges for Multimodal Large Language Models (MLLMs). However, current benchmarks conflate perceptual and reasoning tasks, making it difficult to assess whether MLLMs genuinely understand mathematical diagrams beyond superficial pattern recognition. To address this gap, we introduce MATHGLANCE, a benchmark specifically designed to isolate and evaluate mathematical perception in MLLMs. MATHGLANCE comprises 1.2K images and 1.6K carefully curated questions spanning four perception tasks: shape classification, object counting, relationship identification, and object grounding, covering diverse domains including plane geometry, solid geometry, and graphical representations. Our evaluation of MLLMs reveals that their ability to understand diagrams is notably limited, particularly in fine-grained grounding tasks. In response, we construct GeoPeP, a perception-oriented dataset of 200K structured geometry image-text pairs explicitly annotated with geometric primitives and precise spatial relationships. Training MLLM on GeoPeP leads to significant gains in perceptual accuracy, which in turn substantially improves mathematical reasoning. Our benchmark and dataset establish critical standards for evaluating and advancing multimodal mathematical understanding, providing valuable resources and insights to foster future MLLM research.
☆ High Quality Diffusion Distillation on a Single GPU with Relative and Absolute Position Matching
We introduce relative and absolute position matching (RAPM), a diffusion distillation method resulting in high quality generation that can be trained efficiently on a single GPU. Recent diffusion distillation research has achieved excellent results for high-resolution text-to-image generation with methods such as phased consistency models (PCM) and improved distribution matching distillation (DMD2). However, these methods generally require many GPUs (e.g.~8-64) and significant batchsizes (e.g.~128-2048) during training, resulting in memory and compute requirements that are beyond the resources of some researchers. RAPM provides effective single-GPU diffusion distillation training with a batchsize of 1. The new method attempts to mimic the sampling trajectories of the teacher model by matching the relative and absolute positions. The design of relative positions is inspired by PCM. Two discriminators are introduced accordingly in RAPM, one for matching relative positions and the other for absolute positions. Experimental results on StableDiffusion (SD) V1.5 and SDXL indicate that RAPM with 4 timesteps produces comparable FID scores as the best method with 1 timestep under very limited computational resources.
☆ Emotion Detection and Music Recommendation System
As artificial intelligence becomes more and more ingrained in daily life, we present a novel system that uses deep learning for music recommendation and emotion-based detection. Through the use of facial recognition and the DeepFace framework, our method analyses human emotions in real-time and then plays music that reflects the mood it has discovered. The system uses a webcam to take pictures, analyses the most common facial expression, and then pulls a playlist from local storage that corresponds to the mood it has detected. An engaging and customised experience is ensured by allowing users to manually change the song selection via a dropdown menu or navigation buttons. By continuously looping over the playlist, the technology guarantees continuity. The objective of our system is to improve emotional well-being through music therapy by offering a responsive and automated music-selection experience.
☆ SChanger: Change Detection from a Semantic Change and Spatial Consistency Perspective
Change detection is a key task in Earth observation applications. Recently, deep learning methods have demonstrated strong performance and widespread application. However, change detection faces data scarcity due to the labor-intensive process of accurately aligning remote sensing images of the same area, which limits the performance of deep learning algorithms. To address the data scarcity issue, we develop a fine-tuning strategy called the Semantic Change Network (SCN). We initially pre-train the model on single-temporal supervised tasks to acquire prior knowledge of instance feature extraction. The model then employs a shared-weight Siamese architecture and extended Temporal Fusion Module (TFM) to preserve this prior knowledge and is fine-tuned on change detection tasks. The learned semantics for identifying all instances is changed to focus on identifying only the changes. Meanwhile, we observe that the locations of changes between the two images are spatially identical, a concept we refer to as spatial consistency. We introduce this inductive bias through an attention map that is generated by large-kernel convolutions and applied to the features from both time points. This enhances the modeling of multi-scale changes and helps capture underlying relationships in change detection semantics. We develop a binary change detection model utilizing these two strategies. The model is validated against state-of-the-art methods on six datasets, surpassing all benchmark methods and achieving F1 scores of 92.87%, 86.43%, 68.95%, 97.62%, 84.58%, and 93.20% on the LEVIR-CD, LEVIR-CD+, S2Looking, CDD, SYSU-CD, and WHU-CD datasets, respectively.
☆ Dynamic Motion Blending for Versatile Motion Editing
Text-guided motion editing enables high-level semantic control and iterative modifications beyond traditional keyframe animation. Existing methods rely on limited pre-collected training triplets, which severely hinders their versatility in diverse editing scenarios. We introduce MotionCutMix, an online data augmentation technique that dynamically generates training triplets by blending body part motions based on input text. While MotionCutMix effectively expands the training distribution, the compositional nature introduces increased randomness and potential body part incoordination. To model such a rich distribution, we present MotionReFit, an auto-regressive diffusion model with a motion coordinator. The auto-regressive architecture facilitates learning by decomposing long sequences, while the motion coordinator mitigates the artifacts of motion composition. Our method handles both spatial and temporal motion edits directly from high-level human instructions, without relying on additional specifications or Large Language Models. Through extensive experiments, we show that MotionReFit achieves state-of-the-art performance in text-guided motion editing.
☆ A weakly-supervised deep learning model for fast localisation and delineation of the skeleton, internal organs, and spinal canal on Whole-Body Diffusion-Weighted MRI (WB-DWI)
Background: Apparent Diffusion Coefficient (ADC) values and Total Diffusion Volume (TDV) from Whole-body diffusion-weighted MRI (WB-DWI) are recognized cancer imaging biomarkers. However, manual disease delineation for ADC and TDV measurements is unfeasible in clinical practice, demanding automation. As a first step, we propose an algorithm to generate fast and reproducible probability maps of the skeleton, adjacent internal organs (liver, spleen, urinary bladder, and kidneys), and spinal canal. Methods: We developed an automated deep-learning pipeline based on a 3D patch-based Residual U-Net architecture that localizes and delineates these anatomical structures on WB-DWI. The algorithm was trained using "soft-labels" (non-binary segmentations) derived from a computationally intensive atlas-based approach. For training and validation, we employed a multi-center WB-DWI dataset comprising 532 scans from patients with Advanced Prostate Cancer (APC) or Multiple Myeloma (MM), with testing on 45 patients. Results: Our weakly-supervised deep learning model achieved an average dice score/precision/recall of 0.66/0.6/0.73 for skeletal delineations, 0.8/0.79/0.81 for internal organs, and 0.85/0.79/0.94 for spinal canal, with surface distances consistently below 3 mm. Relative median ADC and log-transformed volume differences between automated and manual expert-defined full-body delineations were below 10% and 4%, respectively. The computational time for generating probability maps was 12x faster than the atlas-based registration algorithm (25 s vs. 5 min). An experienced radiologist rated the model's accuracy "good" or "excellent" on test datasets. Conclusion: Our model offers fast and reproducible probability maps for localizing and delineating body regions on WB-DWI, enabling ADC and TDV quantification, potentially supporting clinicians in disease staging and treatment response assessment.
☆ Demand Estimation with Text and Image Data
We propose a demand estimation method that leverages unstructured text and image data to infer substitution patterns. Using pre-trained deep learning models, we extract embeddings from product images and textual descriptions and incorporate them into a random coefficients logit model. This approach enables researchers to estimate demand even when they lack data on product attributes or when consumers value hard-to-quantify attributes, such as visual design or functional benefits. Using data from a choice experiment, we show that our approach outperforms standard attribute-based models in counterfactual predictions of consumers' second choices. We also apply it across 40 product categories on Amazon.com and consistently find that text and image data help identify close substitutes within each category.
☆ MMMORRF: Multimodal Multilingual Modularized Reciprocal Rank Fusion
Videos inherently contain multiple modalities, including visual events, text overlays, sounds, and speech, all of which are important for retrieval. However, state-of-the-art multimodal language models like VAST and LanguageBind are built on vision-language models (VLMs), and thus overly prioritize visual signals. Retrieval benchmarks further reinforce this bias by focusing on visual queries and neglecting other modalities. We create a search system MMMORRF that extracts text and features from both visual and audio modalities and integrates them with a novel modality-aware weighted reciprocal rank fusion. MMMORRF is both effective and efficient, demonstrating practicality in searching videos based on users' information needs instead of visual descriptive queries. We evaluate MMMORRF on MultiVENT 2.0 and TVR, two multimodal benchmarks designed for more targeted information needs, and find that it improves nDCG@20 by 81% over leading multimodal encoders and 37% over single-modality retrieval, demonstrating the value of integrating diverse modalities.
☆ Flip Learning: Weakly Supervised Erase to Segment Nodules in Breast Ultrasound
Accurate segmentation of nodules in both 2D breast ultrasound (BUS) and 3D automated breast ultrasound (ABUS) is crucial for clinical diagnosis and treatment planning. Therefore, developing an automated system for nodule segmentation can enhance user independence and expedite clinical analysis. Unlike fully-supervised learning, weakly-supervised segmentation (WSS) can streamline the laborious and intricate annotation process. However, current WSS methods face challenges in achieving precise nodule segmentation, as many of them depend on inaccurate activation maps or inefficient pseudo-mask generation algorithms. In this study, we introduce a novel multi-agent reinforcement learning-based WSS framework called Flip Learning, which relies solely on 2D/3D boxes for accurate segmentation. Specifically, multiple agents are employed to erase the target from the box to facilitate classification tag flipping, with the erased region serving as the predicted segmentation mask. The key contributions of this research are as follows: (1) Adoption of a superpixel/supervoxel-based approach to encode the standardized environment, capturing boundary priors and expediting the learning process. (2) Introduction of three meticulously designed rewards, comprising a classification score reward and two intensity distribution rewards, to steer the agents' erasing process precisely, thereby avoiding both under- and over-segmentation. (3) Implementation of a progressive curriculum learning strategy to enable agents to interact with the environment in a progressively challenging manner, thereby enhancing learning efficiency. Extensively validated on the large in-house BUS and ABUS datasets, our Flip Learning method outperforms state-of-the-art WSS methods and foundation models, and achieves comparable performance as fully-supervised learning algorithms.
comment: Accepted by Medical Image Analysis. 24 pages, 13 figures, 18 tabels
☆ GLRD: Global-Local Collaborative Reason and Debate with PSL for 3D Open-Vocabulary Detection
The task of LiDAR-based 3D Open-Vocabulary Detection (3D OVD) requires the detector to learn to detect novel objects from point clouds without off-the-shelf training labels. Previous methods focus on the learning of object-level representations and ignore the scene-level information, thus it is hard to distinguish objects with similar classes. In this work, we propose a Global-Local Collaborative Reason and Debate with PSL (GLRD) framework for the 3D OVD task, considering both local object-level information and global scene-level information. Specifically, LLM is utilized to perform common sense reasoning based on object-level and scene-level information, where the detection result is refined accordingly. To further boost the LLM's ability of precise decisions, we also design a probabilistic soft logic solver (OV-PSL) to search for the optimal solution, and a debate scheme to confirm the class of confusable objects. In addition, to alleviate the uneven distribution of classes, a static balance scheme (SBC) and a dynamic balance scheme (DBC) are designed. In addition, to reduce the influence of noise in data and training, we further propose Reflected Pseudo Labels Generation (RPLG) and Background-Aware Object Localization (BAOL). Extensive experiments conducted on ScanNet and SUN RGB-D demonstrate the superiority of GLRD, where absolute improvements in mean average precision are $+2.82\%$ on SUN RGB-D and $+3.72\%$ on ScanNet in the partial open-vocabulary setting. In the full open-vocabulary setting, the absolute improvements in mean average precision are $+4.03\%$ on ScanNet and $+14.11\%$ on SUN RGB-D.
comment: 15 pages
☆ Benchmarking Machine Learning Methods for Distributed Acoustic Sensing
Distributed acoustic sensing (DAS) technology represents an innovative fiber-optic-based sensing methodology that enables real-time acoustic signal monitoring through the detection of minute perturbations along optical fibers. This sensing approach offers compelling advantages, including extensive measurement ranges, exceptional spatial resolution, and an expansive dynamic measurement spectrum. The integration of machine learning (ML) paradigms presents transformative potential for DAS technology, encompassing critical domains such as data augmentation, sophisticated preprocessing techniques, and advanced acoustic event classification and recognition. By leveraging ML algorithms, DAS systems can transition from traditional data processing methodologies to more automated and intelligent analytical frameworks. The computational intelligence afforded by ML-enhanced DAS technologies facilitates unprecedented monitoring capabilities across diverse critical infrastructure sectors. Particularly noteworthy are the technology's applications in transportation infrastructure, energy management systems, and Natural disaster monitoring frameworks, where the precision of data acquisition and the reliability of intelligent decision-making mechanisms are paramount. This research critically examines the comparative performance characteristics of classical machine learning methodologies and state-of-the-art deep learning models in the context of DAS data recognition and interpretation, offering comprehensive insights into the evolving landscape of intelligent sensing technologies.
☆ Vision as LoRA
We introduce Vision as LoRA (VoRA), a novel paradigm for transforming an LLM into an MLLM. Unlike prevalent MLLM architectures that rely on external vision modules for vision encoding, VoRA internalizes visual capabilities by integrating vision-specific LoRA layers directly into the LLM. This design allows the added parameters to be seamlessly merged into the LLM during inference, eliminating structural complexity and minimizing computational overhead. Moreover, inheriting the LLM's ability of handling flexible context, VoRA can process inputs at arbitrary resolutions. To further strengthen VoRA's visual capabilities, we introduce a block-wise distillation method that transfers visual priors from a pre-trained ViT into the LoRA layers, effectively accelerating training by injecting visual knowledge. Additionally, we apply bi-directional attention masks to better capture the context information of an image. We successfully demonstrate that with additional pre-training data, VoRA can perform comparably with conventional encode-based MLLMs. All training data, codes, and model weights will be released at https://github.com/Hon-Wong/VoRA.
☆ Mitigating Low-Level Visual Hallucinations Requires Self-Awareness: Database, Model and Training Strategy
The rapid development of multimodal large language models has resulted in remarkable advancements in visual perception and understanding, consolidating several tasks into a single visual question-answering framework. However, these models are prone to hallucinations, which limit their reliability as artificial intelligence systems. While this issue is extensively researched in natural language processing and image captioning, there remains a lack of investigation of hallucinations in Low-level Visual Perception and Understanding (HLPU), especially in the context of image quality assessment tasks. We consider that these hallucinations arise from an absence of clear self-awareness within the models. To address this issue, we first introduce the HLPU instruction database, the first instruction database specifically focused on hallucinations in low-level vision tasks. This database contains approximately 200K question-answer pairs and comprises four subsets, each covering different types of instructions. Subsequently, we propose the Self-Awareness Failure Elimination (SAFEQA) model, which utilizes image features, salient region features and quality features to improve the perception and comprehension abilities of the model in low-level vision tasks. Furthermore, we propose the Enhancing Self-Awareness Preference Optimization (ESA-PO) framework to increase the model's awareness of knowledge boundaries, thereby mitigating the incidence of hallucination. Finally, we conduct comprehensive experiments on low-level vision tasks, with the results demonstrating that our proposed method significantly enhances self-awareness of the model in these tasks and reduces hallucinations. Notably, our proposed method improves both accuracy and self-awareness of the proposed model and outperforms close-source models in terms of various evaluation metrics.
☆ BizGen: Advancing Article-level Visual Text Rendering for Infographics Generation CVPR 2025
Recently, state-of-the-art text-to-image generation models, such as Flux and Ideogram 2.0, have made significant progress in sentence-level visual text rendering. In this paper, we focus on the more challenging scenarios of article-level visual text rendering and address a novel task of generating high-quality business content, including infographics and slides, based on user provided article-level descriptive prompts and ultra-dense layouts. The fundamental challenges are twofold: significantly longer context lengths and the scarcity of high-quality business content data. In contrast to most previous works that focus on a limited number of sub-regions and sentence-level prompts, ensuring precise adherence to ultra-dense layouts with tens or even hundreds of sub-regions in business content is far more challenging. We make two key technical contributions: (i) the construction of scalable, high-quality business content dataset, i.e., Infographics-650K, equipped with ultra-dense layouts and prompts by implementing a layer-wise retrieval-augmented infographic generation scheme; and (ii) a layout-guided cross attention scheme, which injects tens of region-wise prompts into a set of cropped region latent space according to the ultra-dense layouts, and refine each sub-regions flexibly during inference using a layout conditional CFG. We demonstrate the strong results of our system compared to previous SOTA systems such as Flux and SD3 on our BizEval prompt set. Additionally, we conduct thorough ablation experiments to verify the effectiveness of each component. We hope our constructed Infographics-650K and BizEval can encourage the broader community to advance the progress of business content generation.
comment: Accepted by CVPR 2025. Project Page: https://bizgen-msra.github.io
☆ AutoRad-Lung: A Radiomic-Guided Prompting Autoregressive Vision-Language Model for Lung Nodule Malignancy Prediction
Lung cancer remains one of the leading causes of cancer-related mortality worldwide. A crucial challenge for early diagnosis is differentiating uncertain cases with similar visual characteristics and closely annotation scores. In clinical practice, radiologists rely on quantitative, hand-crafted Radiomic features extracted from Computed Tomography (CT) images, while recent research has primarily focused on deep learning solutions. More recently, Vision-Language Models (VLMs), particularly Contrastive Language-Image Pre-Training (CLIP)-based models, have gained attention for their ability to integrate textual knowledge into lung cancer diagnosis. While CLIP-Lung models have shown promising results, we identified the following potential limitations: (a) dependence on radiologists' annotated attributes, which are inherently subjective and error-prone, (b) use of textual information only during training, limiting direct applicability at inference, and (c) Convolutional-based vision encoder with randomly initialized weights, which disregards prior knowledge. To address these limitations, we introduce AutoRad-Lung, which couples an autoregressively pre-trained VLM, with prompts generated from hand-crafted Radiomics. AutoRad-Lung uses the vision encoder of the Large-Scale Autoregressive Image Model (AIMv2), pre-trained using a multi-modal autoregressive objective. Given that lung tumors are typically small, irregularly shaped, and visually similar to healthy tissue, AutoRad-Lung offers significant advantages over its CLIP-based counterparts by capturing pixel-level differences. Additionally, we introduce conditional context optimization, which dynamically generates context-specific prompts based on input Radiomics, improving cross-modal alignment.
☆ ARMO: Autoregressive Rigging for Multi-Category Objects
Recent advancements in large-scale generative models have significantly improved the quality and diversity of 3D shape generation. However, most existing methods focus primarily on generating static 3D models, overlooking the potentially dynamic nature of certain shapes, such as humanoids, animals, and insects. To address this gap, we focus on rigging, a fundamental task in animation that establishes skeletal structures and skinning for 3D models. In this paper, we introduce OmniRig, the first large-scale rigging dataset, comprising 79,499 meshes with detailed skeleton and skinning information. Unlike traditional benchmarks that rely on predefined standard poses (e.g., A-pose, T-pose), our dataset embraces diverse shape categories, styles, and poses. Leveraging this rich dataset, we propose ARMO, a novel rigging framework that utilizes an autoregressive model to predict both joint positions and connectivity relationships in a unified manner. By treating the skeletal structure as a complete graph and discretizing it into tokens, we encode the joints using an auto-encoder to obtain a latent embedding and an autoregressive model to predict the tokens. A mesh-conditioned latent diffusion model is used to predict the latent embedding for conditional skeleton generation. Our method addresses the limitations of regression-based approaches, which often suffer from error accumulation and suboptimal connectivity estimation. Through extensive experiments on the OmniRig dataset, our approach achieves state-of-the-art performance in skeleton prediction, demonstrating improved generalization across diverse object categories. The code and dataset will be made public for academic use upon acceptance.
☆ AccidentSim: Generating Physically Realistic Vehicle Collision Videos from Real-World Accident Reports
Collecting real-world vehicle accident videos for autonomous driving research is challenging due to their rarity and complexity. While existing driving video generation methods may produce visually realistic videos, they often fail to deliver physically realistic simulations because they lack the capability to generate accurate post-collision trajectories. In this paper, we introduce AccidentSim, a novel framework that generates physically realistic vehicle collision videos by extracting and utilizing the physical clues and contextual information available in real-world vehicle accident reports. Specifically, AccidentSim leverages a reliable physical simulator to replicate post-collision vehicle trajectories from the physical and contextual information in the accident reports and to build a vehicle collision trajectory dataset. This dataset is then used to fine-tune a language model, enabling it to respond to user prompts and predict physically consistent post-collision trajectories across various driving scenarios based on user descriptions. Finally, we employ Neural Radiance Fields (NeRF) to render high-quality backgrounds, merging them with the foreground vehicles that exhibit physically realistic trajectories to generate vehicle collision videos. Experimental results demonstrate that the videos produced by AccidentSim excel in both visual and physical authenticity.
☆ UWarp: A Whole Slide Image Registration Pipeline to Characterize Scanner-Induced Local Domain Shift
Histopathology slide digitization introduces scanner-induced domain shift that can significantly impact computational pathology models based on deep learning methods. In the state-of-the-art, this shift is often characterized at a broad scale (slide-level or dataset-level) but not patch-level, which limits our comprehension of the impact of localized tissue characteristics on the accuracy of the deep learning models. To address this challenge, we present a domain shift analysis framework based on UWarp, a novel registration tool designed to accurately align histological slides scanned under varying conditions. UWarp employs a hierarchical registration approach, combining global affine transformations with fine-grained local corrections to achieve robust tissue patch alignment. We evaluate UWarp using two private datasets, CypathLung and BosomShieldBreast, containing whole slide images scanned by multiple devices. Our experiments demonstrate that UWarp outperforms existing open-source registration methods, achieving a median target registration error (TRE) of less than 4 pixels (<1 micrometer at 40x magnification) while significantly reducing computational time. Additionally, we apply UWarp to characterize scanner-induced local domain shift in the predictions of Breast-NEOprAIdict, a deep learning model for breast cancer pathological response prediction. We find that prediction variability is strongly correlated with tissue density on a given patch. Our findings highlight the importance of localized domain shift analysis and suggest that UWarp can serve as a valuable tool for improving model robustness and domain adaptation strategies in computational pathology.
☆ Imitating Radiological Scrolling: A Global-Local Attention Model for 3D Chest CT Volumes Multi-Label Anomaly Classification
The rapid increase in the number of Computed Tomography (CT) scan examinations has created an urgent need for automated tools, such as organ segmentation, anomaly classification, and report generation, to assist radiologists with their growing workload. Multi-label classification of Three-Dimensional (3D) CT scans is a challenging task due to the volumetric nature of the data and the variety of anomalies to be detected. Existing deep learning methods based on Convolutional Neural Networks (CNNs) struggle to capture long-range dependencies effectively, while Vision Transformers require extensive pre-training, posing challenges for practical use. Additionally, these existing methods do not explicitly model the radiologist's navigational behavior while scrolling through CT scan slices, which requires both global context understanding and local detail awareness. In this study, we present CT-Scroll, a novel global-local attention model specifically designed to emulate the scrolling behavior of radiologists during the analysis of 3D CT scans. Our approach is evaluated on two public datasets, demonstrating its efficacy through comprehensive experiments and an ablation study that highlights the contribution of each model component.
comment: 13 pages, 4 figures, under review for MIDL 2025
☆ MMGen: Unified Multi-modal Image Generation and Understanding in One Go
A unified diffusion framework for multi-modal generation and understanding has the transformative potential to achieve seamless and controllable image diffusion and other cross-modal tasks. In this paper, we introduce MMGen, a unified framework that integrates multiple generative tasks into a single diffusion model. This includes: (1) multi-modal category-conditioned generation, where multi-modal outputs are generated simultaneously through a single inference process, given category information; (2) multi-modal visual understanding, which accurately predicts depth, surface normals, and segmentation maps from RGB images; and (3) multi-modal conditioned generation, which produces corresponding RGB images based on specific modality conditions and other aligned modalities. Our approach develops a novel diffusion transformer that flexibly supports multi-modal output, along with a simple modality-decoupling strategy to unify various tasks. Extensive experiments and applications demonstrate the effectiveness and superiority of MMGen across diverse tasks and conditions, highlighting its potential for applications that require simultaneous generation and understanding.
comment: Our project page: https://jiepengwang.github.io/MMGen/
☆ Robust Flower Cluster Matching Using The Unscented Transform
Monitoring flowers over time is essential for precision robotic pollination in agriculture. To accomplish this, a continuous spatial-temporal observation of plant growth can be done using stationary RGB-D cameras. However, image registration becomes a serious challenge due to changes in the visual appearance of the plant caused by the pollination process and occlusions from growth and camera angles. Plants flower in a manner that produces distinct clusters on branches. This paper presents a method for matching flower clusters using descriptors generated from RGB-D data and considers allowing for spatial uncertainty within the cluster. The proposed approach leverages the Unscented Transform to efficiently estimate plant descriptor uncertainty tolerances, enabling a robust image-registration process despite temporal changes. The Unscented Transform is used to handle the nonlinear transformations by propagating the uncertainty of flower positions to determine the variations in the descriptor domain. A Monte Carlo simulation is used to validate the Unscented Transform results, confirming our method's effectiveness for flower cluster matching. Therefore, it can facilitate improved robotics pollination in dynamic environments.
comment: *CASE2025 Under Review*
☆ IAP: Improving Continual Learning of Vision-Language Models via Instance-Aware Prompting
Recent pre-trained vision-language models (PT-VLMs) often face a Multi-Domain Class-Incremental Learning (MCIL) scenario in practice, where several classes and domains of multi-modal tasks are incrementally arrived. Without access to previously learned tasks and unseen tasks, memory-constrained MCIL suffers from forward and backward forgetting. To alleviate the above challenges, parameter-efficient fine-tuning techniques (PEFT), such as prompt tuning, are employed to adapt the PT-VLM to the diverse incrementally learned tasks. To achieve effective new task adaptation, existing methods only consider the effect of PEFT strategy selection, but neglect the influence of PEFT parameter setting (e.g., prompting). In this paper, we tackle the challenge of optimizing prompt designs for diverse tasks in MCIL and propose an Instance-Aware Prompting (IAP) framework. Specifically, our Instance-Aware Gated Prompting (IA-GP) module enhances adaptation to new tasks while mitigating forgetting by dynamically assigning prompts across transformer layers at the instance level. Our Instance-Aware Class-Distribution-Driven Prompting (IA-CDDP) improves the task adaptation process by determining an accurate task-label-related confidence score for each instance. Experimental evaluations across 11 datasets, using three performance metrics, demonstrate the effectiveness of our proposed method. Code can be found at https://github.com/FerdinandZJU/IAP.
comment: Code can be found at https://github.com/FerdinandZJU/IAP
Diffusion Counterfactuals for Image Regressors
Counterfactual explanations have been successfully applied to create human interpretable explanations for various black-box models. They are handy for tasks in the image domain, where the quality of the explanations benefits from recent advances in generative models. Although counterfactual explanations have been widely applied to classification models, their application to regression tasks remains underexplored. We present two methods to create counterfactual explanations for image regression tasks using diffusion-based generative models to address challenges in sparsity and quality: 1) one based on a Denoising Diffusion Probabilistic Model that operates directly in pixel-space and 2) another based on a Diffusion Autoencoder operating in latent space. Both produce realistic, semantic, and smooth counterfactuals on CelebA-HQ and a synthetic data set, providing easily interpretable insights into the decision-making process of the regression model and reveal spurious correlations. We find that for regression counterfactuals, changes in features depend on the region of the predicted value. Large semantic changes are needed for significant changes in predicted values, making it harder to find sparse counterfactuals than with classifiers. Moreover, pixel space counterfactuals are more sparse while latent space counterfactuals are of higher quality and allow bigger semantic changes.
comment: 24 Pages, 5 Figures, Accepted at 3rd World Conference on eXplainable Artificial Intelligence (xAI-2025), Code and reproduction instructions available on GitHub, see https://github.com/DevinTDHa/Diffusion-Counterfactuals-for-Image-Regressors
☆ Exploring Robustness of Cortical Morphometry in the presence of white matter lesions, using Diffusion Models for Lesion Filling
Cortical thickness measurements from magnetic resonance imaging, an important biomarker in many neurodegenerative and neurological disorders, are derived by many tools from an initial voxel-wise tissue segmentation. White matter (WM) hypointensities in T1-weighted imaging, such as those arising from multiple sclerosis or small vessel disease, are known to affect the output of brain segmentation methods and therefore bias cortical thickness measurements. These effects are well-documented among traditional brain segmentation tools but have not been studied extensively in tools based on deep-learning segmentations, which promise to be more robust. In this paper, we explore the potential of deep learning to enhance the accuracy and efficiency of cortical thickness measurement in the presence of WM lesions, using a high-quality lesion filling algorithm leveraging denoising diffusion networks. A pseudo-3D U-Net architecture trained on the OASIS dataset to generate synthetic healthy tissue, conditioned on binary lesion masks derived from the MSSEG dataset, allows realistic removal of white matter lesions in multiple sclerosis patients. By applying morphometry methods to patient images before and after lesion filling, we analysed robustness of global and regional cortical thickness measurements in the presence of white matter lesions. Methods based on a deep learning-based segmentation of the brain (Fastsurfer, DL+DiReCT, ANTsPyNet) exhibited greater robustness than those using classical segmentation methods (Freesurfer, ANTs).
☆ TerraTorch: The Geospatial Foundation Models Toolkit RSS 2025
TerraTorch is a fine-tuning and benchmarking toolkit for Geospatial Foundation Models built on PyTorch Lightning and tailored for satellite, weather, and climate data. It integrates domain-specific data modules, pre-defined tasks, and a modular model factory that pairs any backbone with diverse decoder heads. These components allow researchers and practitioners to fine-tune supported models in a no-code fashion by simply editing a training configuration. By consolidating best practices for model development and incorporating the automated hyperparameter optimization extension Iterate, TerraTorch reduces the expertise and time required to fine-tune or benchmark models on new Earth Observation use cases. Furthermore, TerraTorch directly integrates with GEO-Bench, allowing for systematic and reproducible benchmarking of Geospatial Foundation Models. TerraTorch is open sourced under Apache 2.0, available at https://github.com/IBM/terratorch, and can be installed via pip install terratorch.
comment: IGARSS 2025
☆ Beyond Intermediate States: Explaining Visual Redundancy through Language
Multi-modal Large Langue Models (MLLMs) often process thousands of visual tokens, which consume a significant portion of the context window and impose a substantial computational burden. Prior work has empirically explored visual token pruning methods based on MLLMs' intermediate states (e.g., attention scores). However, they have limitations in precisely defining visual redundancy due to their inability to capture the influence of visual tokens on MLLMs' visual understanding (i.e., the predicted probabilities for textual token candidates). To address this issue, we manipulate the visual input and investigate variations in the textual output from both token-centric and context-centric perspectives, achieving intuitive and comprehensive analysis. Experimental results reveal that visual tokens with low ViT-[cls] association and low text-to-image attention scores can contain recognizable information and significantly contribute to images' overall information. To develop a more reliable method for identifying and pruning redundant visual tokens, we integrate these two perspectives and introduce a context-independent condition to identify redundant prototypes from training images, which probes the redundancy of each visual token during inference. Extensive experiments on single-image, multi-image and video comprehension tasks demonstrate the effectiveness of our method, notably achieving 90% to 110% of the performance while pruning 80% to 90% of visual tokens.
☆ TD-BFR: Truncated Diffusion Model for Efficient Blind Face Restoration ICME 2025
Diffusion-based methodologies have shown significant potential in blind face restoration (BFR), leveraging their robust generative capabilities. However, they are often criticized for two significant problems: 1) slow training and inference speed, and 2) inadequate recovery of fine-grained facial details. To address these problems, we propose a novel Truncated Diffusion model for efficient Blind Face Restoration (TD-BFR), a three-stage paradigm tailored for the progressive resolution of degraded images. Specifically, TD-BFR utilizes an innovative truncated sampling method, starting from low-quality (LQ) images at low resolution to enhance sampling speed, and then introduces an adaptive degradation removal module to handle unknown degradations and connect the generation processes across different resolutions. Additionally, we further adapt the priors of pre-trained diffusion models to recover rich facial details. Our method efficiently restores high-quality images in a coarse-to-fine manner and experimental results demonstrate that TD-BFR is, on average, \textbf{4.75$\times$} faster than current state-of-the-art diffusion-based BFR methods while maintaining competitive quality.
comment: Accepted by ICME 2025
☆ GAIA-2: A Controllable Multi-View Generative World Model for Autonomous Driving
Generative models offer a scalable and flexible paradigm for simulating complex environments, yet current approaches fall short in addressing the domain-specific requirements of autonomous driving - such as multi-agent interactions, fine-grained control, and multi-camera consistency. We introduce GAIA-2, Generative AI for Autonomy, a latent diffusion world model that unifies these capabilities within a single generative framework. GAIA-2 supports controllable video generation conditioned on a rich set of structured inputs: ego-vehicle dynamics, agent configurations, environmental factors, and road semantics. It generates high-resolution, spatiotemporally consistent multi-camera videos across geographically diverse driving environments (UK, US, Germany). The model integrates both structured conditioning and external latent embeddings (e.g., from a proprietary driving model) to facilitate flexible and semantically grounded scene synthesis. Through this integration, GAIA-2 enables scalable simulation of both common and rare driving scenarios, advancing the use of generative world models as a core tool in the development of autonomous systems. Videos are available at https://wayve.ai/thinking/gaia-2.
comment: Technical Report
☆ MAR-3D: Progressive Masked Auto-regressor for High-Resolution 3D Generation CVPR 2025
Recent advances in auto-regressive transformers have revolutionized generative modeling across different domains, from language processing to visual generation, demonstrating remarkable capabilities. However, applying these advances to 3D generation presents three key challenges: the unordered nature of 3D data conflicts with sequential next-token prediction paradigm, conventional vector quantization approaches incur substantial compression loss when applied to 3D meshes, and the lack of efficient scaling strategies for higher resolution latent prediction. To address these challenges, we introduce MAR-3D, which integrates a pyramid variational autoencoder with a cascaded masked auto-regressive transformer (Cascaded MAR) for progressive latent upscaling in the continuous space. Our architecture employs random masking during training and auto-regressive denoising in random order during inference, naturally accommodating the unordered property of 3D latent tokens. Additionally, we propose a cascaded training strategy with condition augmentation that enables efficiently up-scale the latent token resolution with fast convergence. Extensive experiments demonstrate that MAR-3D not only achieves superior performance and generalization capabilities compared to existing methods but also exhibits enhanced scaling capabilities compared to joint distribution modeling approaches (e.g., diffusion transformers).
comment: Aceepted to CVPR 2025
☆ Small Object Detection: A Comprehensive Survey on Challenges, Techniques and Real-World Applications
Small object detection (SOD) is a critical yet challenging task in computer vision, with applications like spanning surveillance, autonomous systems, medical imaging, and remote sensing. Unlike larger objects, small objects contain limited spatial and contextual information, making accurate detection difficult. Challenges such as low resolution, occlusion, background interference, and class imbalance further complicate the problem. This survey provides a comprehensive review of recent advancements in SOD using deep learning, focusing on articles published in Q1 journals during 2024-2025. We analyzed challenges, state-of-the-art techniques, datasets, evaluation metrics, and real-world applications. Recent advancements in deep learning have introduced innovative solutions, including multi-scale feature extraction, Super-Resolution (SR) techniques, attention mechanisms, and transformer-based architectures. Additionally, improvements in data augmentation, synthetic data generation, and transfer learning have addressed data scarcity and domain adaptation issues. Furthermore, emerging trends such as lightweight neural networks, knowledge distillation (KD), and self-supervised learning offer promising directions for improving detection efficiency, particularly in resource-constrained environments like Unmanned Aerial Vehicles (UAV)-based surveillance and edge computing. We also review widely used datasets, along with standard evaluation metrics such as mean Average Precision (mAP) and size-specific AP scores. The survey highlights real-world applications, including traffic monitoring, maritime surveillance, industrial defect detection, and precision agriculture. Finally, we discuss open research challenges and future directions, emphasizing the need for robust domain adaptation techniques, better feature fusion strategies, and real-time performance optimization.
☆ Vision-Amplified Semantic Entropy for Hallucination Detection in Medical Visual Question Answering
Multimodal large language models (MLLMs) have demonstrated significant potential in medical Visual Question Answering (VQA). Yet, they remain prone to hallucinations-incorrect responses that contradict input images, posing substantial risks in clinical decision-making. Detecting these hallucinations is essential for establishing trust in MLLMs among clinicians and patients, thereby enabling their real-world adoption. Current hallucination detection methods, especially semantic entropy (SE), have demonstrated promising hallucination detection capacity for LLMs. However, adapting SE to medical MLLMs by incorporating visual perturbations presents a dilemma. Weak perturbations preserve image content and ensure clinical validity, but may be overlooked by medical MLLMs, which tend to over rely on language priors. In contrast, strong perturbations can distort essential diagnostic features, compromising clinical interpretation. To address this issue, we propose Vision Amplified Semantic Entropy (VASE), which incorporates weak image transformations and amplifies the impact of visual input, to improve hallucination detection in medical VQA. We first estimate the semantic predictive distribution under weak visual transformations to preserve clinical validity, and then amplify visual influence by contrasting this distribution with that derived from a distorted image. The entropy of the resulting distribution is estimated as VASE. Experiments on two medical open-ended VQA datasets demonstrate that VASE consistently outperforms existing hallucination detection methods.
comment: 11 pages, 2 figures
☆ MLLM-Selector: Necessity and Diversity-driven High-Value Data Selection for Enhanced Visual Instruction Tuning
Visual instruction tuning (VIT) has emerged as a crucial technique for enabling multi-modal large language models (MLLMs) to follow user instructions adeptly. Yet, a significant gap persists in understanding the attributes of high-quality instruction tuning data and frameworks for its automated selection. To address this, we introduce MLLM-Selector, an automated approach that identifies valuable data for VIT by weighing necessity and diversity. Our process starts by randomly sampling a subset from the VIT data pool to fine-tune a pretrained model, thus creating a seed model with an initial ability to follow instructions. Then, leveraging the seed model, we calculate necessity scores for each sample in the VIT data pool to identify samples pivotal for enhancing model performance. Our findings underscore the importance of mixing necessity and diversity in data choice, leading to the creation of MLLM-Selector, our methodology that fuses necessity scoring with strategic sampling for superior data refinement. Empirical results indicate that within identical experimental conditions, MLLM-Selector surpasses LLaVA-1.5 in some benchmarks with less than 1% of the data and consistently exceeds performance across all validated benchmarks when using less than 50%.
comment: Tech Report
☆ Towards Efficient and General-Purpose Few-Shot Misclassification Detection for Vision-Language Models
Reliable prediction by classifiers is crucial for their deployment in high security and dynamically changing situations. However, modern neural networks often exhibit overconfidence for misclassified predictions, highlighting the need for confidence estimation to detect errors. Despite the achievements obtained by existing methods on small-scale datasets, they all require training from scratch and there are no efficient and effective misclassification detection (MisD) methods, hindering practical application towards large-scale and ever-changing datasets. In this paper, we pave the way to exploit vision language model (VLM) leveraging text information to establish an efficient and general-purpose misclassification detection framework. By harnessing the power of VLM, we construct FSMisD, a Few-Shot prompt learning framework for MisD to refrain from training from scratch and therefore improve tuning efficiency. To enhance misclassification detection ability, we use adaptive pseudo sample generation and a novel negative loss to mitigate the issue of overconfidence by pushing category prompts away from pseudo features. We conduct comprehensive experiments with prompt learning methods and validate the generalization ability across various datasets with domain shift. Significant and consistent improvement demonstrates the effectiveness, efficiency and generalizability of our approach.
comment: preprint
☆ VPO: Aligning Text-to-Video Generation Models with Prompt Optimization
Video generation models have achieved remarkable progress in text-to-video tasks. These models are typically trained on text-video pairs with highly detailed and carefully crafted descriptions, while real-world user inputs during inference are often concise, vague, or poorly structured. This gap makes prompt optimization crucial for generating high-quality videos. Current methods often rely on large language models (LLMs) to refine prompts through in-context learning, but suffer from several limitations: they may distort user intent, omit critical details, or introduce safety risks. Moreover, they optimize prompts without considering the impact on the final video quality, which can lead to suboptimal results. To address these issues, we introduce VPO, a principled framework that optimizes prompts based on three core principles: harmlessness, accuracy, and helpfulness. The generated prompts faithfully preserve user intents and, more importantly, enhance the safety and quality of generated videos. To achieve this, VPO employs a two-stage optimization approach. First, we construct and refine a supervised fine-tuning (SFT) dataset based on principles of safety and alignment. Second, we introduce both text-level and video-level feedback to further optimize the SFT model with preference learning. Our extensive experiments demonstrate that VPO significantly improves safety, alignment, and video quality compared to baseline methods. Moreover, VPO shows strong generalization across video generation models. Furthermore, we demonstrate that VPO could outperform and be combined with RLHF methods on video generation models, underscoring the effectiveness of VPO in aligning video generation models. Our code and data are publicly available at https://github.com/thu-coai/VPO.
☆ Contrastive Learning Guided Latent Diffusion Model for Image-to-Image Translation
The diffusion model has demonstrated superior performance in synthesizing diverse and high-quality images for text-guided image translation. However, there remains room for improvement in both the formulation of text prompts and the preservation of reference image content. First, variations in target text prompts can significantly influence the quality of the generated images, and it is often challenging for users to craft an optimal prompt that fully captures the content of the input image. Second, while existing models can introduce desired modifications to specific regions of the reference image, they frequently induce unintended alterations in areas that should remain unchanged. To address these challenges, we propose pix2pix-zeroCon, a zero-shot diffusion-based method that eliminates the need for additional training by leveraging patch-wise contrastive loss. Specifically, we automatically determine the editing direction in the text embedding space based on the reference image and target prompts. Furthermore, to ensure precise content and structural preservation in the edited image, we introduce cross-attention guiding loss and patch-wise contrastive loss between the generated and original image embeddings within a pre-trained diffusion model. Notably, our approach requires no additional training and operates directly on a pre-trained text-to-image diffusion model. Extensive experiments demonstrate that our method surpasses existing models in image-to-image translation, achieving enhanced fidelity and controllability.
comment: 11 pages, 13 figures
☆ Dissecting and Mitigating Diffusion Bias via Mechanistic Interpretability CVPR 2025
Diffusion models have demonstrated impressive capabilities in synthesizing diverse content. However, despite their high-quality outputs, these models often perpetuate social biases, including those related to gender and race. These biases can potentially contribute to harmful real-world consequences, reinforcing stereotypes and exacerbating inequalities in various social contexts. While existing research on diffusion bias mitigation has predominantly focused on guiding content generation, it often neglects the intrinsic mechanisms within diffusion models that causally drive biased outputs. In this paper, we investigate the internal processes of diffusion models, identifying specific decision-making mechanisms, termed bias features, embedded within the model architecture. By directly manipulating these features, our method precisely isolates and adjusts the elements responsible for bias generation, permitting granular control over the bias levels in the generated content. Through experiments on both unconditional and conditional diffusion models across various social bias attributes, we demonstrate our method's efficacy in managing generation distribution while preserving image quality. We also dissect the discovered model mechanism, revealing different intrinsic features controlling fine-grained aspects of generation, boosting further research on mechanistic interpretability of diffusion models.
comment: CVPR 2025; Project Page: https://foundation-model-research.github.io/difflens
☆ From Trial to Triumph: Advancing Long Video Understanding via Visual Context Sample Scaling and Self-reward Alignment
Multi-modal Large language models (MLLMs) show remarkable ability in video understanding. Nevertheless, understanding long videos remains challenging as the models can only process a finite number of frames in a single inference, potentially omitting crucial visual information. To address the challenge, we propose generating multiple predictions through visual context sampling, followed by a scoring mechanism to select the final prediction. Specifically, we devise a bin-wise sampling strategy that enables MLLMs to generate diverse answers based on various combinations of keyframes, thereby enriching the visual context. To determine the final prediction from the sampled answers, we employ a self-reward by linearly combining three scores: (1) a frequency score indicating the prevalence of each option, (2) a marginal confidence score reflecting the inter-intra sample certainty of MLLM predictions, and (3) a reasoning score for different question types, including clue-guided answering for global questions and temporal self-refocusing for local questions. The frequency score ensures robustness through majority correctness, the confidence-aligned score reflects prediction certainty, and the typed-reasoning score addresses cases with sparse key visual information using tailored strategies. Experiments show that this approach covers the correct answer for a high percentage of long video questions, on seven datasets show that our method improves the performance of three MLLMs.
☆ Lipschitz Constant Meets Condition Number: Learning Robust and Compact Deep Neural Networks
Recent research has revealed that high compression of Deep Neural Networks (DNNs), e.g., massive pruning of the weight matrix of a DNN, leads to a severe drop in accuracy and susceptibility to adversarial attacks. Integration of network pruning into an adversarial training framework has been proposed to promote adversarial robustness. It has been observed that a highly pruned weight matrix tends to be ill-conditioned, i.e., increasing the condition number of the weight matrix. This phenomenon aggravates the vulnerability of a DNN to input noise. Although a highly pruned weight matrix is considered to be able to lower the upper bound of the local Lipschitz constant to tolerate large distortion, the ill-conditionedness of such a weight matrix results in a non-robust DNN model. To overcome this challenge, this work develops novel joint constraints to adjust the weight distribution of networks, namely, the Transformed Sparse Constraint joint with Condition Number Constraint (TSCNC), which copes with smoothing distribution and differentiable constraint functions to reduce condition number and thus avoid the ill-conditionedness of weight matrices. Furthermore, our theoretical analyses unveil the relevance between the condition number and the local Lipschitz constant of the weight matrix, namely, the sharply increasing condition number becomes the dominant factor that restricts the robustness of over-sparsified models. Extensive experiments are conducted on several public datasets, and the results show that the proposed constraints significantly improve the robustness of a DNN with high pruning rates.
comment: 13 pages, 6 figures
☆ Attention Xception UNet (AXUNet): A Novel Combination of CNN and Self-Attention for Brain Tumor Segmentation
Accurate segmentation of glioma brain tumors is crucial for diagnosis and treatment planning. Deep learning techniques offer promising solutions, but optimal model architectures remain under investigation. We used the BraTS 2021 dataset, selecting T1 with contrast enhancement (T1CE), T2, and Fluid-Attenuated Inversion Recovery (FLAIR) sequences for model development. The proposed Attention Xception UNet (AXUNet) architecture integrates an Xception backbone with dot-product self-attention modules, inspired by state-of-the-art (SOTA) large language models such as Google Bard and OpenAI ChatGPT, within a UNet-shaped model. We compared AXUNet with SOTA models. Comparative evaluation on the test set demonstrated improved results over baseline models. Inception-UNet and Xception-UNet achieved mean Dice scores of 90.88 and 93.24, respectively. Attention ResUNet (AResUNet) attained a mean Dice score of 92.80, with the highest score of 84.92 for enhancing tumor (ET) among all models. Attention Gate UNet (AGUNet) yielded a mean Dice score of 90.38. AXUNet outperformed all models with a mean Dice score of 93.73. It demonstrated superior Dice scores across whole tumor (WT) and tumor core (TC) regions, achieving 92.59 for WT, 86.81 for TC, and 84.89 for ET. The integration of the Xception backbone and dot-product self-attention mechanisms in AXUNet showcases enhanced performance in capturing spatial and contextual information. The findings underscore the potential utility of AXUNet in facilitating precise tumor delineation.
☆ Siformer: Feature-isolated Transformer for Efficient Skeleton-based Sign Language Recognition
Sign language recognition (SLR) refers to interpreting sign language glosses from given videos automatically. This research area presents a complex challenge in computer vision because of the rapid and intricate movements inherent in sign languages, which encompass hand gestures, body postures, and even facial expressions. Recently, skeleton-based action recognition has attracted increasing attention due to its ability to handle variations in subjects and backgrounds independently. However, current skeleton-based SLR methods exhibit three limitations: 1) they often neglect the importance of realistic hand poses, where most studies train SLR models on non-realistic skeletal representations; 2) they tend to assume complete data availability in both training or inference phases, and capture intricate relationships among different body parts collectively; 3) these methods treat all sign glosses uniformly, failing to account for differences in complexity levels regarding skeletal representations. To enhance the realism of hand skeletal representations, we present a kinematic hand pose rectification method for enforcing constraints. Mitigating the impact of missing data, we propose a feature-isolated mechanism to focus on capturing local spatial-temporal context. This method captures the context concurrently and independently from individual features, thus enhancing the robustness of the SLR model. Additionally, to adapt to varying complexity levels of sign glosses, we develop an input-adaptive inference approach to optimise computational efficiency and accuracy. Experimental results demonstrate the effectiveness of our approach, as evidenced by achieving a new state-of-the-art (SOTA) performance on WLASL100 and LSA64. For WLASL100, we achieve a top-1 accuracy of 86.50\%, marking a relative improvement of 2.39% over the previous SOTA. For LSA64, we achieve a top-1 accuracy of 99.84%.
comment: 10 pages, ACM Multimedia
☆ Latent Beam Diffusion Models for Decoding Image Sequences
While diffusion models excel at generating high-quality images from text prompts, they struggle with visual consistency in image sequences. Existing methods generate each image independently, leading to disjointed narratives - a challenge further exacerbated in non-linear storytelling, where scenes must connect beyond adjacent frames. We introduce a novel beam search strategy for latent space exploration, enabling conditional generation of full image sequences with beam search decoding. Unlike prior approaches that use fixed latent priors, our method dynamically searches for an optimal sequence of latent representations, ensuring coherent visual transitions. To address beam search's quadratic complexity, we integrate a cross-attention mechanism that efficiently scores search paths and enables pruning, prioritizing alignment with both textual prompts and visual context. Human evaluations confirm that our approach outperforms baseline methods, producing full sequences with superior coherence, visual continuity, and textual alignment. By bridging advances in search optimization and latent space refinement, this work sets a new standard for structured image sequence generation.
☆ Evaluating Facial Expression Recognition Datasets for Deep Learning: A Benchmark Study with Novel Similarity Metrics
This study investigates the key characteristics and suitability of widely used Facial Expression Recognition (FER) datasets for training deep learning models. In the field of affective computing, FER is essential for interpreting human emotions, yet the performance of FER systems is highly contingent on the quality and diversity of the underlying datasets. To address this issue, we compiled and analyzed 24 FER datasets, including those targeting specific age groups such as children, adults, and the elderly, and processed them through a comprehensive normalization pipeline. In addition, we enriched the datasets with automatic annotations for age and gender, enabling a more nuanced evaluation of their demographic properties. To further assess dataset efficacy, we introduce three novel metricsLocal, Global, and Paired Similarity, which quantitatively measure dataset difficulty, generalization capability, and cross-dataset transferability. Benchmark experiments using state-of-the-art neural networks reveal that large-scale, automatically collected datasets (e.g., AffectNet, FER2013) tend to generalize better, despite issues with labeling noise and demographic biases, whereas controlled datasets offer higher annotation quality but limited variability. Our findings provide actionable recommendations for dataset selection and design, advancing the development of more robust, fair, and effective FER systems.
☆ Cherry Yield Forecast: Harvest Prediction for Individual Sweet Cherry Trees
This paper is part of a publication series from the For5G project that has the goal of creating digital twins of sweet cherry trees. At the beginning a brief overview of the revious work in this project is provided. Afterwards the focus shifts to a crucial problem in the fruit farming domain: the difficulty of making reliable yield predictions early in the season. Following three Satin sweet cherry trees along the year 2023 enabled the collection of accurate ground truth data about the development of cherries from dormancy until harvest. The methodology used to collect this data is presented, along with its valuation and visualization. The predictive power of counting objects at all relevant vegetative stages of the fruit development cycle in cherry trees with regards to yield predictions is investigated. It is found that all investigated fruit states are suitable for yield predictions based on linear regression. Conceptionally, there is a trade-off between earliness and external events with the potential to invalidate the prediction. Considering this, two optimal timepoints are suggested that are opening cluster stage before the start of the flowering and the early fruit stage right after the second fruit drop. However, both timepoints are challenging to solve with automated procedures based on image data. Counting developing cherries based on images is exceptionally difficult due to the small fruit size and their tendency to be occluded by leaves. It was not possible to obtain satisfying results relying on a state-of-the-art fruit-counting method. Counting the elements within a bursting bud is also challenging, even when using high resolution cameras. It is concluded that accurate yield prediction for sweet cherry trees is possible when objects are manually counted and that automated features extraction with similar accuracy remains an open problem yet to be solved.
☆ ITA-MDT: Image-Timestep-Adaptive Masked Diffusion Transformer Framework for Image-Based Virtual Try-On CVPR 2025
This paper introduces ITA-MDT, the Image-Timestep-Adaptive Masked Diffusion Transformer Framework for Image-Based Virtual Try-On (IVTON), designed to overcome the limitations of previous approaches by leveraging the Masked Diffusion Transformer (MDT) for improved handling of both global garment context and fine-grained details. The IVTON task involves seamlessly superimposing a garment from one image onto a person in another, creating a realistic depiction of the person wearing the specified garment. Unlike conventional diffusion-based virtual try-on models that depend on large pre-trained U-Net architectures, ITA-MDT leverages a lightweight, scalable transformer-based denoising diffusion model with a mask latent modeling scheme, achieving competitive results while reducing computational overhead. A key component of ITA-MDT is the Image-Timestep Adaptive Feature Aggregator (ITAFA), a dynamic feature aggregator that combines all of the features from the image encoder into a unified feature of the same size, guided by diffusion timestep and garment image complexity. This enables adaptive weighting of features, allowing the model to emphasize either global information or fine-grained details based on the requirements of the denoising stage. Additionally, the Salient Region Extractor (SRE) module is presented to identify complex region of the garment to provide high-resolution local information to the denoising model as an additional condition alongside the global information of the full garment image. This targeted conditioning strategy enhances detail preservation of fine details in highly salient garment regions, optimizing computational resources by avoiding unnecessarily processing entire garment image. Comparative evaluations confirms that ITA-MDT improves efficiency while maintaining strong performance, reaching state-of-the-art results in several metrics.
comment: CVPR 2025, Project Page: https://jiwoohong93.github.io/ita-mdt/
☆ RSRWKV: A Linear-Complexity 2D Attention Mechanism for Efficient Remote Sensing Vision Task
High-resolution remote sensing analysis faces challenges in global context modeling due to scene complexity and scale diversity. While CNNs excel at local feature extraction via parameter sharing, their fixed receptive fields fundamentally restrict long-range dependency modeling. Vision Transformers (ViTs) effectively capture global semantic relationships through self-attention mechanisms but suffer from quadratic computational complexity relative to image resolution, creating critical efficiency bottlenecks for high-resolution imagery. The RWKV model's linear-complexity sequence modeling achieves breakthroughs in NLP but exhibits anisotropic limitations in vision tasks due to its 1D scanning mechanism. To address these challenges, we propose RSRWKV, featuring a novel 2D-WKV scanning mechanism that bridges sequential processing and 2D spatial reasoning while maintaining linear complexity. This enables isotropic context aggregation across multiple directions. The MVC-Shift module enhances multi-scale receptive field coverage, while the ECA module strengthens cross-channel feature interaction and semantic saliency modeling. Experimental results demonstrate RSRWKV's superior performance over CNN and Transformer baselines in classification, detection, and segmentation tasks on NWPU RESISC45, VHR-10.v2, and GLH-Water datasets, offering a scalable solution for high-resolution remote sensing analysis.
☆ Pluggable Style Representation Learning for Multi-Style Transfer
Due to the high diversity of image styles, the scalability to various styles plays a critical role in real-world applications. To accommodate a large amount of styles, previous multi-style transfer approaches rely on enlarging the model size while arbitrary-style transfer methods utilize heavy backbones. However, the additional computational cost introduced by more model parameters hinders these methods to be deployed on resource-limited devices. To address this challenge, in this paper, we develop a style transfer framework by decoupling the style modeling and transferring. Specifically, for style modeling, we propose a style representation learning scheme to encode the style information into a compact representation. Then, for style transferring, we develop a style-aware multi-style transfer network (SaMST) to adapt to diverse styles using pluggable style representations. In this way, our framework is able to accommodate diverse image styles in the learned style representations without introducing additional overhead during inference, thereby maintaining efficiency. Experiments show that our style representation can extract accurate style information. Moreover, qualitative and quantitative results demonstrate that our method achieves state-of-the-art performance in terms of both accuracy and efficiency. The codes are available in https://github.com/The-Learning-And-Vision-Atelier-LAVA/SaMST.
comment: 18 pages, 13 figures, 2 tables
☆ Self-ReS: Self-Reflection in Large Vision-Language Models for Long Video Understanding
Large Vision-Language Models (LVLMs) demonstrate remarkable performance in short-video tasks such as video question answering, but struggle in long-video understanding. The linear frame sampling strategy, conventionally used by LVLMs, fails to account for the non-linear distribution of key events in video data, often introducing redundant or irrelevant information in longer contexts while risking the omission of critical events in shorter ones. To address this, we propose SelfReS, a non-linear spatiotemporal self-reflective sampling method that dynamically selects key video fragments based on user prompts. Unlike prior approaches, SelfReS leverages the inherently sparse attention maps of LVLMs to define reflection tokens, enabling relevance-aware token selection without requiring additional training or external modules. Experiments demonstrate that SelfReS can be seamlessly integrated into strong base LVLMs, improving long-video task accuracy and achieving up to 46% faster inference speed within the same GPU memory budget.
☆ SURGEON: Memory-Adaptive Fully Test-Time Adaptation via Dynamic Activation Sparsity CVPR 2025
Despite the growing integration of deep models into mobile terminals, the accuracy of these models declines significantly due to various deployment interferences. Test-time adaptation (TTA) has emerged to improve the performance of deep models by adapting them to unlabeled target data online. Yet, the significant memory cost, particularly in resource-constrained terminals, impedes the effective deployment of most backward-propagation-based TTA methods. To tackle memory constraints, we introduce SURGEON, a method that substantially reduces memory cost while preserving comparable accuracy improvements during fully test-time adaptation (FTTA) without relying on specific network architectures or modifications to the original training procedure. Specifically, we propose a novel dynamic activation sparsity strategy that directly prunes activations at layer-specific dynamic ratios during adaptation, allowing for flexible control of learning ability and memory cost in a data-sensitive manner. Among this, two metrics, Gradient Importance and Layer Activation Memory, are considered to determine the layer-wise pruning ratios, reflecting accuracy contribution and memory efficiency, respectively. Experimentally, our method surpasses the baselines by not only reducing memory usage but also achieving superior accuracy, delivering SOTA performance across diverse datasets, architectures, and tasks.
comment: Accepted to CVPR 2025
☆ Consistency Trajectory Matching for One-Step Generative Super-Resolution
Current diffusion-based super-resolution (SR) approaches achieve commendable performance at the cost of high inference overhead. Therefore, distillation techniques are utilized to accelerate the multi-step teacher model into one-step student model. Nevertheless, these methods significantly raise training costs and constrain the performance of the student model by the teacher model. To overcome these tough challenges, we propose Consistency Trajectory Matching for Super-Resolution (CTMSR), a distillation-free strategy that is able to generate photo-realistic SR results in one step. Concretely, we first formulate a Probability Flow Ordinary Differential Equation (PF-ODE) trajectory to establish a deterministic mapping from low-resolution (LR) images with noise to high-resolution (HR) images. Then we apply the Consistency Training (CT) strategy to directly learn the mapping in one step, eliminating the necessity of pre-trained diffusion model. To further enhance the performance and better leverage the ground-truth during the training process, we aim to align the distribution of SR results more closely with that of the natural images. To this end, we propose to minimize the discrepancy between their respective PF-ODE trajectories from the LR image distribution by our meticulously designed Distribution Trajectory Matching (DTM) loss, resulting in improved realism of our recovered HR images. Comprehensive experimental results demonstrate that the proposed methods can attain comparable or even superior capabilities on both synthetic and real datasets while maintaining minimal inference latency.
☆ VideoGEM: Training-free Action Grounding in Videos
Vision-language foundation models have shown impressive capabilities across various zero-shot tasks, including training-free localization and grounding, primarily focusing on localizing objects in images. However, leveraging those capabilities to localize actions and events in videos is challenging, as actions have less physical outline and are usually described by higher-level concepts. In this work, we propose VideoGEM, the first training-free spatial action grounding method based on pretrained image- and video-language backbones. Namely, we adapt the self-self attention formulation of GEM to spatial activity grounding. We observe that high-level semantic concepts, such as actions, usually emerge in the higher layers of the image- and video-language models. We, therefore, propose a layer weighting in the self-attention path to prioritize higher layers. Additionally, we introduce a dynamic weighting method to automatically tune layer weights to capture each layer`s relevance to a specific prompt. Finally, we introduce a prompt decomposition, processing action, verb, and object prompts separately, resulting in a better spatial localization of actions. We evaluate the proposed approach on three image- and video-language backbones, CLIP, OpenCLIP, and ViCLIP, and on four video grounding datasets, V-HICO, DALY, YouCook-Interactions, and GroundingYouTube, showing that the proposed training-free approach is able to outperform current trained state-of-the-art approaches for spatial video grounding.
☆ Progressive Focused Transformer for Single Image Super-Resolution
Transformer-based methods have achieved remarkable results in image super-resolution tasks because they can capture non-local dependencies in low-quality input images. However, this feature-intensive modeling approach is computationally expensive because it calculates the similarities between numerous features that are irrelevant to the query features when obtaining attention weights. These unnecessary similarity calculations not only degrade the reconstruction performance but also introduce significant computational overhead. How to accurately identify the features that are important to the current query features and avoid similarity calculations between irrelevant features remains an urgent problem. To address this issue, we propose a novel and effective Progressive Focused Transformer (PFT) that links all isolated attention maps in the network through Progressive Focused Attention (PFA) to focus attention on the most important tokens. PFA not only enables the network to capture more critical similar features, but also significantly reduces the computational cost of the overall network by filtering out irrelevant features before calculating similarities. Extensive experiments demonstrate the effectiveness of the proposed method, achieving state-of-the-art performance on various single image super-resolution benchmarks.
☆ Euclidean Distance to Convex Polyhedra and Application to Class Representation in Spectral Images
With the aim of estimating the abundance map from observations only, linear unmixing approaches are not always suitable to spectral images, especially when the number of bands is too small or when the spectra of the observed data are too correlated. To address this issue in the general case, we present a novel approach which provides an adapted spatial density function based on any arbitrary linear classifier. A robust mathematical formulation for computing the Euclidean distance to polyhedral sets is presented, along with an efficient algorithm that provides the exact minimum-norm point in a polyhedron. An empirical evaluation on the widely-used Samson hyperspectral dataset demonstrates that the proposed method surpasses state-of-the-art approaches in reconstructing abundance maps. Furthermore, its application to spectral images of a Lithium-ion battery, incompatible with linear unmixing models, validates the method's generality and effectiveness.
☆ Dynamic Pyramid Network for Efficient Multimodal Large Language Model
Multimodal large language models (MLLMs) have demonstrated impressive performance in various vision-language (VL) tasks, but their expensive computations still limit the real-world application. To address this issue, recent efforts aim to compress the visual features to save the computational costs of MLLMs. However, direct visual compression methods, e.g. efficient projectors, inevitably destroy the visual semantics in MLLM, especially in difficult samples. To overcome this shortcoming, we propose a novel dynamic pyramid network (DPN) for efficient MLLMs. Specifically, DPN formulates MLLM as a hierarchical structure where visual features are gradually compressed with increasing depth. In this case, even with a high compression ratio, fine-grained visual information can still be perceived in shallow layers. To maximize the benefit of DPN, we further propose an innovative Dynamic Pooling Experts (DPE) that can dynamically choose the optimal visual compression rate according to input features. With this design, harder samples will be assigned larger computations, thus preserving the model performance. To validate our approach, we conduct extensive experiments on two popular MLLMs and ten benchmarks. Experimental results show that DPN can save up to 56% average FLOPs on LLaVA while further achieving +0.74% performance gains. Besides, the generalization ability of DPN is also validated on the existing high-resolution MLLM called LLaVA-HR. Our source codes are anonymously released at https://github.com/aihao2000/DPN-LLaVA.
☆ Recovering Dynamic 3D Sketches from Videos CVPR 2025
Understanding 3D motion from videos presents inherent challenges due to the diverse types of movement, ranging from rigid and deformable objects to articulated structures. To overcome this, we propose Liv3Stroke, a novel approach for abstracting objects in motion with deformable 3D strokes. The detailed movements of an object may be represented by unstructured motion vectors or a set of motion primitives using a pre-defined articulation from a template model. Just as a free-hand sketch can intuitively visualize scenes or intentions with a sparse set of lines, we utilize a set of parametric 3D curves to capture a set of spatially smooth motion elements for general objects with unknown structures. We first extract noisy, 3D point cloud motion guidance from video frames using semantic features, and our approach deforms a set of curves to abstract essential motion features as a set of explicit 3D representations. Such abstraction enables an understanding of prominent components of motions while maintaining robustness to environmental factors. Our approach allows direct analysis of 3D object movements from video, tackling the uncertainty that typically occurs when translating real-world motion into recorded footage. The project page is accessible via: https://jaeah.me/liv3stroke_web}
comment: Accepted to CVPR 2025
☆ EditCLIP: Representation Learning for Image Editing
We introduce EditCLIP, a novel representation-learning approach for image editing. Our method learns a unified representation of edits by jointly encoding an input image and its edited counterpart, effectively capturing their transformation. To evaluate its effectiveness, we employ EditCLIP to solve two tasks: exemplar-based image editing and automated edit evaluation. In exemplar-based image editing, we replace text-based instructions in InstructPix2Pix with EditCLIP embeddings computed from a reference exemplar image pair. Experiments demonstrate that our approach outperforms state-of-the-art methods while being more efficient and versatile. For automated evaluation, EditCLIP assesses image edits by measuring the similarity between the EditCLIP embedding of a given image pair and either a textual editing instruction or the EditCLIP embedding of another reference image pair. Experiments show that EditCLIP aligns more closely with human judgments than existing CLIP-based metrics, providing a reliable measure of edit quality and structural preservation.
comment: Project page: https://qianwangx.github.io/EditCLIP/
☆ AI-Driven MRI Spine Pathology Detection: A Comprehensive Deep Learning Approach for Automated Diagnosis in Diverse Clinical Settings
Study Design This study presents the development of an autonomous AI system for MRI spine pathology detection, trained on a dataset of 2 million MRI spine scans sourced from diverse healthcare facilities across India. The AI system integrates advanced architectures, including Vision Transformers, U-Net with cross-attention, MedSAM, and Cascade R-CNN, enabling comprehensive classification, segmentation, and detection of 43 distinct spinal pathologies. The dataset is balanced across age groups, genders, and scanner manufacturers to ensure robustness and adaptability. Subgroup analyses were conducted to validate the model's performance across different patient demographics, imaging conditions, and equipment types. Performance The AI system achieved up to 97.9 percent multi-pathology detection, demonstrating consistent performance across age, gender, and manufacturer subgroups. The normal vs. abnormal classification achieved 98.0 percent accuracy, and the system was deployed across 13 major healthcare enterprises in India, encompassing diagnostic centers, large hospitals, and government facilities. During deployment, it processed approximately 100,000 plus MRI spine scans, leading to reduced reporting times and increased diagnostic efficiency by automating the identification of common spinal conditions. Conclusion The AI system's high precision and recall validate its capability as a reliable tool for autonomous normal/abnormal classification, pathology segmentation, and detection. Its scalability and adaptability address critical diagnostic gaps, optimize radiology workflows, and improve patient care across varied healthcare environments in India.
comment: 20 pages , 3 figurea
☆ SpikeDerain: Unveiling Clear Videos from Rainy Sequences Using Color Spike Streams
Restoring clear frames from rainy videos presents a significant challenge due to the rapid motion of rain streaks. Traditional frame-based visual sensors, which capture scene content synchronously, struggle to capture the fast-moving details of rain accurately. In recent years, neuromorphic sensors have introduced a new paradigm for dynamic scene perception, offering microsecond temporal resolution and high dynamic range. However, existing multimodal methods that fuse event streams with RGB images face difficulties in handling the complex spatiotemporal interference of raindrops in real scenes, primarily due to hardware synchronization errors and computational redundancy. In this paper, we propose a Color Spike Stream Deraining Network (SpikeDerain), capable of reconstructing spike streams of dynamic scenes and accurately removing rain streaks. To address the challenges of data scarcity in real continuous rainfall scenes, we design a physically interpretable rain streak synthesis model that generates parameterized continuous rain patterns based on arbitrary background images. Experimental results demonstrate that the network, trained with this synthetic data, remains highly robust even under extreme rainfall conditions. These findings highlight the effectiveness and robustness of our method across varying rainfall levels and datasets, setting new standards for video deraining tasks. The code will be released soon.
☆ Wan: Open and Advanced Large-Scale Video Generative Models
This report presents Wan, a comprehensive and open suite of video foundation models designed to push the boundaries of video generation. Built upon the mainstream diffusion transformer paradigm, Wan achieves significant advancements in generative capabilities through a series of innovations, including our novel VAE, scalable pre-training strategies, large-scale data curation, and automated evaluation metrics. These contributions collectively enhance the model's performance and versatility. Specifically, Wan is characterized by four key features: Leading Performance: The 14B model of Wan, trained on a vast dataset comprising billions of images and videos, demonstrates the scaling laws of video generation with respect to both data and model size. It consistently outperforms the existing open-source models as well as state-of-the-art commercial solutions across multiple internal and external benchmarks, demonstrating a clear and significant performance superiority. Comprehensiveness: Wan offers two capable models, i.e., 1.3B and 14B parameters, for efficiency and effectiveness respectively. It also covers multiple downstream applications, including image-to-video, instruction-guided video editing, and personal video generation, encompassing up to eight tasks. Consumer-Grade Efficiency: The 1.3B model demonstrates exceptional resource efficiency, requiring only 8.19 GB VRAM, making it compatible with a wide range of consumer-grade GPUs. Openness: We open-source the entire series of Wan, including source code and all models, with the goal of fostering the growth of the video generation community. This openness seeks to significantly expand the creative possibilities of video production in the industry and provide academia with high-quality video foundation models. All the code and models are available at https://github.com/Wan-Video/Wan2.1.
comment: 60 pages, 33 figures
☆ Enabling Heterogeneous Adversarial Transferability via Feature Permutation Attacks PAKDD 2025
Adversarial attacks in black-box settings are highly practical, with transfer-based attacks being the most effective at generating adversarial examples (AEs) that transfer from surrogate models to unseen target models. However, their performance significantly degrades when transferring across heterogeneous architectures -- such as CNNs, MLPs, and Vision Transformers (ViTs) -- due to fundamental architectural differences. To address this, we propose Feature Permutation Attack (FPA), a zero-FLOP, parameter-free method that enhances adversarial transferability across diverse architectures. FPA introduces a novel feature permutation (FP) operation, which rearranges pixel values in selected feature maps to simulate long-range dependencies, effectively making CNNs behave more like ViTs and MLPs. This enhances feature diversity and improves transferability both across heterogeneous architectures and within homogeneous CNNs. Extensive evaluations on 14 state-of-the-art architectures show that FPA achieves maximum absolute gains in attack success rates of 7.68% on CNNs, 14.57% on ViTs, and 14.48% on MLPs, outperforming existing black-box attacks. Additionally, FPA is highly generalizable and can seamlessly integrate with other transfer-based attacks to further boost their performance. Our findings establish FPA as a robust, efficient, and computationally lightweight strategy for enhancing adversarial transferability across heterogeneous architectures.
comment: PAKDD 2025. Main Track
☆ Instruction-Oriented Preference Alignment for Enhancing Multi-Modal Comprehension Capability of MLLMs
Preference alignment has emerged as an effective strategy to enhance the performance of Multimodal Large Language Models (MLLMs) following supervised fine-tuning. While existing preference alignment methods predominantly target hallucination factors, they overlook the factors essential for multi-modal comprehension capabilities, often narrowing their improvements on hallucination mitigation. To bridge this gap, we propose Instruction-oriented Preference Alignment (IPA), a scalable framework designed to automatically construct alignment preferences grounded in instruction fulfillment efficacy. Our method involves an automated preference construction coupled with a dedicated verification process that identifies instruction-oriented factors, avoiding significant variability in response representations. Additionally, IPA incorporates a progressive preference collection pipeline, further recalling challenging samples through model self-evolution and reference-guided refinement. Experiments conducted on Qwen2VL-7B demonstrate IPA's effectiveness across multiple benchmarks, including hallucination evaluation, visual question answering, and text understanding tasks, highlighting its capability to enhance general comprehension.
comment: Technical report
☆ Perceptually Accurate 3D Talking Head Generation: New Definitions, Speech-Mesh Representation, and Evaluation Metrics
Recent advancements in speech-driven 3D talking head generation have made significant progress in lip synchronization. However, existing models still struggle to capture the perceptual alignment between varying speech characteristics and corresponding lip movements. In this work, we claim that three criteria -- Temporal Synchronization, Lip Readability, and Expressiveness -- are crucial for achieving perceptually accurate lip movements. Motivated by our hypothesis that a desirable representation space exists to meet these three criteria, we introduce a speech-mesh synchronized representation that captures intricate correspondences between speech signals and 3D face meshes. We found that our learned representation exhibits desirable characteristics, and we plug it into existing models as a perceptual loss to better align lip movements to the given speech. In addition, we utilize this representation as a perceptual metric and introduce two other physically grounded lip synchronization metrics to assess how well the generated 3D talking heads align with these three criteria. Experiments show that training 3D talking head generation models with our perceptual loss significantly improve all three aspects of perceptually accurate lip synchronization. Codes and datasets are available at https://perceptual-3d-talking-head.github.io/.
3D Convolutional Neural Networks for Improved Detection of Intracranial bleeding in CT Imaging
Background: Intracranial bleeding (IB) is a life-threatening condition caused by traumatic brain injuries, including epidural, subdural, subarachnoid, and intraparenchymal hemorrhages. Rapid and accurate detection is crucial to prevent severe complications. Traditional imaging can be slow and prone to variability, especially in high-pressure scenarios. Artificial Intelligence (AI) provides a solution by quickly analyzing medical images, identifying subtle hemorrhages, and flagging urgent cases. By enhancing diagnostic speed and accuracy, AI improves workflows and patient care. This article explores AI's role in transforming IB detection in emergency settings. Methods: A U-shaped 3D Convolutional Neural Network (CNN) automates IB detection and classification in volumetric CT scans. Advanced preprocessing, including CLAHE and intensity normalization, enhances image quality. The architecture preserves spatial and contextual details for precise segmentation. A dataset of 2,912 annotated CT scans was used for training and evaluation. Results: The model achieved high performance across major bleed types, with precision, recall, and accuracy exceeding 90 percent in most cases 96 percent precision for epidural hemorrhages and 94 percent accuracy for subarachnoid hemorrhages. Its ability to classify and localize hemorrhages highlights its clinical reliability. Conclusion: This U-shaped 3D CNN offers a scalable solution for automating IB detection, reducing diagnostic delays, and improving emergency care outcomes. Future work will expand dataset diversity, optimize real-time processing, and integrate multimodal data for enhanced clinical applicability.
comment: 12 pages,4 figures
☆ Attribute-formed Class-specific Concept Space: Endowing Language Bottleneck Model with Better Interpretability and Scalability CVPR 2025
Language Bottleneck Models (LBMs) are proposed to achieve interpretable image recognition by classifying images based on textual concept bottlenecks. However, current LBMs simply list all concepts together as the bottleneck layer, leading to the spurious cue inference problem and cannot generalized to unseen classes. To address these limitations, we propose the Attribute-formed Language Bottleneck Model (ALBM). ALBM organizes concepts in the attribute-formed class-specific space, where concepts are descriptions of specific attributes for specific classes. In this way, ALBM can avoid the spurious cue inference problem by classifying solely based on the essential concepts of each class. In addition, the cross-class unified attribute set also ensures that the concept spaces of different classes have strong correlations, as a result, the learned concept classifier can be easily generalized to unseen classes. Moreover, to further improve interpretability, we propose Visual Attribute Prompt Learning (VAPL) to extract visual features on fine-grained attributes. Furthermore, to avoid labor-intensive concept annotation, we propose the Description, Summary, and Supplement (DSS) strategy to automatically generate high-quality concept sets with a complete and precise attribute. Extensive experiments on 9 widely used few-shot benchmarks demonstrate the interpretability, transferability, and performance of our approach. The code and collected concept sets are available at https://github.com/tiggers23/ALBM.
comment: This paper has been accepted to CVPR 2025
☆ Traversing Distortion-Perception Tradeoff using a Single Score-Based Generative Model
The distortion-perception (DP) tradeoff reveals a fundamental conflict between distortion metrics (e.g., MSE and PSNR) and perceptual quality. Recent research has increasingly concentrated on evaluating denoising algorithms within the DP framework. However, existing algorithms either prioritize perceptual quality by sacrificing acceptable distortion, or focus on minimizing MSE for faithful restoration. When the goal shifts or noisy measurements vary, adapting to different points on the DP plane needs retraining or even re-designing the model. Inspired by recent advances in solving inverse problems using score-based generative models, we explore the potential of flexibly and optimally traversing DP tradeoffs using a single pre-trained score-based model. Specifically, we introduce a variance-scaled reverse diffusion process and theoretically characterize the marginal distribution. We then prove that the proposed sample process is an optimal solution to the DP tradeoff for conditional Gaussian distribution. Experimental results on two-dimensional and image datasets illustrate that a single score network can effectively and flexibly traverse the DP tradeoff for general denoising problems.
comment: Accepted by IEEE/CVF Conference on Computer Vision and Pattern Recognition 2025
☆ Context-Aware Weakly Supervised Image Manipulation Localization with SAM Refinement
Malicious image manipulation poses societal risks, increasing the importance of effective image manipulation detection methods. Recent approaches in image manipulation detection have largely been driven by fully supervised approaches, which require labor-intensive pixel-level annotations. Thus, it is essential to explore weakly supervised image manipulation localization methods that only require image-level binary labels for training. However, existing weakly supervised image manipulation methods overlook the importance of edge information for accurate localization, leading to suboptimal localization performance. To address this, we propose a Context-Aware Boundary Localization (CABL) module to aggregate boundary features and learn context-inconsistency for localizing manipulated areas. Furthermore, by leveraging Class Activation Mapping (CAM) and Segment Anything Model (SAM), we introduce the CAM-Guided SAM Refinement (CGSR) module to generate more accurate manipulation localization maps. By integrating two modules, we present a novel weakly supervised framework based on a dual-branch Transformer-CNN architecture. Our method achieves outstanding localization performance across multiple datasets.
☆ CryoSAMU: Enhancing 3D Cryo-EM Density Maps of Protein Structures at Intermediate Resolution with Structure-Aware Multimodal U-Nets
Enhancing cryogenic electron microscopy (cryo-EM) 3D density maps at intermediate resolution (4-8 {\AA}) is crucial in protein structure determination. Recent advances in deep learning have led to the development of automated approaches for enhancing experimental cryo-EM density maps. Yet, these methods are not optimized for intermediate-resolution maps and rely on map density features alone. To address this, we propose CryoSAMU, a novel method designed to enhance 3D cryo-EM density maps of protein structures using structure-aware multimodal U-Nets and trained on curated intermediate-resolution density maps. We comprehensively evaluate CryoSAMU across various metrics and demonstrate its competitive performance compared to state-of-the-art methods. Notably, CryoSAMU achieves significantly faster processing speed, showing promise for future practical applications. Our code is available at https://github.com/chenwei-zhang/CryoSAMU.
comment: 18 pages, 6 main figures, 2 supplementary figures, 3 main tables, 4 supplementary tables
☆ RelTriple: Learning Plausible Indoor Layouts by Integrating Relationship Triples into the Diffusion Process
The generation of indoor furniture layouts has significant applications in augmented reality, smart homes, and architectural design. Successful furniture arrangement requires proper physical relationships (e.g., collision avoidance) and spacing relationships between furniture and their functional zones to be respected. However, manually defined relationships are almost always incomplete and can produce unrealistic layouts. This work instead extracts spacing relationships automatically based on a hierarchical analysis and adopts the Delaunay Triangulation to produce important triple relationships. Compared to pairwise relationship modeling, triple relationships account for interactions and space utilization among multiple objects. To this end, we introduce RelTriple, a novel approach that enhances furniture distribution by learning spacing relationships between objects and regions. We formulate triple relationships as object-to-object (O2O) losses and object-to-region (O2R) losses and integrate them directly into the training process of generative diffusion. Our approach consistently improves over existing state-of-the-art methods in visual results evaluation metrics on unconditional layout generation, floorplan-conditioned layout generation, and scene rearrangement, achieving at least 12% on the introduced spatial relationship metric and superior spatial coherence and practical usability.
☆ InsViE-1M: Effective Instruction-based Video Editing with Elaborate Dataset Construction
Instruction-based video editing allows effective and interactive editing of videos using only instructions without extra inputs such as masks or attributes. However, collecting high-quality training triplets (source video, edited video, instruction) is a challenging task. Existing datasets mostly consist of low-resolution, short duration, and limited amount of source videos with unsatisfactory editing quality, limiting the performance of trained editing models. In this work, we present a high-quality Instruction-based Video Editing dataset with 1M triplets, namely InsViE-1M. We first curate high-resolution and high-quality source videos and images, then design an effective editing-filtering pipeline to construct high-quality editing triplets for model training. For a source video, we generate multiple edited samples of its first frame with different intensities of classifier-free guidance, which are automatically filtered by GPT-4o with carefully crafted guidelines. The edited first frame is propagated to subsequent frames to produce the edited video, followed by another round of filtering for frame quality and motion evaluation. We also generate and filter a variety of video editing triplets from high-quality images. With the InsViE-1M dataset, we propose a multi-stage learning strategy to train our InsViE model, progressively enhancing its instruction following and editing ability. Extensive experiments demonstrate the advantages of our InsViE-1M dataset and the trained model over state-of-the-art works. Codes are available at InsViE.
☆ Faster Parameter-Efficient Tuning with Token Redundancy Reduction CVPR 2025
Parameter-efficient tuning (PET) aims to transfer pre-trained foundation models to downstream tasks by learning a small number of parameters. Compared to traditional fine-tuning, which updates the entire model, PET significantly reduces storage and transfer costs for each task regardless of exponentially increasing pre-trained model capacity. However, most PET methods inherit the inference latency of their large backbone models and often introduce additional computational overhead due to additional modules (e.g. adapters), limiting their practicality for compute-intensive applications. In this paper, we propose Faster Parameter-Efficient Tuning (FPET), a novel approach that enhances inference speed and training efficiency while maintaining high storage efficiency. Specifically, we introduce a plug-and-play token redundancy reduction module delicately designed for PET. This module refines tokens from the self-attention layer using an adapter to learn the accurate similarity between tokens and cuts off the tokens through a fully-differentiable token merging strategy, which uses a straight-through estimator for optimal token reduction. Experimental results prove that our FPET achieves faster inference and higher memory efficiency than the pre-trained backbone while keeping competitive performance on par with state-of-the-art PET methods.
comment: CVPR 2025 Camera-ready
☆ ViLBench: A Suite for Vision-Language Process Reward Modeling
Process-supervised reward models serve as a fine-grained function that provides detailed step-wise feedback to model responses, facilitating effective selection of reasoning trajectories for complex tasks. Despite its advantages, evaluation on PRMs remains less explored, especially in the multimodal domain. To address this gap, this paper first benchmarks current vision large language models (VLLMs) as two types of reward models: output reward models (ORMs) and process reward models (PRMs) on multiple vision-language benchmarks, which reveal that neither ORM nor PRM consistently outperforms across all tasks, and superior VLLMs do not necessarily yield better rewarding performance. To further advance evaluation, we introduce ViLBench, a vision-language benchmark designed to require intensive process reward signals. Notably, OpenAI's GPT-4o with Chain-of-Thought (CoT) achieves only 27.3% accuracy, indicating the benchmark's challenge for current VLLMs. Lastly, we preliminarily showcase a promising pathway towards bridging the gap between general VLLMs and reward models -- by collecting 73.6K vision-language process reward data using an enhanced tree-search algorithm, our 3B model is able to achieve an average improvement of 3.3% over standard CoT and up to 2.5% compared to its untrained counterpart on ViLBench by selecting OpenAI o1's generations. We release the implementations at https://ucsc-vlaa.github.io/ViLBench with our code, model, and data.
☆ EGVD: Event-Guided Video Diffusion Model for Physically Realistic Large-Motion Frame Interpolation
Video frame interpolation (VFI) in scenarios with large motion remains challenging due to motion ambiguity between frames. While event cameras can capture high temporal resolution motion information, existing event-based VFI methods struggle with limited training data and complex motion patterns. In this paper, we introduce Event-Guided Video Diffusion Model (EGVD), a novel framework that leverages the powerful priors of pre-trained stable video diffusion models alongside the precise temporal information from event cameras. Our approach features a Multi-modal Motion Condition Generator (MMCG) that effectively integrates RGB frames and event signals to guide the diffusion process, producing physically realistic intermediate frames. We employ a selective fine-tuning strategy that preserves spatial modeling capabilities while efficiently incorporating event-guided temporal information. We incorporate input-output normalization techniques inspired by recent advances in diffusion modeling to enhance training stability across varying noise levels. To improve generalization, we construct a comprehensive dataset combining both real and simulated event data across diverse scenarios. Extensive experiments on both real and simulated datasets demonstrate that EGVD significantly outperforms existing methods in handling large motion and challenging lighting conditions, achieving substantial improvements in perceptual quality metrics (27.4% better LPIPS on Prophesee and 24.1% on BSRGB) while maintaining competitive fidelity measures. Code and datasets available at: https://github.com/OpenImagingLab/EGVD.
☆ Mamba-3D as Masked Autoencoders for Accurate and Data-Efficient Analysis of Medical Ultrasound Videos
Ultrasound videos are an important form of clinical imaging data, and deep learning-based automated analysis can improve diagnostic accuracy and clinical efficiency. However, the scarcity of labeled data and the inherent challenges of video analysis have impeded the advancement of related methods. In this work, we introduce E-ViM$^3$, a data-efficient Vision Mamba network that preserves the 3D structure of video data, enhancing long-range dependencies and inductive biases to better model space-time correlations. With our design of Enclosure Global Tokens (EGT), the model captures and aggregates global features more effectively than competing methods. To further improve data efficiency, we employ masked video modeling for self-supervised pre-training, with the proposed Spatial-Temporal Chained (STC) masking strategy designed to adapt to various video scenarios. Experiments demonstrate that E-ViM$^3$ performs as the state-of-the-art in two high-level semantic analysis tasks across four datasets of varying sizes: EchoNet-Dynamic, CAMUS, MICCAI-BUV, and WHBUS. Furthermore, our model achieves competitive performance with limited labels, highlighting its potential impact on real-world clinical applications.
☆ LogicQA: Logical Anomaly Detection with Vision Language Model Generated Questions
Anomaly Detection (AD) focuses on detecting samples that differ from the standard pattern, making it a vital tool in process control. Logical anomalies may appear visually normal yet violate predefined constraints on object presence, arrangement, or quantity, depending on reasoning and explainability. We introduce LogicQA, a framework that enhances AD by providing industrial operators with explanations for logical anomalies. LogicQA compiles automatically generated questions into a checklist and collects responses to identify violations of logical constraints. LogicQA is training-free, annotation-free, and operates in a few-shot setting. We achieve state-of-the-art (SOTA) Logical AD performance on public benchmarks, MVTec LOCO AD, with an AUROC of 87.6 percent and an F1-max of 87.0 percent along with the explanations of anomalies. Also, our approach has shown outstanding performance on semiconductor SEM corporate data, further validating its effectiveness in industrial applications.
☆ Incremental Object Keypoint Learning CVPR
Existing progress in object keypoint estimation primarily benefits from the conventional supervised learning paradigm based on numerous data labeled with pre-defined keypoints. However, these well-trained models can hardly detect the undefined new keypoints in test time, which largely hinders their feasibility for diverse downstream tasks. To handle this, various solutions are explored but still suffer from either limited generalizability or transferability. Therefore, in this paper, we explore a novel keypoint learning paradigm in that we only annotate new keypoints in the new data and incrementally train the model, without retaining any old data, called Incremental object Keypoint Learning (IKL). A two-stage learning scheme as a novel baseline tailored to IKL is developed. In the first Knowledge Association stage, given the data labeled with only new keypoints, an auxiliary KA-Net is trained to automatically associate the old keypoints to these new ones based on their spatial and intrinsic anatomical relations. In the second Mutual Promotion stage, based on a keypoint-oriented spatial distillation loss, we jointly leverage the auxiliary KA-Net and the old model for knowledge consolidation to mutually promote the estimation of all old and new keypoints. Owing to the investigation of the correlations between new and old keypoints, our proposed method can not just effectively mitigate the catastrophic forgetting of old keypoints, but may even further improve the estimation of the old ones and achieve a positive transfer beyond anti-forgetting. Such an observation has been solidly verified by extensive experiments on different keypoint datasets, where our method exhibits superiority in alleviating the forgetting issue and boosting performance while enjoying labeling efficiency even under the low-shot data regime.
comment: The IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) 2025
☆ Unconditional Priors Matter! Improving Conditional Generation of Fine-Tuned Diffusion Models
Classifier-Free Guidance (CFG) is a fundamental technique in training conditional diffusion models. The common practice for CFG-based training is to use a single network to learn both conditional and unconditional noise prediction, with a small dropout rate for conditioning. However, we observe that the joint learning of unconditional noise with limited bandwidth in training results in poor priors for the unconditional case. More importantly, these poor unconditional noise predictions become a serious reason for degrading the quality of conditional generation. Inspired by the fact that most CFG-based conditional models are trained by fine-tuning a base model with better unconditional generation, we first show that simply replacing the unconditional noise in CFG with that predicted by the base model can significantly improve conditional generation. Furthermore, we show that a diffusion model other than the one the fine-tuned model was trained on can be used for unconditional noise replacement. We experimentally verify our claim with a range of CFG-based conditional models for both image and video generation, including Zero-1-to-3, Versatile Diffusion, DiT, DynamiCrafter, and InstructPix2Pix.
☆ Leveraging 3D Geometric Priors in 2D Rotation Symmetry Detection CVPR 2025
Symmetry plays a vital role in understanding structural patterns, aiding object recognition and scene interpretation. This paper focuses on rotation symmetry, where objects remain unchanged when rotated around a central axis, requiring detection of rotation centers and supporting vertices. Traditional methods relied on hand-crafted feature matching, while recent segmentation models based on convolutional neural networks detect rotation centers but struggle with 3D geometric consistency due to viewpoint distortions. To overcome this, we propose a model that directly predicts rotation centers and vertices in 3D space and projects the results back to 2D while preserving structural integrity. By incorporating a vertex reconstruction stage enforcing 3D geometric priors -- such as equal side lengths and interior angles -- our model enhances robustness and accuracy. Experiments on the DENDI dataset show superior performance in rotation axis detection and validate the impact of 3D priors through ablation studies.
comment: Accepted to CVPR 2025
☆ TraNCE: Transformative Non-linear Concept Explainer for CNNs
Convolutional neural networks (CNNs) have succeeded remarkably in various computer vision tasks. However, they are not intrinsically explainable. While the feature-level understanding of CNNs reveals where the models looked, concept-based explainability methods provide insights into what the models saw. However, their assumption of linear reconstructability of image activations fails to capture the intricate relationships within these activations. Their Fidelity-only approach to evaluating global explanations also presents a new concern. For the first time, we address these limitations with the novel Transformative Nonlinear Concept Explainer (TraNCE) for CNNs. Unlike linear reconstruction assumptions made by existing methods, TraNCE captures the intricate relationships within the activations. This study presents three original contributions to the CNN explainability literature: (i) An automatic concept discovery mechanism based on variational autoencoders (VAEs). This transformative concept discovery process enhances the identification of meaningful concepts from image activations. (ii) A visualization module that leverages the Bessel function to create a smooth transition between prototypical image pixels, revealing not only what the CNN saw but also what the CNN avoided, thereby mitigating the challenges of concept duplication as documented in previous works. (iii) A new metric, the Faith score, integrates both Coherence and Fidelity for a comprehensive evaluation of explainer faithfulness and consistency.
☆ TC-GS: Tri-plane based compression for 3D Gaussian Splatting ICME 2025
Recently, 3D Gaussian Splatting (3DGS) has emerged as a prominent framework for novel view synthesis, providing high fidelity and rapid rendering speed. However, the substantial data volume of 3DGS and its attributes impede its practical utility, requiring compression techniques for reducing memory cost. Nevertheless, the unorganized shape of 3DGS leads to difficulties in compression. To formulate unstructured attributes into normative distribution, we propose a well-structured tri-plane to encode Gaussian attributes, leveraging the distribution of attributes for compression. To exploit the correlations among adjacent Gaussians, K-Nearest Neighbors (KNN) is used when decoding Gaussian distribution from the Tri-plane. We also introduce Gaussian position information as a prior of the position-sensitive decoder. Additionally, we incorporate an adaptive wavelet loss, aiming to focus on the high-frequency details as iterations increase. Our approach has achieved results that are comparable to or surpass that of SOTA 3D Gaussians Splatting compression work in extensive experiments across multiple datasets. The codes are released at https://github.com/timwang2001/TC-GS.
comment: Accepted by ICME 2025
☆ DINeMo: Learning Neural Mesh Models with no 3D Annotations
Category-level 3D/6D pose estimation is a crucial step towards comprehensive 3D scene understanding, which would enable a broad range of applications in robotics and embodied AI. Recent works explored neural mesh models that approach a range of 2D and 3D tasks from an analysis-by-synthesis perspective. Despite the largely enhanced robustness to partial occlusion and domain shifts, these methods depended heavily on 3D annotations for part-contrastive learning, which confines them to a narrow set of categories and hinders efficient scaling. In this work, we present DINeMo, a novel neural mesh model that is trained with no 3D annotations by leveraging pseudo-correspondence obtained from large visual foundation models. We adopt a bidirectional pseudo-correspondence generation method, which produce pseudo correspondence utilize both local appearance features and global context information. Experimental results on car datasets demonstrate that our DINeMo outperforms previous zero- and few-shot 3D pose estimation by a wide margin, narrowing the gap with fully-supervised methods by 67.3%. Our DINeMo also scales effectively and efficiently when incorporating more unlabeled images during training, which demonstrate the advantages over supervised learning methods that rely on 3D annotations. Our project page is available at https://analysis-by-synthesis.github.io/DINeMo/.
comment: Technical report
☆ Video Motion Graphs
We present Video Motion Graphs, a system designed to generate realistic human motion videos. Using a reference video and conditional signals such as music or motion tags, the system synthesizes new videos by first retrieving video clips with gestures matching the conditions and then generating interpolation frames to seamlessly connect clip boundaries. The core of our approach is HMInterp, a robust Video Frame Interpolation (VFI) model that enables seamless interpolation of discontinuous frames, even for complex motion scenarios like dancing. HMInterp i) employs a dual-branch interpolation approach, combining a Motion Diffusion Model for human skeleton motion interpolation with a diffusion-based video frame interpolation model for final frame generation. ii) adopts condition progressive training to effectively leverage identity strong and weak conditions, such as images and pose. These designs ensure both high video texture quality and accurate motion trajectory. Results show that our Video Motion Graphs outperforms existing generative- and retrieval-based methods for multi-modal conditioned human motion video generation. Project page can be found at https://h-liu1997.github.io/Video-Motion-Graphs/
comment: 14 pages,10 figures
☆ Qwen2.5-Omni Technical Report
In this report, we present Qwen2.5-Omni, an end-to-end multimodal model designed to perceive diverse modalities, including text, images, audio, and video, while simultaneously generating text and natural speech responses in a streaming manner. To enable the streaming of multimodal information inputs, both audio and visual encoders utilize a block-wise processing approach. To synchronize the timestamps of video inputs with audio, we organize the audio and video sequentially in an interleaved manner and propose a novel position embedding approach, named TMRoPE(Time-aligned Multimodal RoPE). To concurrently generate text and speech while avoiding interference between the two modalities, we propose \textbf{Thinker-Talker} architecture. In this framework, Thinker functions as a large language model tasked with text generation, while Talker is a dual-track autoregressive model that directly utilizes the hidden representations from the Thinker to produce audio tokens as output. Both the Thinker and Talker models are designed to be trained and inferred in an end-to-end manner. For decoding audio tokens in a streaming manner, we introduce a sliding-window DiT that restricts the receptive field, aiming to reduce the initial package delay. Qwen2.5-Omni is comparable with the similarly sized Qwen2.5-VL and outperforms Qwen2-Audio. Furthermore, Qwen2.5-Omni achieves state-of-the-art performance on multimodal benchmarks like Omni-Bench. Notably, Qwen2.5-Omni's performance in end-to-end speech instruction following is comparable to its capabilities with text inputs, as evidenced by benchmarks such as MMLU and GSM8K. As for speech generation, Qwen2.5-Omni's streaming Talker outperforms most existing streaming and non-streaming alternatives in robustness and naturalness.
☆ Synthetic-to-Real Self-supervised Robust Depth Estimation via Learning with Motion and Structure Priors
Self-supervised depth estimation from monocular cameras in diverse outdoor conditions, such as daytime, rain, and nighttime, is challenging due to the difficulty of learning universal representations and the severe lack of labeled real-world adverse data. Previous methods either rely on synthetic inputs and pseudo-depth labels or directly apply daytime strategies to adverse conditions, resulting in suboptimal results. In this paper, we present the first synthetic-to-real robust depth estimation framework, incorporating motion and structure priors to capture real-world knowledge effectively. In the synthetic adaptation, we transfer motion-structure knowledge inside cost volumes for better robust representation, using a frozen daytime model to train a depth estimator in synthetic adverse conditions. In the innovative real adaptation, which targets to fix synthetic-real gaps, models trained earlier identify the weather-insensitive regions with a designed consistency-reweighting strategy to emphasize valid pseudo-labels. We introduce a new regularization by gathering explicit depth distributions to constrain the model when facing real-world data. Experiments show that our method outperforms the state-of-the-art across diverse conditions in multi-frame and single-frame evaluations. We achieve improvements of 7.5% and 4.3% in AbsRel and RMSE on average for nuScenes and Robotcar datasets (daytime, nighttime, rain). In zero-shot evaluation of DrivingStereo (rain, fog), our method generalizes better than the previous ones.
☆ BEAR: A Video Dataset For Fine-grained Behaviors Recognition Oriented with Action and Environment Factors ICME2025
Behavior recognition is an important task in video representation learning. An essential aspect pertains to effective feature learning conducive to behavior recognition. Recently, researchers have started to study fine-grained behavior recognition, which provides similar behaviors and encourages the model to concern with more details of behaviors with effective features for distinction. However, previous fine-grained behaviors limited themselves to controlling partial information to be similar, leading to an unfair and not comprehensive evaluation of existing works. In this work, we develop a new video fine-grained behavior dataset, named BEAR, which provides fine-grained (i.e. similar) behaviors that uniquely focus on two primary factors defining behavior: Environment and Action. It includes two fine-grained behavior protocols including Fine-grained Behavior with Similar Environments and Fine-grained Behavior with Similar Actions as well as multiple sub-protocols as different scenarios. Furthermore, with this new dataset, we conduct multiple experiments with different behavior recognition models. Our research primarily explores the impact of input modality, a critical element in studying the environmental and action-based aspects of behavior recognition. Our experimental results yield intriguing insights that have substantial implications for further research endeavors.
comment: Accept by ICME2025
☆ Reasoning and Learning a Perceptual Metric for Self-Training of Reflective Objects in Bin-Picking with a Low-cost Camera
Bin-picking of metal objects using low-cost RGB-D cameras often suffers from sparse depth information and reflective surface textures, leading to errors and the need for manual labeling. To reduce human intervention, we propose a two-stage framework consisting of a metric learning stage and a self-training stage. Specifically, to automatically process data captured by a low-cost camera (LC), we introduce a Multi-object Pose Reasoning (MoPR) algorithm that optimizes pose hypotheses under depth, collision, and boundary constraints. To further refine pose candidates, we adopt a Symmetry-aware Lie-group based Bayesian Gaussian Mixture Model (SaL-BGMM), integrated with the Expectation-Maximization (EM) algorithm, for symmetry-aware filtering. Additionally, we propose a Weighted Ranking Information Noise Contrastive Estimation (WR-InfoNCE) loss to enable the LC to learn a perceptual metric from reconstructed data, supporting self-training on untrained or even unseen objects. Experimental results show that our approach outperforms several state-of-the-art methods on both the ROBI dataset and our newly introduced Self-ROBI dataset.
comment: 9 pages, 10 figures
Assessing SAM for Tree Crown Instance Segmentation from Drone Imagery ICLR 2025
The potential of tree planting as a natural climate solution is often undermined by inadequate monitoring of tree planting projects. Current monitoring methods involve measuring trees by hand for each species, requiring extensive cost, time, and labour. Advances in drone remote sensing and computer vision offer great potential for mapping and characterizing trees from aerial imagery, and large pre-trained vision models, such as the Segment Anything Model (SAM), may be a particularly compelling choice given limited labeled data. In this work, we compare SAM methods for the task of automatic tree crown instance segmentation in high resolution drone imagery of young tree plantations. We explore the potential of SAM for this task, and find that methods using SAM out-of-the-box do not outperform a custom Mask R-CNN, even with well-designed prompts, but that there is potential for methods which tune SAM further. We also show that predictions can be improved by adding Digital Surface Model (DSM) information as an input.
comment: ICLR 2025 ML4RS workshop
☆ Beyond Words: Advancing Long-Text Image Generation via Multimodal Autoregressive Models
Recent advancements in autoregressive and diffusion models have led to strong performance in image generation with short scene text words. However, generating coherent, long-form text in images, such as paragraphs in slides or documents, remains a major challenge for current generative models. We present the first work specifically focused on long text image generation, addressing a critical gap in existing text-to-image systems that typically handle only brief phrases or single sentences. Through comprehensive analysis of state-of-the-art autoregressive generation models, we identify the image tokenizer as a critical bottleneck in text generating quality. To address this, we introduce a novel text-focused, binary tokenizer optimized for capturing detailed scene text features. Leveraging our tokenizer, we develop \ModelName, a multimodal autoregressive model that excels in generating high-quality long-text images with unprecedented fidelity. Our model offers robust controllability, enabling customization of text properties such as font style, size, color, and alignment. Extensive experiments demonstrate that \ModelName~significantly outperforms SD3.5 Large~\cite{sd3} and GPT4o~\cite{gpt4o} with DALL-E 3~\cite{dalle3} in generating long text accurately, consistently, and flexibly. Beyond its technical achievements, \ModelName~opens up exciting opportunities for innovative applications like interleaved document and PowerPoint generation, establishing a new frontier in long-text image generating.
comment: 16 pages
☆ Cross-Modal Prototype Allocation: Unsupervised Slide Representation Learning via Patch-Text Contrast in Computational Pathology
With the rapid advancement of pathology foundation models (FMs), the representation learning of whole slide images (WSIs) attracts increasing attention. Existing studies develop high-quality patch feature extractors and employ carefully designed aggregation schemes to derive slide-level representations. However, mainstream weakly supervised slide representation learning methods, primarily based on multiple instance learning (MIL), are tailored to specific downstream tasks, which limits their generalizability. To address this issue, some studies explore unsupervised slide representation learning. However, these approaches focus solely on the visual modality of patches, neglecting the rich semantic information embedded in textual data. In this work, we propose ProAlign, a cross-modal unsupervised slide representation learning framework. Specifically, we leverage a large language model (LLM) to generate descriptive text for the prototype types present in a WSI, introducing patch-text contrast to construct initial prototype embeddings. Furthermore, we propose a parameter-free attention aggregation strategy that utilizes the similarity between patches and these prototypes to form unsupervised slide embeddings applicable to a wide range of downstream tasks. Extensive experiments on four public datasets show that ProAlign outperforms existing unsupervised frameworks and achieves performance comparable to some weakly supervised models.
comment: 11pages,3 figures
☆ Rethinking Vision-Language Model in Face Forensics: Multi-Modal Interpretable Forged Face Detector
Deepfake detection is a long-established research topic vital for mitigating the spread of malicious misinformation. Unlike prior methods that provide either binary classification results or textual explanations separately, we introduce a novel method capable of generating both simultaneously. Our method harnesses the multi-modal learning capability of the pre-trained CLIP and the unprecedented interpretability of large language models (LLMs) to enhance both the generalization and explainability of deepfake detection. Specifically, we introduce a multi-modal face forgery detector (M2F2-Det) that employs tailored face forgery prompt learning, incorporating the pre-trained CLIP to improve generalization to unseen forgeries. Also, M2F2-Det incorporates an LLM to provide detailed textual explanations of its detection decisions, enhancing interpretability by bridging the gap between natural language and subtle cues of facial forgeries. Empirically, we evaluate M2F2-Det on both detection and explanation generation tasks, where it achieves state-of-the-art performance, demonstrating its effectiveness in identifying and explaining diverse forgeries.
comment: 8 figures; 6 tables
☆ Network Inversion for Generating Confidently Classified Counterfeits
In machine learning, especially with vision classifiers, generating inputs that are confidently classified by the model is essential for understanding its decision boundaries and behavior. However, creating such samples that are confidently classified yet distinct from the training data distribution is a challenge. Traditional methods often modify existing inputs, but they don't always ensure confident classification. In this work, we extend network inversion techniques to generate Confidently Classified Counterfeits-synthetic samples that are confidently classified by the model despite being significantly different from the training data. We achieve this by modifying the generator's conditioning mechanism from soft vector conditioning to one-hot vector conditioning and applying Kullback-Leibler divergence (KLD) between the one-hot vectors and the classifier's output distribution. This encourages the generator to produce samples that are both plausible and confidently classified. Generating Confidently Classified Counterfeits is crucial for ensuring the safety and reliability of machine learning systems, particularly in safety-critical applications where models must exhibit confidence only on data within the training distribution. By generating such counterfeits, we challenge the assumption that high-confidence predictions are always indicative of in-distribution data, providing deeper insights into the model's limitations and decision-making process.
☆ Spectrum from Defocus: Fast Spectral Imaging with Chromatic Focal Stack
Hyperspectral cameras face harsh trade-offs between spatial, spectral, and temporal resolution in an inherently low-photon regime. Computational imaging systems break through these trade-offs with compressive sensing, but require complex optics and/or extensive compute. We present Spectrum from Defocus (SfD), a chromatic focal sweep method that recovers state-of-the-art hyperspectral images with a small system of off-the-shelf optics and < 1 second of compute. Our camera uses two lenses and a grayscale sensor to preserve nearly all incident light in a chromatically-aberrated focal stack. Our physics-based iterative algorithm efficiently demixes, deconvolves, and denoises the blurry grayscale focal stack into a sharp spectral image. The combination of photon efficiency, optical simplicity, and physical modeling makes SfD a promising solution for fast, compact, interpretable hyperspectral imaging.
☆ Devil is in the Uniformity: Exploring Diverse Learners within Transformer for Image Restoration
Transformer-based approaches have gained significant attention in image restoration, where the core component, i.e, Multi-Head Attention (MHA), plays a crucial role in capturing diverse features and recovering high-quality results. In MHA, heads perform attention calculation independently from uniform split subspaces, and a redundancy issue is triggered to hinder the model from achieving satisfactory outputs. In this paper, we propose to improve MHA by exploring diverse learners and introducing various interactions between heads, which results in a Hierarchical multI-head atteNtion driven Transformer model, termed HINT, for image restoration. HINT contains two modules, i.e., the Hierarchical Multi-Head Attention (HMHA) and the Query-Key Cache Updating (QKCU) module, to address the redundancy problem that is rooted in vanilla MHA. Specifically, HMHA extracts diverse contextual features by employing heads to learn from subspaces of varying sizes and containing different information. Moreover, QKCU, comprising intra- and inter-layer schemes, further reduces the redundancy problem by facilitating enhanced interactions between attention heads within and across layers. Extensive experiments are conducted on 12 benchmarks across 5 image restoration tasks, including low-light enhancement, dehazing, desnowing, denoising, and deraining, to demonstrate the superiority of HINT. The source code is available in the supplementary materials.
comment: 11 pages, 10 figures
☆ Guiding Human-Object Interactions with Rich Geometry and Relations CVPR 2025
Human-object interaction (HOI) synthesis is crucial for creating immersive and realistic experiences for applications such as virtual reality. Existing methods often rely on simplified object representations, such as the object's centroid or the nearest point to a human, to achieve physically plausible motions. However, these approaches may overlook geometric complexity, resulting in suboptimal interaction fidelity. To address this limitation, we introduce ROG, a novel diffusion-based framework that models the spatiotemporal relationships inherent in HOIs with rich geometric detail. For efficient object representation, we select boundary-focused and fine-detail key points from the object mesh, ensuring a comprehensive depiction of the object's geometry. This representation is used to construct an interactive distance field (IDF), capturing the robust HOI dynamics. Furthermore, we develop a diffusion-based relation model that integrates spatial and temporal attention mechanisms, enabling a better understanding of intricate HOI relationships. This relation model refines the generated motion's IDF, guiding the motion generation process to produce relation-aware and semantically aligned movements. Experimental evaluations demonstrate that ROG significantly outperforms state-of-the-art methods in the realism and semantic accuracy of synthesized HOIs.
comment: CVPR 2025.Project website: https://lalalfhdh.github.io/rog_page/
☆ EVolSplat: Efficient Volume-based Gaussian Splatting for Urban View Synthesis CVPR2025
Novel view synthesis of urban scenes is essential for autonomous driving-related applications.Existing NeRF and 3DGS-based methods show promising results in achieving photorealistic renderings but require slow, per-scene optimization. We introduce EVolSplat, an efficient 3D Gaussian Splatting model for urban scenes that works in a feed-forward manner. Unlike existing feed-forward, pixel-aligned 3DGS methods, which often suffer from issues like multi-view inconsistencies and duplicated content, our approach predicts 3D Gaussians across multiple frames within a unified volume using a 3D convolutional network. This is achieved by initializing 3D Gaussians with noisy depth predictions, and then refining their geometric properties in 3D space and predicting color based on 2D textures. Our model also handles distant views and the sky with a flexible hemisphere background model. This enables us to perform fast, feed-forward reconstruction while achieving real-time rendering. Experimental evaluations on the KITTI-360 and Waymo datasets show that our method achieves state-of-the-art quality compared to existing feed-forward 3DGS- and NeRF-based methods.
comment: CVPR2025
☆ Operating Room Workflow Analysis via Reasoning Segmentation over Digital Twins
Analyzing operating room (OR) workflows to derive quantitative insights into OR efficiency is important for hospitals to maximize patient care and financial sustainability. Prior work on OR-level workflow analysis has relied on end-to-end deep neural networks. While these approaches work well in constrained settings, they are limited to the conditions specified at development time and do not offer the flexibility necessary to accommodate the OR workflow analysis needs of various OR scenarios (e.g., large academic center vs. rural provider) without data collection, annotation, and retraining. Reasoning segmentation (RS) based on foundation models offers this flexibility by enabling automated analysis of OR workflows from OR video feeds given only an implicit text query related to the objects of interest. Due to the reliance on large language model (LLM) fine-tuning, current RS approaches struggle with reasoning about semantic/spatial relationships and show limited generalization to OR video due to variations in visual characteristics and domain-specific terminology. To address these limitations, we first propose a novel digital twin (DT) representation that preserves both semantic and spatial relationships between the various OR components. Then, building on this foundation, we propose ORDiRS (Operating Room Digital twin representation for Reasoning Segmentation), an LLM-tuning-free RS framework that reformulates RS into a "reason-retrieval-synthesize" paradigm. Finally, we present ORDiRS-Agent, an LLM-based agent that decomposes OR workflow analysis queries into manageable RS sub-queries and generates responses by combining detailed textual explanations with supporting visual evidence from RS. Experimental results on both an in-house and a public OR dataset demonstrate that our ORDiRS achieves a cIoU improvement of 6.12%-9.74% compared to the existing state-of-the-arts.
☆ Reconstructing Gridded Data from Higher Autocorrelations
The higher-order autocorrelations of integer-valued or rational-valued gridded data sets appear naturally in X-ray crystallography, and have applications in computer vision systems, correlation tomography, correlation spectroscopy, and pattern recognition. In this paper, we consider the problem of reconstructing a gridded data set from its higher-order autocorrelations. We describe an explicit reconstruction algorithm, and prove that the autocorrelations up to order 3r + 3 are always sufficient to determine the data up to translation, where r is the dimension of the grid. We also provide examples of rational-valued gridded data sets which are not determined by their autocorrelations up to order 3r + 2.
comment: 13 pages, 1 figure
☆ Forensic Self-Descriptions Are All You Need for Zero-Shot Detection, Open-Set Source Attribution, and Clustering of AI-generated Images CVPR
The emergence of advanced AI-based tools to generate realistic images poses significant challenges for forensic detection and source attribution, especially as new generative techniques appear rapidly. Traditional methods often fail to generalize to unseen generators due to reliance on features specific to known sources during training. To address this problem, we propose a novel approach that explicitly models forensic microstructures - subtle, pixel-level patterns unique to the image creation process. Using only real images in a self-supervised manner, we learn a set of diverse predictive filters to extract residuals that capture different aspects of these microstructures. By jointly modeling these residuals across multiple scales, we obtain a compact model whose parameters constitute a unique forensic self-description for each image. This self-description enables us to perform zero-shot detection of synthetic images, open-set source attribution of images, and clustering based on source without prior knowledge. Extensive experiments demonstrate that our method achieves superior accuracy and adaptability compared to competing techniques, advancing the state of the art in synthetic media forensics.
comment: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) 2025
☆ MVFNet: Multipurpose Video Forensics Network using Multiple Forms of Forensic Evidence WACV
While videos can be falsified in many different ways, most existing forensic networks are specialized to detect only a single manipulation type (e.g. deepfake, inpainting). This poses a significant issue as the manipulation used to falsify a video is not known a priori. To address this problem, we propose MVFNet - a multipurpose video forensics network capable of detecting multiple types of manipulations including inpainting, deepfakes, splicing, and editing. Our network does this by extracting and jointly analyzing a broad set of forensic feature modalities that capture both spatial and temporal anomalies in falsified videos. To reliably detect and localize fake content of all shapes and sizes, our network employs a novel Multi-Scale Hierarchical Transformer module to identify forensic inconsistencies across multiple spatial scales. Experimental results show that our network obtains state-of-the-art performance in general scenarios where multiple different manipulations are possible, and rivals specialized detectors in targeted scenarios.
comment: Proceedings of the Winter Conference on Applications of Computer Vision (WACV) 2025
☆ Eyes Tell the Truth: GazeVal Highlights Shortcomings of Generative AI in Medical Imaging
The demand for high-quality synthetic data for model training and augmentation has never been greater in medical imaging. However, current evaluations predominantly rely on computational metrics that fail to align with human expert recognition. This leads to synthetic images that may appear realistic numerically but lack clinical authenticity, posing significant challenges in ensuring the reliability and effectiveness of AI-driven medical tools. To address this gap, we introduce GazeVal, a practical framework that synergizes expert eye-tracking data with direct radiological evaluations to assess the quality of synthetic medical images. GazeVal leverages gaze patterns of radiologists as they provide a deeper understanding of how experts perceive and interact with synthetic data in different tasks (i.e., diagnostic or Turing tests). Experiments with sixteen radiologists revealed that 96.6% of the generated images (by the most recent state-of-the-art AI algorithm) were identified as fake, demonstrating the limitations of generative AI in producing clinically accurate images.
☆ LATTE-MV: Learning to Anticipate Table Tennis Hits from Monocular Videos CVPR 2025
Physical agility is a necessary skill in competitive table tennis, but by no means sufficient. Champions excel in this fast-paced and highly dynamic environment by anticipating their opponent's intent - buying themselves the necessary time to react. In this work, we take one step towards designing such an anticipatory agent. Previous works have developed systems capable of real-time table tennis gameplay, though they often do not leverage anticipation. Among the works that forecast opponent actions, their approaches are limited by dataset size and variety. Our paper contributes (1) a scalable system for reconstructing monocular video of table tennis matches in 3D and (2) an uncertainty-aware controller that anticipates opponent actions. We demonstrate in simulation that our policy improves the ball return rate against high-speed hits from 49.9% to 59.0% as compared to a baseline non-anticipatory policy.
comment: CVPR 2025
☆ Prototype Guided Backdoor Defense
Deep learning models are susceptible to {\em backdoor attacks} involving malicious attackers perturbing a small subset of training data with a {\em trigger} to causes misclassifications. Various triggers have been used, including semantic triggers that are easily realizable without requiring the attacker to manipulate the image. The emergence of generative AI has eased the generation of varied poisoned samples. Robustness across types of triggers is crucial to effective defense. We propose Prototype Guided Backdoor Defense (PGBD), a robust post-hoc defense that scales across different trigger types, including previously unsolved semantic triggers. PGBD exploits displacements in the geometric spaces of activations to penalize movements toward the trigger. This is done using a novel sanitization loss of a post-hoc fine-tuning step. The geometric approach scales easily to all types of attacks. PGBD achieves better performance across all settings. We also present the first defense against a new semantic attack on celebrity face images. Project page: \hyperlink{https://venkatadithya9.github.io/pgbd.github.io/}{this https URL}.
☆ Feature Modulation for Semi-Supervised Domain Generalization without Domain Labels
Semi-supervised domain generalization (SSDG) leverages a small fraction of labeled data alongside unlabeled data to enhance model generalization. Most of the existing SSDG methods rely on pseudo-labeling (PL) for unlabeled data, often assuming access to domain labels-a privilege not always available. However, domain shifts introduce domain noise, leading to inconsistent PLs that degrade model performance. Methods derived from FixMatch suffer particularly from lower PL accuracy, reducing the effectiveness of unlabeled data. To address this, we tackle the more challenging domain-label agnostic SSDG, where domain labels for unlabeled data are not available during training. First, we propose a feature modulation strategy that enhances class-discriminative features while suppressing domain-specific information. This modulation shifts features toward Similar Average Representations-a modified version of class prototypes-that are robust across domains, encouraging the classifier to distinguish between closely related classes and feature extractor to form tightly clustered, domain-invariant representations. Second, to mitigate domain noise and improve pseudo-label accuracy, we introduce a loss-scaling function that dynamically lowers the fixed confidence threshold for pseudo-labels, optimizing the use of unlabeled data. With these key innovations, our approach achieves significant improvements on four major domain generalization benchmarks-even without domain labels. We will make the code available.
☆ BioX-CPath: Biologically-driven Explainable Diagnostics for Multistain IHC Computational Pathology CVPR 2025
The development of biologically interpretable and explainable models remains a key challenge in computational pathology, particularly for multistain immunohistochemistry (IHC) analysis. We present BioX-CPath, an explainable graph neural network architecture for whole slide image (WSI) classification that leverages both spatial and semantic features across multiple stains. At its core, BioX-CPath introduces a novel Stain-Aware Attention Pooling (SAAP) module that generates biologically meaningful, stain-aware patient embeddings. Our approach achieves state-of-the-art performance on both Rheumatoid Arthritis and Sjogren's Disease multistain datasets. Beyond performance metrics, BioX-CPath provides interpretable insights through stain attention scores, entropy measures, and stain interaction scores, that permit measuring model alignment with known pathological mechanisms. This biological grounding, combined with strong classification performance, makes BioX-CPath particularly suitable for clinical applications where interpretability is key. Source code and documentation can be found at: https://github.com/AmayaGS/BioX-CPath.
comment: Accepted for publication at CVPR 2025
☆ VinaBench: Benchmark for Faithful and Consistent Visual Narratives CVPR 2025
Visual narrative generation transforms textual narratives into sequences of images illustrating the content of the text. However, generating visual narratives that are faithful to the input text and self-consistent across generated images remains an open challenge, due to the lack of knowledge constraints used for planning the stories. In this work, we propose a new benchmark, VinaBench, to address this challenge. Our benchmark annotates the underlying commonsense and discourse constraints in visual narrative samples, offering systematic scaffolds for learning the implicit strategies of visual storytelling. Based on the incorporated narrative constraints, we further propose novel metrics to closely evaluate the consistency of generated narrative images and the alignment of generations with the input textual narrative. Our results across three generative vision models demonstrate that learning with VinaBench's knowledge constraints effectively improves the faithfulness and cohesion of generated visual narratives.
comment: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR 2025)
☆ Unified Multimodal Discrete Diffusion
Multimodal generative models that can understand and generate across multiple modalities are dominated by autoregressive (AR) approaches, which process tokens sequentially from left to right, or top to bottom. These models jointly handle images, text, video, and audio for various tasks such as image captioning, question answering, and image generation. In this work, we explore discrete diffusion models as a unified generative formulation in the joint text and image domain, building upon their recent success in text generation. Discrete diffusion models offer several advantages over AR models, including improved control over quality versus diversity of generated samples, the ability to perform joint multimodal inpainting (across both text and image domains), and greater controllability in generation through guidance. Leveraging these benefits, we present the first Unified Multimodal Discrete Diffusion (UniDisc) model which is capable of jointly understanding and generating text and images for a variety of downstream tasks. We compare UniDisc to multimodal AR models, performing a scaling analysis and demonstrating that UniDisc outperforms them in terms of both performance and inference-time compute, enhanced controllability, editability, inpainting, and flexible trade-off between inference time and generation quality. Code and additional visualizations are available at https://unidisc.github.io.
comment: Project Website: https://unidisc.github.io
☆ Generating Synthetic Data with Formal Privacy Guarantees: State of the Art and the Road Ahead
Privacy-preserving synthetic data offers a promising solution to harness segregated data in high-stakes domains where information is compartmentalized for regulatory, privacy, or institutional reasons. This survey provides a comprehensive framework for understanding the landscape of privacy-preserving synthetic data, presenting the theoretical foundations of generative models and differential privacy followed by a review of state-of-the-art methods across tabular data, images, and text. Our synthesis of evaluation approaches highlights the fundamental trade-off between utility for down-stream tasks and privacy guarantees, while identifying critical research gaps: the lack of realistic benchmarks representing specialized domains and insufficient empirical evaluations required to contextualise formal guarantees. Through empirical analysis of four leading methods on five real-world datasets from specialized domains, we demonstrate significant performance degradation under realistic privacy constraints ($\epsilon \leq 4$), revealing a substantial gap between results reported on general domain benchmarks and performance on domain-specific data. %Our findings highlight key challenges including unaccounted privacy leakage, insufficient empirical verification of formal guarantees, and a critical deficit of realistic benchmarks. These challenges underscore the need for robust evaluation frameworks, standardized benchmarks for specialized domains, and improved techniques to address the unique requirements of privacy-sensitive fields such that this technology can deliver on its considerable potential.
comment: 23 pages + references + Appendix. Preprint
♻ ☆ PhysAnimator: Physics-Guided Generative Cartoon Animation CVPR 2025
Creating hand-drawn animation sequences is labor-intensive and demands professional expertise. We introduce PhysAnimator, a novel approach for generating physically plausible meanwhile anime-stylized animation from static anime illustrations. Our method seamlessly integrates physics-based simulations with data-driven generative models to produce dynamic and visually compelling animations. To capture the fluidity and exaggeration characteristic of anime, we perform image-space deformable body simulations on extracted mesh geometries. We enhance artistic control by introducing customizable energy strokes and incorporating rigging point support, enabling the creation of tailored animation effects such as wind interactions. Finally, we extract and warp sketches from the simulation sequence, generating a texture-agnostic representation, and employ a sketch-guided video diffusion model to synthesize high-quality animation frames. The resulting animations exhibit temporal consistency and visual plausibility, demonstrating the effectiveness of our method in creating dynamic anime-style animations. See our project page for more demos: https://xpandora.github.io/PhysAnimator/
comment: Accepted by CVPR 2025
OTTER: A Vision-Language-Action Model with Text-Aware Visual Feature Extraction
Vision-Language-Action (VLA) models aim to predict robotic actions based on visual observations and language instructions. Existing approaches require fine-tuning pre-trained visionlanguage models (VLMs) as visual and language features are independently fed into downstream policies, degrading the pre-trained semantic alignments. We propose OTTER, a novel VLA architecture that leverages these existing alignments through explicit, text-aware visual feature extraction. Instead of processing all visual features, OTTER selectively extracts and passes only task-relevant visual features that are semantically aligned with the language instruction to the policy transformer. This allows OTTER to keep the pre-trained vision-language encoders frozen. Thereby, OTTER preserves and utilizes the rich semantic understanding learned from large-scale pre-training, enabling strong zero-shot generalization capabilities. In simulation and real-world experiments, OTTER significantly outperforms existing VLA models, demonstrating strong zeroshot generalization to novel objects and environments. Video, code, checkpoints, and dataset: https://ottervla.github.io/.
♻ ☆ R-LiViT: A LiDAR-Visual-Thermal Dataset Enabling Vulnerable Road User Focused Roadside Perception ICCV2025
In autonomous driving, the integration of roadside perception systems is essential for overcoming occlusion challenges and enhancing the safety of Vulnerable Road Users (VRUs). While LiDAR and visual (RGB) sensors are commonly used, thermal imaging remains underrepresented in datasets, despite its acknowledged advantages for VRU detection in extreme lighting conditions. In this paper, we present R-LiViT, the first dataset to combine LiDAR, RGB, and thermal imaging from a roadside perspective, with a strong focus on VRUs. R-LiViT captures three intersections during both day and night, ensuring a diverse dataset. It includes 10,000 LiDAR frames and 2,400 temporally and spatially aligned RGB and thermal images across over 150 traffic scenarios, with 6 and 8 annotated classes respectively, providing a comprehensive resource for tasks such as object detection and tracking. The dataset and the code for reproducing our evaluation results are made publicly available.
comment: 10 pages, 7 figures, submitted to ICCV2025
♻ ☆ DexHandDiff: Interaction-aware Diffusion Planning for Adaptive Dexterous Manipulation CVPR 2025
Dexterous manipulation with contact-rich interactions is crucial for advanced robotics. While recent diffusion-based planning approaches show promise for simple manipulation tasks, they often produce unrealistic ghost states (e.g., the object automatically moves without hand contact) or lack adaptability when handling complex sequential interactions. In this work, we introduce DexHandDiff, an interaction-aware diffusion planning framework for adaptive dexterous manipulation. DexHandDiff models joint state-action dynamics through a dual-phase diffusion process which consists of pre-interaction contact alignment and post-contact goal-directed control, enabling goal-adaptive generalizable dexterous manipulation. Additionally, we incorporate dynamics model-based dual guidance and leverage large language models for automated guidance function generation, enhancing generalizability for physical interactions and facilitating diverse goal adaptation through language cues. Experiments on physical interaction tasks such as door opening, pen and block re-orientation, object relocation, and hammer striking demonstrate DexHandDiff's effectiveness on goals outside training distributions, achieving over twice the average success rate (59.2% vs. 29.5%) compared to existing methods. Our framework achieves an average of 70.7% success rate on goal adaptive dexterous tasks, highlighting its robustness and flexibility in contact-rich manipulation.
comment: Accepted by CVPR 2025. Camera ready version. Previous DexDiffuser. Project page: https://dexdiffuser.github.io/
♻ ☆ Networking Systems for Video Anomaly Detection: A Tutorial and Survey
The increasing utilization of surveillance cameras in smart cities, coupled with the surge of online video applications, has heightened concerns regarding public security and privacy protection, which propelled automated Video Anomaly Detection (VAD) into a fundamental research task within the Artificial Intelligence (AI) community. With the advancements in deep learning and edge computing, VAD has made significant progress and advances synergized with emerging applications in smart cities and video internet, which has moved beyond the conventional research scope of algorithm engineering to deployable Networking Systems for VAD (NSVAD), a practical hotspot for intersection exploration in the AI, IoVT, and computing fields. In this article, we delineate the foundational assumptions, learning frameworks, and applicable scenarios of various deep learning-driven VAD routes, offering an exhaustive tutorial for novices in NSVAD. In addition, this article elucidates core concepts by reviewing recent advances and typical solutions and aggregating available research resources accessible at https://github.com/fdjingliu/NSVAD. Lastly, this article projects future development trends and discusses how the integration of AI and computing technologies can address existing research challenges and promote open opportunities, serving as an insightful guide for prospective researchers and engineers.
comment: Revised to ACM Computing Surveys, under review, for more information and supplementary material, please see https://github.com/fdjingliu/NSVAD
♻ ☆ Data Augmentation in Earth Observation: A Diffusion Model Approach
High-quality Earth Observation (EO) imagery is essential for accurate analysis and informed decision making across sectors. However, data scarcity caused by atmospheric conditions, seasonal variations, and limited geographical coverage hinders the effective application of Artificial Intelligence (AI) in EO. Traditional data augmentation techniques, which rely on basic parameterized image transformations, often fail to introduce sufficient diversity across key semantic axes. These axes include natural changes such as snow and floods, human impacts like urbanization and roads, and disasters such as wildfires and storms, which limits the accuracy of AI models in EO applications. To address this, we propose a four-stage data augmentation approach that integrates diffusion models to enhance semantic diversity. Our method employs meta-prompts for instruction generation, vision-language models for rich captioning, EO-specific diffusion model fine-tuning, and iterative data augmentation. Extensive experiments using four augmentation techniques demonstrate that our approach consistently outperforms established methods, generating semantically diverse EO images and improving AI model performance.
comment: 25 pages, 12 figures
♻ ☆ Harmony: A Joint Self-Supervised and Weakly-Supervised Framework for Learning General Purpose Visual Representations
Vision-language contrastive learning frameworks like CLIP enable learning representations from natural language supervision, and provide strong zero-shot classification capabilities. However, due to the nature of the supervisory signal in these paradigms, they lack the ability to learn localized features, leading to degraded performance on dense prediction tasks like segmentation and detection. On the other hand, self-supervised learning methods have shown the ability to learn granular representations, complementing the high-level features in vision-language training. In this work, we present Harmony, a framework that combines vision-language training with discriminative and generative self-supervision to learn visual features that can be generalized across different vision downstream tasks. Our framework is specifically designed to work on web-scraped data by not relying on negative examples and addressing the one-to-one correspondence issue using soft CLIP targets generated by an EMA model. We comprehensively evaluate Harmony across various vision downstream tasks and find that it significantly outperforms the baseline CLIP and the previously leading joint self and weakly-supervised methods, MaskCLIP and SLIP. Specifically, when comparing against these methods, Harmony shows superior performance in fine-tuning and zero-shot classification on ImageNet-1k, semantic segmentation on ADE20K, and both object detection and instance segmentation on MS-COCO, when pre-training a ViT-B on CC3M. We also show that Harmony outperforms other self-supervised learning methods like iBOT and MAE across all tasks evaluated. Our code is publicly at https://github.com/MohammedSB/Harmony}{https://github.com/MohammedSB/Harmony available.
comment: 22 pages, 4 figures
♻ ☆ Towards Scalable Foundation Model for Multi-modal and Hyperspectral Geospatial Data
Geospatial raster data, such as that collected by satellite-based imaging systems at different times and spectral bands, hold immense potential for enabling a wide range of high-impact applications. This potential stems from the rich information that is spatially and temporally contextualized across multiple channels and sensing modalities. Recent work has adapted existing self-supervised learning approaches for such geospatial data. However, they fall short of scalable model architectures, leading to inflexibility and computational inefficiencies when faced with an increasing number of channels and modalities. To address these limitations, we introduce Low-rank Efficient Spatial-Spectral Vision Transformer with three key innovations: i) the LESS Attention Block that approximates high-dimensional spatial-spectral attention through Kronecker's product of the low-dimensional spatial and spectral attention components; ii) the Continuous Positional-Channel Embedding Layer that preserves both the continuity and physical characteristics of each spatial-spectral patch; and iii) the Perception Field Mask that exploits local spatial dependencies by constraining attention to neighboring patches. To evaluate the proposed innovations, we construct GFM-Bench, which serves as a comprehensive benchmark for such geospatial raster data. We pretrain LESS ViT using a Hyperspectral Masked Autoencoder framework with integrated positional and channel masking strategies. Experimental results demonstrate that our proposed method achieves competitive performance against state-of-the-art multi-modal geospatial foundation models while outperforming them on cross-satellite generalization tasks with higher computational efficiency. The flexibility and extensibility of our framework make it a promising direction for future geospatial data analysis tasks that involve a wide range of modalities and channels.
♻ ☆ COSMOS: Cross-Modality Self-Distillation for Vision Language Pre-training CVPR 2025
Vision-Language Models (VLMs) trained with contrastive loss have achieved significant advancements in various vision and language tasks. However, the global nature of the contrastive loss makes VLMs focus predominantly on foreground objects, neglecting other crucial information in the image, which limits their effectiveness in downstream tasks. To address these challenges, we propose COSMOS: CrOSs-MOdality Self-distillation for vision-language pre-training that integrates a novel text-cropping strategy and cross-attention module into a self-supervised learning framework. We create global and local views of images and texts (i.e., multi-modal augmentations), which are essential for self-distillation in VLMs. We further introduce a cross-attention module, enabling COSMOS to learn comprehensive cross-modal representations optimized via a cross-modality self-distillation loss. COSMOS consistently outperforms previous strong baselines on various zero-shot downstream tasks, including retrieval, classification, and semantic segmentation. Additionally, it surpasses CLIP-based models trained on larger datasets in visual perception and contextual understanding tasks. Code is available at https://github.com/ExplainableML/cosmos.
comment: CVPR 2025
♻ ☆ The mathematics of adversarial attacks in AI -- Why deep learning is unstable despite the existence of stable neural networks
The unprecedented success of deep learning (DL) makes it unchallenged when it comes to classification problems. However, it is well established that the current DL methodology produces universally unstable neural networks (NNs). The instability problem has caused an enormous research effort -- with a vast literature on so-called adversarial attacks -- yet there has been no solution to the problem. Our paper addresses why there has been no solution to the problem, as we prove the following mathematical paradox: any training procedure based on training neural networks for classification problems with a fixed architecture will yield neural networks that are either inaccurate or unstable (if accurate) -- despite the provable existence of both accurate and stable neural networks for the same classification problems. The key is that the stable and accurate neural networks must have variable dimensions depending on the input, in particular, variable dimensions is a necessary condition for stability. Our result points towards the paradox that accurate and stable neural networks exist, however, modern algorithms do not compute them. This yields the question: if the existence of neural networks with desirable properties can be proven, can one also find algorithms that compute them? There are cases in mathematics where provable existence implies computability, but will this be the case for neural networks? The contrary is true, as we demonstrate how neural networks can provably exist as approximate minimisers to standard optimisation problems with standard cost functions, however, no randomised algorithm can compute them with probability better than 1/2.
comment: 31 pages, 1 figure. Revised to make minor changes to notation and references
♻ ☆ MC-LLaVA: Multi-Concept Personalized Vision-Language Model
Current vision-language models (VLMs) show exceptional abilities across diverse tasks, such as visual question answering. To enhance user experience, recent studies investigate VLM personalization to understand user-provided concepts. However, they mainly focus on single-concept personalization, neglecting the existence and interplay of multiple concepts, which limits real-world applicability. This paper proposes the first multi-concept personalization paradigm, MC-LLaVA. Specifically, MC-LLaVA employs a multi-concept instruction tuning strategy, effectively integrating multiple concepts in a single training step. To reduce the costs related to joint training, we propose a personalized textual prompt that uses visual token information to initialize concept tokens. Additionally, we introduce a personalized visual prompt during inference, aggregating location confidence maps for enhanced recognition and grounding capabilities. To advance multi-concept personalization research, we further contribute a high-quality instruction tuning dataset. We carefully collect images with multiple characters and objects from movies and manually generate question-answer samples for multi-concept scenarios, featuring superior diversity. Comprehensive qualitative and quantitative experiments demonstrate that MC-LLaVA can achieve impressive multi-concept personalized responses, paving the way for VLMs to become better user-specific assistants. The code and dataset will be publicly available at https://github.com/arctanxarc/MC-LLaVA.
♻ ☆ Intuitive Axial Augmentation Using Polar-Sine-Based Piecewise Distortion for Medical Slice-Wise Segmentation
Most data-driven models for medical image analysis rely on universal augmentations to improve accuracy. Experimental evidence has confirmed their effectiveness, but the unclear mechanism underlying them poses a barrier to the widespread acceptance and trust in such methods within the medical community. We revisit and acknowledge the unique characteristics of medical images apart from traditional digital images, and consequently, proposed a medical-specific augmentation algorithm that is more elastic and aligns well with radiology scan procedure. The method performs piecewise affine with sinusoidal distorted ray according to radius on polar coordinates, thus simulating uncertain postures of human lying flat on the scanning table. Our method could generate human visceral distribution without affecting the fundamental relative position on axial plane. Two non-adaptive algorithms, namely Meta-based Scan Table Removal and Similarity-Guided Parameter Search, are introduced to bolster robustness of our augmentation method. In contrast to other methodologies, our method is highlighted for its intuitive design and ease of understanding for medical professionals, thereby enhancing its applicability in clinical scenarios. Experiments show our method improves accuracy with two modality across multiple famous segmentation frameworks without requiring more data samples. Our preview code is available in: https://github.com/MGAMZ/PSBPD.
comment: Published at Smart Health
♻ ☆ 4DRGS: 4D Radiative Gaussian Splatting for Efficient 3D Vessel Reconstruction from Sparse-View Dynamic DSA Images
Reconstructing 3D vessel structures from sparse-view dynamic digital subtraction angiography (DSA) images enables accurate medical assessment while reducing radiation exposure. Existing methods often produce suboptimal results or require excessive computation time. In this work, we propose 4D radiative Gaussian splatting (4DRGS) to achieve high-quality reconstruction efficiently. In detail, we represent the vessels with 4D radiative Gaussian kernels. Each kernel has time-invariant geometry parameters, including position, rotation, and scale, to model static vessel structures. The time-dependent central attenuation of each kernel is predicted from a compact neural network to capture the temporal varying response of contrast agent flow. We splat these Gaussian kernels to synthesize DSA images via X-ray rasterization and optimize the model with real captured ones. The final 3D vessel volume is voxelized from the well-trained kernels. Moreover, we introduce accumulated attenuation pruning and bounded scaling activation to improve reconstruction quality. Extensive experiments on real-world patient data demonstrate that 4DRGS achieves impressive results in 5 minutes training, which is 32x faster than the state-of-the-art method. This underscores the potential of 4DRGS for real-world clinics.
comment: IPMI 2025 Oral; Zhentao Liu and Ruyi Zha made equal contributions
♻ ☆ DeSplat: Decomposed Gaussian Splatting for Distractor-Free Rendering
Gaussian splatting enables fast novel view synthesis in static 3D environments. However, reconstructing real-world environments remains challenging as distractors or occluders break the multi-view consistency assumption required for accurate 3D reconstruction. Most existing methods rely on external semantic information from pre-trained models, introducing additional computational overhead as pre-processing steps or during optimization. In this work, we propose a novel method, DeSplat, that directly separates distractors and static scene elements purely based on volume rendering of Gaussian primitives. We initialize Gaussians within each camera view for reconstructing the view-specific distractors to separately model the static 3D scene and distractors in the alpha compositing stages. DeSplat yields an explicit scene separation of static elements and distractors, achieving comparable results to prior distractor-free approaches without sacrificing rendering speed. We demonstrate DeSplat's effectiveness on three benchmark data sets for distractor-free novel view synthesis. See the project website at https://aaltoml.github.io/desplat/.
♻ ☆ Black-Box Forgery Attacks on Semantic Watermarks for Diffusion Models CVPR
Integrating watermarking into the generation process of latent diffusion models (LDMs) simplifies detection and attribution of generated content. Semantic watermarks, such as Tree-Rings and Gaussian Shading, represent a novel class of watermarking techniques that are easy to implement and highly robust against various perturbations. However, our work demonstrates a fundamental security vulnerability of semantic watermarks. We show that attackers can leverage unrelated models, even with different latent spaces and architectures (UNet vs DiT), to perform powerful and realistic forgery attacks. Specifically, we design two watermark forgery attacks. The first imprints a targeted watermark into real images by manipulating the latent representation of an arbitrary image in an unrelated LDM to get closer to the latent representation of a watermarked image. We also show that this technique can be used for watermark removal. The second attack generates new images with the target watermark by inverting a watermarked image and re-generating it with an arbitrary prompt. Both attacks just need a single reference image with the target watermark. Overall, our findings question the applicability of semantic watermarks by revealing that attackers can easily forge or remove these watermarks under realistic conditions.
comment: 28 pages, 22 figures, 8 tables, to be published in The IEEE/CVF Conference on Computer Vision and Pattern Recognition 2025 (CVPR)
♻ ☆ Unleashing Vecset Diffusion Model for Fast Shape Generation
3D shape generation has greatly flourished through the development of so-called "native" 3D diffusion, particularly through the Vecset Diffusion Model (VDM). While recent advancements have shown promising results in generating high-resolution 3D shapes, VDM still struggles with high-speed generation. Challenges exist because of difficulties not only in accelerating diffusion sampling but also VAE decoding in VDM, areas under-explored in previous works. To address these challenges, we present FlashVDM, a systematic framework for accelerating both VAE and DiT in VDM. For DiT, FlashVDM enables flexible diffusion sampling with as few as 5 inference steps and comparable quality, which is made possible by stabilizing consistency distillation with our newly introduced Progressive Flow Distillation. For VAE, we introduce a lightning vecset decoder equipped with Adaptive KV Selection, Hierarchical Volume Decoding, and Efficient Network Design. By exploiting the locality of the vecset and the sparsity of shape surface in the volume, our decoder drastically lowers FLOPs, minimizing the overall decoding overhead. We apply FlashVDM to Hunyuan3D-2 to obtain Hunyuan3D-2 Turbo. Through systematic evaluation, we show that our model significantly outperforms existing fast 3D generation methods, achieving comparable performance to the state-of-the-art while reducing inference time by over 45x for reconstruction and 32x for generation. Code and models are available at https://github.com/Tencent/FlashVDM.
comment: Technical report
♻ ☆ Comparison of marker-less 2D image-based methods for infant pose estimation
In this study we compare the performance of available generic- and infant-pose estimators for a video-based automated general movement assessment (GMA), and the choice of viewing angle for optimal recordings, i.e., conventional diagonal view used in GMA vs. top-down view. We used 4500 annotated video-frames from 75 recordings of infant spontaneous motor functions from 4 to 26 weeks. To determine which pose estimation method and camera angle yield the best pose estimation accuracy on infants in a GMA related setting, the distance to human annotations and the percentage of correct key-points (PCK) were computed and compared. The results show that the best performing generic model trained on adults, ViTPose, also performs best on infants. We see no improvement from using infant-pose estimators over the generic pose estimators on our infant dataset. However, when retraining a generic model on our data, there is a significant improvement in pose estimation accuracy. The pose estimation accuracy obtained from the top-down view is significantly better than that obtained from the diagonal view, especially for the detection of the hip key-points. The results also indicate limited generalization capabilities of infant-pose estimators to other infant datasets, which hints that one should be careful when choosing infant pose estimators and using them on infant datasets which they were not trained on. While the standard GMA method uses a diagonal view for assessment, pose estimation accuracy significantly improves using a top-down view. This suggests that a top-down view should be included in recording setups for automated GMA research.
♻ ☆ DashGaussian: Optimizing 3D Gaussian Splatting in 200 Seconds CVPR2025
3D Gaussian Splatting (3DGS) renders pixels by rasterizing Gaussian primitives, where the rendering resolution and the primitive number, concluded as the optimization complexity, dominate the time cost in primitive optimization. In this paper, we propose DashGaussian, a scheduling scheme over the optimization complexity of 3DGS that strips redundant complexity to accelerate 3DGS optimization. Specifically, we formulate 3DGS optimization as progressively fitting 3DGS to higher levels of frequency components in the training views, and propose a dynamic rendering resolution scheme that largely reduces the optimization complexity based on this formulation. Besides, we argue that a specific rendering resolution should cooperate with a proper primitive number for a better balance between computing redundancy and fitting quality, where we schedule the growth of the primitives to synchronize with the rendering resolution. Extensive experiments show that our method accelerates the optimization of various 3DGS backbones by 45.7% on average while preserving the rendering quality.
comment: Accepted by CVPR2025. Project page: https://dashgaussian.github.io
♻ ☆ Generating Multimodal Driving Scenes via Next-Scene Prediction CVPR 2025
Generative models in Autonomous Driving (AD) enable diverse scene creation, yet existing methods fall short by only capturing a limited range of modalities, restricting the capability of generating controllable scenes for comprehensive evaluation of AD systems. In this paper, we introduce a multimodal generation framework that incorporates four major data modalities, including a novel addition of map modality. With tokenized modalities, our scene sequence generation framework autoregressively predicts each scene while managing computational demands through a two-stage approach. The Temporal AutoRegressive (TAR) component captures inter-frame dynamics for each modality while the Ordered AutoRegressive (OAR) component aligns modalities within each scene by sequentially predicting tokens in a fixed order. To maintain coherence between map and ego-action modalities, we introduce the Action-aware Map Alignment (AMA) module, which applies a transformation based on the ego-action to maintain coherence between these modalities. Our framework effectively generates complex, realistic driving scenes over extended sequences, ensuring multimodal consistency and offering fine-grained control over scene elements. Project page: https://yanhaowu.github.io/UMGen/
comment: CVPR 2025
♻ ☆ PG-SAM: Prior-Guided SAM with Medical for Multi-organ Segmentation
Segment Anything Model (SAM) demonstrates powerful zero-shot capabilities; however, its accuracy and robustness significantly decrease when applied to medical image segmentation. Existing methods address this issue through modality fusion, integrating textual and image information to provide more detailed priors. In this study, we argue that the granularity of text and the domain gap affect the accuracy of the priors. Furthermore, the discrepancy between high-level abstract semantics and pixel-level boundary details in images can introduce noise into the fusion process. To address this, we propose Prior-Guided SAM (PG-SAM), which employs a fine-grained modality prior aligner to leverage specialized medical knowledge for better modality alignment. The core of our method lies in efficiently addressing the domain gap with fine-grained text from a medical LLM. Meanwhile, it also enhances the priors' quality after modality alignment, ensuring more accurate segmentation. In addition, our decoder enhances the model's expressive capabilities through multi-level feature fusion and iterative mask optimizer operations, supporting unprompted learning. We also propose a unified pipeline that effectively supplies high-quality semantic information to SAM. Extensive experiments on the Synapse dataset demonstrate that the proposed PG-SAM achieves state-of-the-art performance. Our code is released at https://github.com/logan-0623/PG-SAM.
♻ ☆ Scale-Equivariant Imaging: Self-Supervised Learning for Image Super-Resolution and Deblurring
Self-supervised methods have recently proved to be nearly as effective as supervised ones in various imaging inverse problems, paving the way for learning-based approaches in scientific and medical imaging applications where ground truth data is hard or expensive to obtain. These methods critically rely on invariance to translations and/or rotations of the image distribution to learn from incomplete measurement data alone. However, existing approaches fail to obtain competitive performances in the problems of image super-resolution and deblurring, which play a key role in most imaging systems. In this work, we show that invariance to roto-translations is insufficient to learn from measurements that only contain low-frequency information. Instead, we propose scale-equivariant imaging, a new self-supervised approach that leverages the fact that many image distributions are approximately scale-invariant, enabling the recovery of high-frequency information lost in the measurement process. We demonstrate throughout a series of experiments on real datasets that the proposed method outperforms other self-supervised approaches, and obtains performances on par with fully supervised learning.
♻ ☆ CLIP in Medical Imaging: A Survey
Contrastive Language-Image Pre-training (CLIP), a simple yet effective pre-training paradigm, successfully introduces text supervision to vision models. It has shown promising results across various tasks due to its generalizability and interpretability. The use of CLIP has recently gained increasing interest in the medical imaging domain, serving as a pre-training paradigm for image-text alignment, or a critical component in diverse clinical tasks. With the aim of facilitating a deeper understanding of this promising direction, this survey offers an in-depth exploration of the CLIP within the domain of medical imaging, regarding both refined CLIP pre-training and CLIP-driven applications. In this paper, we (1) first start with a brief introduction to the fundamentals of CLIP methodology; (2) then investigate the adaptation of CLIP pre-training in the medical imaging domain, focusing on how to optimize CLIP given characteristics of medical images and reports; (3) further explore practical utilization of CLIP pre-trained models in various tasks, including classification, dense prediction, and cross-modal tasks; and (4) finally discuss existing limitations of CLIP in the context of medical imaging, and propose forward-looking directions to address the demands of medical imaging domain. Studies featuring technical and practical value are both investigated. We expect this survey will provide researchers with a holistic understanding of the CLIP paradigm and its potential implications. The project page of this survey can also be found on https://github.com/zhaozh10/Awesome-CLIP-in-Medical-Imaging.
comment: Project page available at https://github.com/zhaozh10/Awesome-CLIP-in-Medical-Imaging
♻ ☆ TechCoach: Towards Technical-Point-Aware Descriptive Action Coaching
To guide a learner in mastering action skills, it is crucial for a coach to 1) reason through the learner's action execution and technical points (TechPoints), and 2) provide detailed, comprehensible feedback on what is done well and what can be improved. However, existing score-based action assessment methods are still far from reaching this practical scenario. To bridge this gap, we investigate a new task termed Descriptive Action Coaching (DescCoach) which requires the model to provide detailed commentary on what is done well and what can be improved beyond a simple quality score for action execution. To this end, we first build a new dataset named EE4D-DescCoach. Through an automatic annotation pipeline, our dataset goes beyond the existing action assessment datasets by providing detailed TechPoint-level commentary. Furthermore, we propose TechCoach, a new framework that explicitly incorporates TechPoint-level reasoning into the DescCoach process. The central to our method lies in the Context-aware TechPoint Reasoner, which enables TechCoach to learn TechPoint-related quality representation by querying visual context under the supervision of TechPoint-level coaching commentary. By leveraging the visual context and the TechPoint-related quality representation, a unified TechPoint-aware Action Assessor is then employed to provide the overall coaching commentary together with the quality score. Combining all of these, we establish a new benchmark for DescCoach and evaluate the effectiveness of our method through extensive experiments. The data and code will be made publicly available.
comment: 21 pages, 16 figures
♻ ☆ Perception of Visual Content: Differences Between Humans and Foundation Models
Human-annotated content is often used to train machine learning (ML) models. However, recently, language and multi-modal foundational models have been used to replace and scale-up human annotator's efforts. This study compares human-generated and ML-generated annotations of images representing diverse socio-economic contexts. We aim to understand differences in perception and identify potential biases in content interpretation. Our dataset comprises images of people from various geographical regions and income levels, covering various daily activities and home environments. We compare human and ML-generated annotations semantically and evaluate their impact on predictive models. Our results show highest similarity between ML captions and human labels from a low-level perspective, i.e., types of words that appear and sentence structures, but all three annotations are alike in how similar or dissimilar they perceive images across different regions. Additionally, ML Captions resulted in best overall region classification performance, while ML Objects and ML Captions performed best overall for income regression. The varying performance of annotation sets highlights the notion that all annotations are important, and that human-generated annotations are yet to be replaceable.
comment: 12 pages, 5 figures, 5 tables; updated version for a Revise-and-Resubmit at ICWSM 2025. This version includes a larger and more diverse dataset, leading to updated results
♻ ☆ Cutting Voxel Projector a New Approach to Construct 3D Cone Beam CT Operator
We introduce a novel class of projectors for 3D cone beam tomographic reconstruction. Analytical formulas are derived to compute the relationship between the volume of a voxel projected onto a detector pixel and its contribution to the line integral of attenuation recorded by that pixel. Based on these formulas, we construct a near-exact projector and backprojector, particularly suited for algebraic reconstruction techniques and hierarchical reconstruction approaches with nonuniform voxel grids. Unlike traditional projectors, which assume a uniform grid with fixed voxel sizes, our method enables local refinement of voxels, allowing for adaptive grid resolution and improved reconstruction quality in regions of interest. We have implemented this cutting voxel projector along with a relaxed, speed-optimized version and compared them to two established projectors: a ray-tracing projector based on Siddon's algorithm and a TT footprint projector. Our results demonstrate that the cutting voxel projector achieves higher accuracy than the TT projector, especially for large cone beam angles. Furthermore, the relaxed version of the cutting voxel projector offers a significant speed advantage, while maintaining comparable accuracy. In contrast, Siddon's algorithm, tuned to achieve the same accuracy, is considerably slower than the cutting voxel projector. All algorithms are implemented in a GPU optimized open-source framework for algebraic reconstruction. GitHub repository of the project https://github.com/kulvait/KCT_cbct.
comment: 12 pages, 5 figures
♻ ☆ A Survey on Event-driven 3D Reconstruction: Development under Different Categories
Event cameras have gained increasing attention for 3D reconstruction due to their high temporal resolution, low latency, and high dynamic range. They capture per-pixel brightness changes asynchronously, allowing accurate reconstruction under fast motion and challenging lighting conditions. In this survey, we provide a comprehensive review of event-driven 3D reconstruction methods, including stereo, monocular, and multimodal systems. We further categorize recent developments based on geometric, learning-based, and hybrid approaches. Emerging trends, such as neural radiance fields and 3D Gaussian splatting with event data, are also covered. The related works are structured chronologically to illustrate the innovations and progression within the field. To support future research, we also highlight key research gaps and future research directions in dataset, experiment, evaluation, event representation, etc.
comment: 6 pages, 1 figure, 6 tables, submitted to an anonymous conference under double-blind review
♻ ☆ Bayesian Modeling of Zero-Shot Classifications for Urban Flood Detection
Street scene datasets, collected from Street View or dashboard cameras, offer a promising means of detecting urban objects and incidents like street flooding. However, a major challenge in using these datasets is their lack of reliable labels: there are myriad types of incidents, many types occur rarely, and ground-truth measures of where incidents occur are lacking. Here, we propose BayFlood, a two-stage approach which circumvents this difficulty. First, we perform zero-shot classification of where incidents occur using a pretrained vision-language model (VLM). Second, we fit a spatial Bayesian model on the VLM classifications. The zero-shot approach avoids the need to annotate large training sets, and the Bayesian model provides frequent desiderata in urban settings - principled measures of uncertainty, smoothing across locations, and incorporation of external data like stormwater accumulation zones. We comprehensively validate this two-stage approach, showing that VLMs provide strong zero-shot signal for floods across multiple cities and time periods, the Bayesian model improves out-of-sample prediction relative to baseline methods, and our inferred flood risk correlates with known external predictors of risk. Having validated our approach, we show it can be used to improve urban flood detection: our analysis reveals 113,738 people who are at high risk of flooding overlooked by current methods, identifies demographic biases in existing methods, and suggests locations for new flood sensors. More broadly, our results showcase how Bayesian modeling of zero-shot LM annotations represents a promising paradigm because it avoids the need to collect large labeled datasets and leverages the power of foundation models while providing the expressiveness and uncertainty quantification of Bayesian models.
comment: In review
♻ ☆ Fantastic Copyrighted Beasts and How (Not) to Generate Them
Recent studies show that image and video generation models can be prompted to reproduce copyrighted content from their training data, raising serious legal concerns about copyright infringement. Copyrighted characters (e.g., Mario, Batman) present a significant challenge: at least one lawsuit has already awarded damages based on the generation of such characters. Consequently, commercial services like DALL-E have started deploying interventions. However, little research has systematically examined these problems: (1) Can users easily prompt models to generate copyrighted characters, even if it is unintentional?; (2) How effective are the existing mitigation strategies? To address these questions, we introduce a novel evaluation framework with metrics that assess both the generated image's similarity to copyrighted characters and its consistency with user intent, grounded in a set of popular copyrighted characters from diverse studios and regions. We show that state-of-the-art image and video generation models can still generate characters even if characters' names are not explicitly mentioned, sometimes with only two generic keywords (e.g., prompting with "videogame, plumber" consistently generates Nintendo's Mario character). We also introduce semi-automatic techniques to identify such keywords or descriptions that trigger character generation. Using this framework, we evaluate mitigation strategies, including prompt rewriting and new approaches we propose. Our findings reveal that common methods, such as DALL-E's prompt rewriting, are insufficient alone and require supplementary strategies like negative prompting. Our work provides empirical grounding for discussions on copyright mitigation strategies and offers actionable insights for model deployers implementing these safeguards.
♻ ☆ Towards End-to-End Neuromorphic Voxel-based 3D Object Reconstruction Without Physical Priors ICME
Neuromorphic cameras, also known as event cameras, are asynchronous brightness-change sensors that can capture extremely fast motion without suffering from motion blur, making them particularly promising for 3D reconstruction in extreme environments. However, existing research on 3D reconstruction using monocular neuromorphic cameras is limited, and most of the methods rely on estimating physical priors and employ complex multi-step pipelines. In this work, we propose an end-to-end method for dense voxel 3D reconstruction using neuromorphic cameras that eliminates the need to estimate physical priors. Our method incorporates a novel event representation to enhance edge features, enabling the proposed feature-enhancement model to learn more effectively. Additionally, we introduced Optimal Binarization Threshold Selection Principle as a guideline for future related work, using the optimal reconstruction results achieved with threshold optimization as the benchmark. Our method achieves a 54.6% improvement in reconstruction accuracy compared to the baseline method.
comment: 6 pages, 15 figures, 5 tables, accepted by IEEE International Conference on Multimedia & Expo (ICME) 2025
♻ ☆ Towards Real-World Test-Time Adaptation: Tri-Net Self-Training with Balanced Normalization AAAI 2024
Test-Time Adaptation aims to adapt source domain model to testing data at inference stage with success demonstrated in adapting to unseen corruptions. However, these attempts may fail under more challenging real-world scenarios. Existing works mainly consider real-world test-time adaptation under non-i.i.d. data stream and continual domain shift. In this work, we first complement the existing real-world TTA protocol with a globally class imbalanced testing set. We demonstrate that combining all settings together poses new challenges to existing methods. We argue the failure of state-of-the-art methods is first caused by indiscriminately adapting normalization layers to imbalanced testing data. To remedy this shortcoming, we propose a balanced batchnorm layer to swap out the regular batchnorm at inference stage. The new batchnorm layer is capable of adapting without biasing towards majority classes. We are further inspired by the success of self-training (ST) in learning from unlabeled data and adapt ST for test-time adaptation. However, ST alone is prone to over adaption which is responsible for the poor performance under continual domain shift. Hence, we propose to improve self-training under continual domain shift by regularizing model updates with an anchored loss. The final TTA model, termed as TRIBE, is built upon a tri-net architecture with balanced batchnorm layers. We evaluate TRIBE on four datasets representing real-world TTA settings. TRIBE consistently achieves the state-of-the-art performance across multiple evaluation protocols. The code is available at https://github.com/Gorilla-Lab-SCUT/TRIBE.
comment: Accepted by AAAI 2024. 19 pages, 7 figures and 22 tables
♻ ☆ Inference-Time Scaling for Flow Models via Stochastic Generation and Rollover Budget Forcing
We propose an inference-time scaling approach for pretrained flow models. Recently, inference-time scaling has gained significant attention in LLMs and diffusion models, improving sample quality or better aligning outputs with user preferences by leveraging additional computation. For diffusion models, particle sampling has allowed more efficient scaling due to the stochasticity at intermediate denoising steps. On the contrary, while flow models have gained popularity as an alternative to diffusion models--offering faster generation and high-quality outputs in state-of-the-art image and video generative models--efficient inference-time scaling methods used for diffusion models cannot be directly applied due to their deterministic generative process. To enable efficient inference-time scaling for flow models, we propose three key ideas: 1) SDE-based generation, enabling particle sampling in flow models, 2) Interpolant conversion, broadening the search space and enhancing sample diversity, and 3) Rollover Budget Forcing (RBF), an adaptive allocation of computational resources across timesteps to maximize budget utilization. Our experiments show that SDE-based generation, particularly variance-preserving (VP) interpolant-based generation, improves the performance of particle sampling methods for inference-time scaling in flow models. Additionally, we demonstrate that RBF with VP-SDE achieves the best performance, outperforming all previous inference-time scaling approaches.
comment: Project page: https://flow-inference-time-scaling.github.io/
♻ ☆ Aligning Visual Contrastive learning models via Preference Optimization
Contrastive learning models have demonstrated impressive abilities to capture semantic similarities by aligning representations in the embedding space. However, their performance can be limited by the quality of the training data and its inherent biases. While Preference Optimization (PO) methods such as Reinforcement Learning from Human Feedback (RLHF) and Direct Preference Optimization (DPO) have been applied to align generative models with human preferences, their use in contrastive learning has yet to be explored. This paper introduces a novel method for training contrastive learning models using different PO methods to break down complex concepts. Our method systematically aligns model behavior with desired preferences, enhancing performance on the targeted task. In particular, we focus on enhancing model robustness against typographic attacks and inductive biases, commonly seen in contrastive vision-language models like CLIP. Our experiments demonstrate that models trained using PO outperform standard contrastive learning techniques while retaining their ability to handle adversarial challenges and maintain accuracy on other downstream tasks. This makes our method well-suited for tasks requiring fairness, robustness, and alignment with specific preferences. We evaluate our method for tackling typographic attacks on images and explore its ability to disentangle gender concepts and mitigate gender bias, showcasing the versatility of our approach.
♻ ☆ MMGDreamer: Mixed-Modality Graph for Geometry-Controllable 3D Indoor Scene Generation AAAI 2025
Controllable 3D scene generation has extensive applications in virtual reality and interior design, where the generated scenes should exhibit high levels of realism and controllability in terms of geometry. Scene graphs provide a suitable data representation that facilitates these applications. However, current graph-based methods for scene generation are constrained to text-based inputs and exhibit insufficient adaptability to flexible user inputs, hindering the ability to precisely control object geometry. To address this issue, we propose MMGDreamer, a dual-branch diffusion model for scene generation that incorporates a novel Mixed-Modality Graph, visual enhancement module, and relation predictor. The mixed-modality graph allows object nodes to integrate textual and visual modalities, with optional relationships between nodes. It enhances adaptability to flexible user inputs and enables meticulous control over the geometry of objects in the generated scenes. The visual enhancement module enriches the visual fidelity of text-only nodes by constructing visual representations using text embeddings. Furthermore, our relation predictor leverages node representations to infer absent relationships between nodes, resulting in more coherent scene layouts. Extensive experimental results demonstrate that MMGDreamer exhibits superior control of object geometry, achieving state-of-the-art scene generation performance. Project page: https://yangzhifeio.github.io/project/MMGDreamer.
comment: Accepted by AAAI 2025 Main Track
♻ ☆ Unlocking the Hidden Potential of CLIP in Generalizable Deepfake Detection
This paper tackles the challenge of detecting partially manipulated facial deepfakes, which involve subtle alterations to specific facial features while retaining the overall context, posing a greater detection difficulty than fully synthetic faces. We leverage the Contrastive Language-Image Pre-training (CLIP) model, specifically its ViT-L/14 visual encoder, to develop a generalizable detection method that performs robustly across diverse datasets and unknown forgery techniques with minimal modifications to the original model. The proposed approach utilizes parameter-efficient fine-tuning (PEFT) techniques, such as LN-tuning, to adjust a small subset of the model's parameters, preserving CLIP's pre-trained knowledge and reducing overfitting. A tailored preprocessing pipeline optimizes the method for facial images, while regularization strategies, including L2 normalization and metric learning on a hyperspherical manifold, enhance generalization. Trained on the FaceForensics++ dataset and evaluated in a cross-dataset fashion on Celeb-DF-v2, DFDC, FFIW, and others, the proposed method achieves competitive detection accuracy comparable to or outperforming much more complex state-of-the-art techniques. This work highlights the efficacy of CLIP's visual encoder in facial deepfake detection and establishes a simple, powerful baseline for future research, advancing the field of generalizable deepfake detection. The code is available at: https://github.com/yermandy/deepfake-detection
♻ ☆ Point-Cache: Test-time Dynamic and Hierarchical Cache for Robust and Generalizable Point Cloud Analysis CVPR 2025
This paper proposes a general solution to enable point cloud recognition models to handle distribution shifts at test time. Unlike prior methods, which rely heavily on training data (often inaccessible during online inference) and are limited to recognizing a fixed set of point cloud classes predefined during training, we explore a more practical and challenging scenario: adapting the model solely based on online test data to recognize both previously seen classes and novel, unseen classes at test time. To this end, we develop \textbf{Point-Cache}, a hierarchical cache model that captures essential clues of online test samples, particularly focusing on the global structure of point clouds and their local-part details. Point-Cache, which serves as a rich 3D knowledge base, is dynamically managed to prioritize the inclusion of high-quality samples. Designed as a plug-and-play module, our method can be flexibly integrated into large multimodal 3D models to support open-vocabulary point cloud recognition. Notably, our solution operates with efficiency comparable to zero-shot inference, as it is entirely training-free. Point-Cache demonstrates substantial gains across 8 challenging benchmarks and 4 representative large 3D models, highlighting its effectiveness. Code is available at https://github.com/auniquesun/Point-Cache.
comment: Accepted by CVPR 2025; 24 pages, 14 figures, 18 tables
♻ ☆ MARVEL-40M+: Multi-Level Visual Elaboration for High-Fidelity Text-to-3D Content Creation
Generating high-fidelity 3D content from text prompts remains a significant challenge in computer vision due to the limited size, diversity, and annotation depth of the existing datasets. To address this, we introduce MARVEL-40M+, an extensive dataset with 40 million text annotations for over 8.9 million 3D assets aggregated from seven major 3D datasets. Our contribution is a novel multi-stage annotation pipeline that integrates open-source pretrained multi-view VLMs and LLMs to automatically produce multi-level descriptions, ranging from detailed (150-200 words) to concise semantic tags (10-20 words). This structure supports both fine-grained 3D reconstruction and rapid prototyping. Furthermore, we incorporate human metadata from source datasets into our annotation pipeline to add domain-specific information in our annotation and reduce VLM hallucinations. Additionally, we develop MARVEL-FX3D, a two-stage text-to-3D pipeline. We fine-tune Stable Diffusion with our annotations and use a pretrained image-to-3D network to generate 3D textured meshes within 15s. Extensive evaluations show that MARVEL-40M+ significantly outperforms existing datasets in annotation quality and linguistic diversity, achieving win rates of 72.41% by GPT-4 and 73.40% by human evaluators. Project page is available at https://sankalpsinha-cmos.github.io/MARVEL/.
♻ ☆ MuseTalk: Real-Time High-Fidelity Video Dubbing via Spatio-Temporal Sampling
Real-time video dubbing that preserves identity consistency while achieving accurate lip synchronization remains a critical challenge. Existing approaches face a trilemma: diffusion-based methods achieve high visual fidelity but suffer from prohibitive computational costs, while GAN-based solutions sacrifice lip-sync accuracy or dental details for real-time performance. We present MuseTalk, a novel two-stage training framework that resolves this trade-off through latent space optimization and spatio-temporal data sampling strategy. Our key innovations include: (1) During the Facial Abstract Pretraining stage, we propose Informative Frame Sampling to temporally align reference-source pose pairs, eliminating redundant feature interference while preserving identity cues. (2) In the Lip-Sync Adversarial Finetuning stage, we employ Dynamic Margin Sampling to spatially select the most suitable lip-movement-promoting regions, balancing audio-visual synchronization and dental clarity. (3) MuseTalk establishes an effective audio-visual feature fusion framework in the latent space, delivering 30 FPS output at 256*256 resolution on an NVIDIA V100 GPU. Extensive experiments demonstrate that MuseTalk outperforms state-of-the-art methods in visual fidelity while achieving comparable lip-sync accuracy. %The codes and models will be made publicly available upon acceptance. The code is made available at \href{https://github.com/TMElyralab/MuseTalk}{https://github.com/TMElyralab/MuseTalk}
comment: 15 pages, 4 figures
♻ ☆ PHT-CAD: Efficient CAD Parametric Primitive Analysis with Progressive Hierarchical Tuning
Computer-Aided Design (CAD) plays a pivotal role in industrial manufacturing, yet 2D Parametric Primitive Analysis (PPA) remains underexplored due to two key challenges: structural constraint reasoning and advanced semantic understanding. To tackle these challenges, we first propose an Efficient Hybrid Parametrization (EHP) for better representing 2D engineering drawings. EHP contains four types of atomic component i.e., point, line, circle, and arc). Additionally, we propose PHT-CAD, a novel 2D PPA framework that harnesses the modality alignment and reasoning capabilities of Vision-Language Models (VLMs) for precise engineering drawing analysis. In PHT-CAD, we introduce four dedicated regression heads to predict corresponding atomic components. To train PHT-CAD, a three-stage training paradigm Progressive Hierarchical Tuning (PHT) is proposed to progressively enhance PHT-CAD's capability to perceive individual primitives, infer structural constraints, and align annotation layers with their corresponding geometric representations. Considering that existing datasets lack complete annotation layers and real-world engineering drawings, we introduce ParaCAD, the first large-scale benchmark that explicitly integrates both the geometric and annotation layers. ParaCAD comprises over 10 million annotated drawings for training and 3,000 real-world industrial drawings with complex topological structures and physical constraints for test. Extensive experiments demonstrate the effectiveness of PHT-CAD and highlight the practical significance of ParaCAD in advancing 2D PPA research.
♻ ☆ DEIM: DETR with Improved Matching for Fast Convergence CVPR 2025
We introduce DEIM, an innovative and efficient training framework designed to accelerate convergence in real-time object detection with Transformer-based architectures (DETR). To mitigate the sparse supervision inherent in one-to-one (O2O) matching in DETR models, DEIM employs a Dense O2O matching strategy. This approach increases the number of positive samples per image by incorporating additional targets, using standard data augmentation techniques. While Dense O2O matching speeds up convergence, it also introduces numerous low-quality matches that could affect performance. To address this, we propose the Matchability-Aware Loss (MAL), a novel loss function that optimizes matches across various quality levels, enhancing the effectiveness of Dense O2O. Extensive experiments on the COCO dataset validate the efficacy of DEIM. When integrated with RT-DETR and D-FINE, it consistently boosts performance while reducing training time by 50%. Notably, paired with RT-DETRv2, DEIM achieves 53.2% AP in a single day of training on an NVIDIA 4090 GPU. Additionally, DEIM-trained real-time models outperform leading real-time object detectors, with DEIM-D-FINE-L and DEIM-D-FINE-X achieving 54.7% and 56.5% AP at 124 and 78 FPS on an NVIDIA T4 GPU, respectively, without the need for additional data. We believe DEIM sets a new baseline for advancements in real-time object detection. Our code and pre-trained models are available at https://github.com/ShihuaHuang95/DEIM.
comment: CVPR 2025
♻ ☆ Decoupling Fine Detail and Global Geometry for Compressed Depth Map Super-Resolution CVPR 2025
Recovering high-quality depth maps from compressed sources has gained significant attention due to the limitations of consumer-grade depth cameras and the bandwidth restrictions during data transmission. However, current methods still suffer from two challenges. First, bit-depth compression produces a uniform depth representation in regions with subtle variations, hindering the recovery of detailed information. Second, densely distributed random noise reduces the accuracy of estimating the global geometric structure of the scene. To address these challenges, we propose a novel framework, termed geometry-decoupled network (GDNet), for compressed depth map super-resolution that decouples the high-quality depth map reconstruction process by handling global and detailed geometric features separately. To be specific, we propose the fine geometry detail encoder (FGDE), which is designed to aggregate fine geometry details in high-resolution low-level image features while simultaneously enriching them with complementary information from low-resolution context-level image features. In addition, we develop the global geometry encoder (GGE) that aims at suppressing noise and extracting global geometric information effectively via constructing compact feature representation in a low-rank space. We conduct experiments on multiple benchmark datasets, demonstrating that our GDNet significantly outperforms current methods in terms of geometric consistency and detail recovery. In the ECCV 2024 AIM Compressed Depth Upsampling Challenge, our solution won the 1st place award. Our codes are available at: https://github.com/Ian0926/GDNet.
comment: Accepted by CVPR 2025 & The 1st place award for the ECCV 2024 AIM Compressed Depth Upsampling Challenge
♻ ☆ NLPrompt: Noise-Label Prompt Learning for Vision-Language Models
The emergence of vision-language foundation models, such as CLIP, has revolutionized image-text representation, enabling a broad range of applications via prompt learning. Despite its promise, real-world datasets often contain noisy labels that can degrade prompt learning performance. In this paper, we demonstrate that using mean absolute error (MAE) loss in prompt learning, named PromptMAE, significantly enhances robustness against noisy labels while maintaining high accuracy. Though MAE is straightforward and recognized for its robustness, it is rarely used in noisy-label learning due to its slow convergence and poor performance outside prompt learning scenarios. To elucidate the robustness of PromptMAE, we leverage feature learning theory to show that MAE can suppress the influence of noisy samples, thereby improving the signal-to-noise ratio and enhancing overall robustness. Additionally, we introduce PromptOT, a prompt-based optimal transport data purification method to enhance the robustness further. PromptOT employs text features in vision-language models as prototypes to construct an optimal transportation matrix. This matrix effectively partitions datasets into clean and noisy subsets, allowing for the application of cross-entropy loss to the clean subset and MAE loss to the noisy subset. Our Noise-Label Prompt Learning method, named NLPrompt, offers a simple and efficient approach that leverages the expressive representations and precise alignment capabilities of vision-language models for robust prompt learning. We validate NLPrompt through extensive experiments across various noise settings, demonstrating significant performance improvements.
♻ ☆ MozzaVID: Mozzarella Volumetric Image Dataset
Influenced by the complexity of volumetric imaging, there is a shortage of established datasets useful for benchmarking volumetric deep-learning models. As a consequence, new and existing models are not easily comparable, limiting the development of architectures optimized specifically for volumetric data. To counteract this trend, we introduce MozzaVID - a large, clean, and versatile volumetric classification dataset. Our dataset contains X-ray computed tomography (CT) images of mozzarella microstructure and enables the classification of 25 cheese types and 149 cheese samples. We provide data in three different resolutions, resulting in three dataset instances containing from 591 to 37,824 images. While being general-purpose, the dataset also facilitates investigating mozzarella structure properties. The structure of food directly affects its functional properties and thus its consumption experience. Understanding food structure helps tune the production and mimicking it enables sustainable alternatives to animal-derived food products. The complex and disordered nature of food structures brings a unique challenge, where a choice of appropriate imaging method, scale, and sample size is not trivial. With this dataset we aim to address these complexities, contributing to more robust structural analysis models. The dataset can be downloaded from: https://archive.compute.dtu.dk/files/public/projects/MozzaVID/.
♻ ☆ ONER: Online Experience Replay for Incremental Anomaly Detection
Incremental anomaly detection aims to sequentially identify defects in industrial product lines but suffers from catastrophic forgetting, primarily due to knowledge overwriting during parameter updates and feature conflicts between tasks. In this work, We propose ONER (ONline Experience Replay), an end-to-end framework that addresses these issues by synergistically integrating two types of experience: (1) decomposed prompts, which dynamically generate image-conditioned prompts from reusable modules to retain prior knowledge thus prevent knowledge overwriting, and (2) semantic prototypes, which enforce separability in latent feature spaces at pixel and image levels to mitigate cross-task feature conflicts. Extensive experiments demonstrate the superiority of ONER, achieving state-of-the-art performance with +4.4% Pixel AUROC and +28.3% Pixel AUPR improvements on the MVTec AD dataset over prior methods. Remarkably, ONER achieves this with only 0.019M parameters and 5 training epochs per task, confirming its efficiency and stability for real-world industrial deployment.
♻ ☆ DiTCtrl: Exploring Attention Control in Multi-Modal Diffusion Transformer for Tuning-Free Multi-Prompt Longer Video Generation CVPR 2025
Sora-like video generation models have achieved remarkable progress with a Multi-Modal Diffusion Transformer MM-DiT architecture. However, the current video generation models predominantly focus on single-prompt, struggling to generate coherent scenes with multiple sequential prompts that better reflect real-world dynamic scenarios. While some pioneering works have explored multi-prompt video generation, they face significant challenges including strict training data requirements, weak prompt following, and unnatural transitions. To address these problems, we propose DiTCtrl, a training-free multi-prompt video generation method under MM-DiT architectures for the first time. Our key idea is to take the multi-prompt video generation task as temporal video editing with smooth transitions. To achieve this goal, we first analyze MM-DiT's attention mechanism, finding that the 3D full attention behaves similarly to that of the cross/self-attention blocks in the UNet-like diffusion models, enabling mask-guided precise semantic control across different prompts with attention sharing for multi-prompt video generation. Based on our careful design, the video generated by DiTCtrl achieves smooth transitions and consistent object motion given multiple sequential prompts without additional training. Besides, we also present MPVBench, a new benchmark specially designed for multi-prompt video generation to evaluate the performance of multi-prompt generation. Extensive experiments demonstrate that our method achieves state-of-the-art performance without additional training.
comment: CVPR 2025; 21 pages, 23 figures, Project page: https://onevfall.github.io/project_page/ditctrl ; GitHub repository: https://github.com/TencentARC/DiTCtrl
♻ ☆ Oasis: One Image is All You Need for Multimodal Instruction Data Synthesis
The success of multi-modal large language models (MLLMs) has been largely attributed to the large-scale training data. However, the training data of many MLLMs is unavailable due to privacy concerns. The expensive and labor-intensive process of collecting multi-modal data further exacerbates the problem. Is it possible to synthesize multi-modal training data automatically without compromising diversity and quality? In this paper, we propose a new method, Oasis, to synthesize high-quality multi-modal data with only images. Oasis breaks through traditional methods by prompting only images to the MLLMs, thus extending the data diversity by a large margin. Our method features a delicate quality control method which ensures the data quality. We collected over 500k data and conducted incremental experiments on LLaVA-NeXT. Extensive experiments demonstrate that our method can significantly improve the performance of MLLMs. The image-based synthesis also allows us to focus on the specific-domain ability of MLLMs. Code and dataset are publicly available at https://github.com/Letian2003/MM_INF.
♻ ☆ Referring Video Object Segmentation via Language-aligned Track Selection
Referring video object segmentation (RVOS) requires tracking and segmenting an object throughout a video according to a given natural language expression, demanding both complex motion understanding and the alignment of visual representations with language descriptions. Given these challenges, the recently proposed Segment Anything Model 2 (SAM2) emerges as a potential candidate due to its ability to generate coherent segmentation mask tracks across video frames, and provide an inherent spatio-temporal objectness in its object token representations. In this paper, we introduce SOLA (Selection by Object Language Alignment), a novel framework that leverages SAM2 object tokens as compact video-level object representations, which are aligned with language features through a lightweight track selection module. To effectively facilitate this alignment, we propose an IoU-based pseudo-labeling strategy, which bridges the modality gap between SAM2 representations with language features. Extensive experiments show that SOLA achieves state-of-the-art performance on the MeViS dataset and demonstrate that SOLA offers an effective solution for RVOS. Our project page is available at: https://cvlab-kaist.github.io/SOLA.
comment: Project page is available at https://cvlab-kaist.github.io/SOLA
♻ ☆ Unleashing HyDRa: Hybrid Fusion, Depth Consistency and Radar for Unified 3D Perception ICRA 2025
Low-cost, vision-centric 3D perception systems for autonomous driving have made significant progress in recent years, narrowing the gap to expensive LiDAR-based methods. The primary challenge in becoming a fully reliable alternative lies in robust depth prediction capabilities, as camera-based systems struggle with long detection ranges and adverse lighting and weather conditions. In this work, we introduce HyDRa, a novel camera-radar fusion architecture for diverse 3D perception tasks. Building upon the principles of dense BEV (Bird's Eye View)-based architectures, HyDRa introduces a hybrid fusion approach to combine the strengths of complementary camera and radar features in two distinct representation spaces. Our Height Association Transformer module leverages radar features already in the perspective view to produce more robust and accurate depth predictions. In the BEV, we refine the initial sparse representation by a Radar-weighted Depth Consistency. HyDRa achieves a new state-of-the-art for camera-radar fusion of 64.2 NDS (+1.8) and 58.4 AMOTA (+1.5) on the public nuScenes dataset. Moreover, our new semantically rich and spatially accurate BEV features can be directly converted into a powerful occupancy representation, beating all previous camera-based methods on the Occ3D benchmark by an impressive 3.7 mIoU. Code and models are available at https://github.com/phi-wol/hydra.
comment: Accepted to ICRA 2025
♻ ☆ ARFlow: Human Action-Reaction Flow Matching with Physical Guidance
Human action-reaction synthesis, a fundamental challenge in modeling causal human interactions, plays a critical role in applications ranging from virtual reality to social robotics. While diffusion-based models have demonstrated promising performance, they exhibit two key limitations for interaction synthesis: reliance on complex noise-to-reaction generators with intricate conditional mechanisms, and frequent physical violations in generated motions. To address these issues, we propose Action-Reaction Flow Matching (ARFlow), a novel framework that establishes direct action-to-reaction mappings, eliminating the need for complex conditional mechanisms. Our approach introduces two key innovations: an x1-prediction method that directly outputs human motions instead of velocity fields, enabling explicit constraint enforcement; and a training-free, gradient-based physical guidance mechanism that effectively prevents body penetration artifacts during sampling. Extensive experiments on NTU120 and Chi3D datasets demonstrate that ARFlow not only outperforms existing methods in terms of Fr\'echet Inception Distance and motion diversity but also significantly reduces body collisions, as measured by our new Intersection Volume and Intersection Frequency metrics.
comment: Project Page: https://arflow2025.github.io/
♻ ☆ CogVideoX: Text-to-Video Diffusion Models with An Expert Transformer ICLR2025
We present CogVideoX, a large-scale text-to-video generation model based on diffusion transformer, which can generate 10-second continuous videos aligned with text prompt, with a frame rate of 16 fps and resolution of 768 * 1360 pixels. Previous video generation models often had limited movement and short durations, and is difficult to generate videos with coherent narratives based on text. We propose several designs to address these issues. First, we propose a 3D Variational Autoencoder (VAE) to compress videos along both spatial and temporal dimensions, to improve both compression rate and video fidelity. Second, to improve the text-video alignment, we propose an expert transformer with the expert adaptive LayerNorm to facilitate the deep fusion between the two modalities. Third, by employing a progressive training and multi-resolution frame pack technique, CogVideoX is adept at producing coherent, long-duration, different shape videos characterized by significant motions. In addition, we develop an effective text-video data processing pipeline that includes various data preprocessing strategies and a video captioning method, greatly contributing to the generation quality and semantic alignment. Results show that CogVideoX demonstrates state-of-the-art performance across both multiple machine metrics and human evaluations. The model weight of both 3D Causal VAE, Video caption model and CogVideoX are publicly available at https://github.com/THUDM/CogVideo.
comment: Accepted by ICLR2025
♻ ☆ TopoBDA: Towards Bezier Deformable Attention for Road Topology Understanding
Understanding road topology is crucial for autonomous driving. This paper introduces TopoBDA (Topology with Bezier Deformable Attention), a novel approach that enhances road topology comprehension by leveraging Bezier Deformable Attention (BDA). TopoBDA processes multi-camera 360-degree imagery to generate Bird's Eye View (BEV) features, which are refined through a transformer decoder employing BDA. BDA utilizes Bezier control points to drive the deformable attention mechanism, improving the detection and representation of elongated and thin polyline structures, such as lane centerlines. Additionally, TopoBDA integrates two auxiliary components: an instance mask formulation loss and a one-to-many set prediction loss strategy, to further refine centerline detection and enhance road topology understanding. Experimental evaluations on the OpenLane-V2 dataset demonstrate that TopoBDA outperforms existing methods, achieving state-of-the-art results in centerline detection and topology reasoning. TopoBDA also achieves the best results on the OpenLane-V1 dataset in 3D lane detection. Further experiments on integrating multi-modal data -- such as LiDAR, radar, and SDMap -- show that multimodal inputs can further enhance performance in road topology understanding.
♻ ☆ HumanDiT: Pose-Guided Diffusion Transformer for Long-form Human Motion Video Generation
Human motion video generation has advanced significantly, while existing methods still struggle with accurately rendering detailed body parts like hands and faces, especially in long sequences and intricate motions. Current approaches also rely on fixed resolution and struggle to maintain visual consistency. To address these limitations, we propose HumanDiT, a pose-guided Diffusion Transformer (DiT)-based framework trained on a large and wild dataset containing 14,000 hours of high-quality video to produce high-fidelity videos with fine-grained body rendering. Specifically, (i) HumanDiT, built on DiT, supports numerous video resolutions and variable sequence lengths, facilitating learning for long-sequence video generation; (ii) we introduce a prefix-latent reference strategy to maintain personalized characteristics across extended sequences. Furthermore, during inference, HumanDiT leverages Keypoint-DiT to generate subsequent pose sequences, facilitating video continuation from static images or existing videos. It also utilizes a Pose Adapter to enable pose transfer with given sequences. Extensive experiments demonstrate its superior performance in generating long-form, pose-accurate videos across diverse scenarios.
comment: https://agnjason.github.io/HumanDiT-page/
♻ ☆ Socratic Planner: Self-QA-Based Zero-Shot Planning for Embodied Instruction Following ICRA 2025
Embodied Instruction Following (EIF) is the task of executing natural language instructions by navigating and interacting with objects in interactive environments. A key challenge in EIF is compositional task planning, typically addressed through supervised learning or few-shot in-context learning with labeled data. To this end, we introduce the Socratic Planner, a self-QA-based zero-shot planning method that infers an appropriate plan without any further training. The Socratic Planner first facilitates self-questioning and answering by the Large Language Model (LLM), which in turn helps generate a sequence of subgoals. While executing the subgoals, an embodied agent may encounter unexpected situations, such as unforeseen obstacles. The Socratic Planner then adjusts plans based on dense visual feedback through a visually-grounded re-planning mechanism. Experiments demonstrate the effectiveness of the Socratic Planner, outperforming current state-of-the-art planning models on the ALFRED benchmark across all metrics, particularly excelling in long-horizon tasks that demand complex inference. We further demonstrate its real-world applicability through deployment on a physical robot for long-horizon tasks.
comment: 8 pages, 6 figures, published to ICRA 2025
♻ ☆ One Framework to Rule Them All: Unifying RL-Based and RL-Free Methods in RLHF
In this article, we primarily examine a variety of RL-based and RL-free methods designed to address Reinforcement Learning from Human Feedback (RLHF) and Large Reasoning Models (LRMs). We begin with a concise overview of the typical steps involved in RLHF and LRMs. Next, we reinterpret several RL-based and RL-free algorithms through the perspective of neural structured bandit prediction, providing a clear conceptual framework that uncovers a deeper connection between these seemingly distinct approaches. Following this, we briefly review some core principles of reinforcement learning, drawing attention to an often-overlooked aspect in existing RLHF studies. This leads to a detailed derivation of the standard RLHF objective within a full RL context, demonstrating its equivalence to neural structured bandit prediction. Finally, by reinvestigating the principles behind Proximal Policy Optimization (PPO), we pinpoint areas needing adjustment, which culminates in the introduction of the Generalized Reinforce Optimization (GRO) framework, seamlessly integrating RL-based and RL-free methods in RLHF. We look forward to the community's efforts to empirically validate GRO and invite constructive feedback.
♻ ☆ MMRL: Multi-Modal Representation Learning for Vision-Language Models CVPR 2025
Large-scale pre-trained Vision-Language Models (VLMs) have become essential for transfer learning across diverse tasks. However, adapting these models with limited few-shot data often leads to overfitting, diminishing their performance on new tasks. To tackle this issue, we propose a novel Multi-Modal Representation Learning (MMRL) framework that introduces a shared, learnable, and modality-agnostic representation space. MMRL projects the space tokens to text and image representation tokens, facilitating more effective multi-modal interactions. Unlike previous approaches that solely optimize class token features, MMRL integrates representation tokens at higher layers of the encoders--where dataset-specific features are more prominent--while preserving generalized knowledge in the lower layers. During training, both representation and class features are optimized, with trainable projection layer applied to the representation tokens, whereas the class token projection layer remains frozen to retain pre-trained knowledge. Furthermore, a regularization term is introduced to align the class features and text features with the zero-shot features from the frozen VLM, thereby safeguarding the model's generalization capacity. For inference, a decoupling strategy is employed, wherein both representation and class features are utilized for base classes, while only the class features, which retain more generalized knowledge, are used for new tasks. Extensive experiments across 15 datasets demonstrate that MMRL outperforms state-of-the-art methods, achieving a balanced trade-off between task-specific adaptation and generalization. Code is available at https://github.com/yunncheng/MMRL.
comment: Accepted by CVPR 2025
♻ ☆ RoboSpatial: Teaching Spatial Understanding to 2D and 3D Vision-Language Models for Robotics CVPR 2025
Spatial understanding is a crucial capability that enables robots to perceive their surroundings, reason about their environment, and interact with it meaningfully. In modern robotics, these capabilities are increasingly provided by vision-language models. However, these models face significant challenges in spatial reasoning tasks, as their training data are based on general-purpose image datasets that often lack sophisticated spatial understanding. For example, datasets frequently do not capture reference frame comprehension, yet effective spatial reasoning requires understanding whether to reason from ego-, world-, or object-centric perspectives. To address this issue, we introduce RoboSpatial, a large-scale dataset for spatial understanding in robotics. It consists of real indoor and tabletop scenes, captured as 3D scans and egocentric images, and annotated with rich spatial information relevant to robotics. The dataset includes 1M images, 5k 3D scans, and 3M annotated spatial relationships, and the pairing of 2D egocentric images with 3D scans makes it both 2D- and 3D- ready. Our experiments show that models trained with RoboSpatial outperform baselines on downstream tasks such as spatial affordance prediction, spatial relationship prediction, and robot manipulation.
comment: CVPR 2025
♻ ☆ TopV-Nav: Unlocking the Top-View Spatial Reasoning Potential of MLLM for Zero-shot Object Navigation
The Zero-Shot Object Navigation (ZSON) task requires embodied agents to find a previously unseen object by navigating in unfamiliar environments. Such a goal-oriented exploration heavily relies on the ability to perceive, understand, and reason based on the spatial information of the environment. However, current LLM-based approaches convert visual observations to language descriptions and reason in the linguistic space, leading to the loss of spatial information. In this paper, we introduce TopV-Nav, an MLLM-based method that directly reasons on the top-view map with sufficient spatial information. To fully unlock the MLLM's spatial reasoning potential in top-view perspective, we propose the Adaptive Visual Prompt Generation (AVPG) method to adaptively construct semantically-rich top-view map. It enables the agent to directly utilize spatial information contained in the top-view map to conduct thorough reasoning. Besides, we design a Dynamic Map Scaling (DMS) mechanism to dynamically zoom top-view map at preferred scales, enhancing local fine-grained reasoning. Additionally, we devise a Potential Target Driven (PTD) mechanism to predict and to utilize target locations, facilitating global and human-like exploration. Experiments on MP3D and HM3D datasets demonstrate the superiority of our TopV-Nav.
comment: 10 pages
♻ ☆ Fine-Grained Domain Generalization with Feature Structuralization
Fine-grained domain generalization (FGDG) is a more challenging task than traditional DG tasks due to its small inter-class variations and relatively large intra-class disparities. When domain distribution changes, the vulnerability of subtle features leads to a severe deterioration in model performance. Nevertheless, humans inherently demonstrate the capacity for generalizing to out-of-distribution data, leveraging structured multi-granularity knowledge that emerges from discerning the commonality and specificity within categories. Likewise, we propose a Feature Structuralized Domain Generalization (FSDG) model, wherein features experience structuralization into common, specific, and confounding segments, harmoniously aligned with their relevant semantic concepts, to elevate performance in FGDG. Specifically, feature structuralization (FS) is accomplished through joint optimization of five constraints: a decorrelation function applied to disentangled segments, three constraints ensuring common feature consistency and specific feature distinctiveness, and a prediction calibration term. By imposing these stipulations, FSDG is prompted to disentangle and align features based on multi-granularity knowledge, facilitating robust subtle distinctions among categories. Extensive experimentation on three benchmarks consistently validates the superiority of FSDG over state-of-the-art counterparts, with an average improvement of 6.2% in FGDG performance. Beyond that, the explainability analysis on explicit concept matching intensity between the shared concepts among categories and the model channels, along with experiments on various mainstream model architectures, substantiates the validity of FS.
♻ ☆ OverLoCK: An Overview-first-Look-Closely-next ConvNet with Context-Mixing Dynamic Kernels CVPR 2025
Top-down attention plays a crucial role in the human vision system, wherein the brain initially obtains a rough overview of a scene to discover salient cues (i.e., overview first), followed by a more careful finer-grained examination (i.e., look closely next). However, modern ConvNets remain confined to a pyramid structure that successively downsamples the feature map for receptive field expansion, neglecting this crucial biomimetic principle. We present OverLoCK, the first pure ConvNet backbone architecture that explicitly incorporates a top-down attention mechanism. Unlike pyramid backbone networks, our design features a branched architecture with three synergistic sub-networks: 1) a Base-Net that encodes low/mid-level features; 2) a lightweight Overview-Net that generates dynamic top-down attention through coarse global context modeling (i.e., overview first); and 3) a robust Focus-Net that performs finer-grained perception guided by top-down attention (i.e., look closely next). To fully unleash the power of top-down attention, we further propose a novel context-mixing dynamic convolution (ContMix) that effectively models long-range dependencies while preserving inherent local inductive biases even when the input resolution increases, addressing critical limitations in existing convolutions. Our OverLoCK exhibits a notable performance improvement over existing methods. For instance, OverLoCK-T achieves a Top-1 accuracy of 84.2%, significantly surpassing ConvNeXt-B while using only around one-third of the FLOPs/parameters. On object detection, our OverLoCK-S clearly surpasses MogaNet-B by 1% in AP^b. On semantic segmentation, our OverLoCK-T remarkably improves UniRepLKNet-T by 1.7% in mIoU. Code is publicly available at https://rb.gy/wit4jh.
comment: Accepted by CVPR 2025
♻ ☆ FUSE: Label-Free Image-Event Joint Monocular Depth Estimation via Frequency-Decoupled Alignment and Degradation-Robust Fusion
Image-event joint depth estimation methods leverage complementary modalities for robust perception, yet face challenges in generalizability stemming from two factors: 1) limited annotated image-event-depth datasets causing insufficient cross-modal supervision, and 2) inherent frequency mismatches between static images and dynamic event streams with distinct spatiotemporal patterns, leading to ineffective feature fusion. To address this dual challenge, we propose Frequency-decoupled Unified Self-supervised Encoder (FUSE) with two synergistic components: The Parameter-efficient Self-supervised Transfer (PST) establishes cross-modal knowledge transfer through latent space alignment with image foundation models, effectively mitigating data scarcity by enabling joint encoding without depth ground truth. Complementing this, we propose the Frequency-Decoupled Fusion module (FreDFuse) to explicitly decouple high-frequency edge features from low-frequency structural components, resolving modality-specific frequency mismatches through physics-aware fusion. This combined approach enables FUSE to construct a universal image-event encoder that only requires lightweight decoder adaptation for target datasets. Extensive experiments demonstrate state-of-the-art performance with 14% and 24.9% improvements in Abs.Rel on MVSEC and DENSE datasets. The framework exhibits remarkable zero-shot adaptability to challenging scenarios including extreme lighting and motion blur, significantly advancing real-world deployment capabilities. The source code for our method is publicly available at: https://github.com/sunpihai-up/FUSE
comment: 8 pages, 6 figures
♻ ☆ DAWN: Dynamic Frame Avatar with Non-autoregressive Diffusion Framework for Talking Head Video Generation
Talking head generation intends to produce vivid and realistic talking head videos from a single portrait and speech audio clip. Although significant progress has been made in diffusion-based talking head generation, almost all methods rely on autoregressive strategies, which suffer from limited context utilization beyond the current generation step, error accumulation, and slower generation speed. To address these challenges, we present DAWN (Dynamic frame Avatar With Non-autoregressive diffusion), a framework that enables all-at-once generation of dynamic-length video sequences. Specifically, it consists of two main components: (1) audio-driven holistic facial dynamics generation in the latent motion space, and (2) audio-driven head pose and blink generation. Extensive experiments demonstrate that our method generates authentic and vivid videos with precise lip motions, and natural pose/blink movements. Additionally, with a high generation speed, DAWN possesses strong extrapolation capabilities, ensuring the stable production of high-quality long videos. These results highlight the considerable promise and potential impact of DAWN in the field of talking head video generation. Furthermore, we hope that DAWN sparks further exploration of non-autoregressive approaches in diffusion models. Our code will be publicly available at https://github.com/Hanbo-Cheng/DAWN-pytorch.
♻ ☆ Stealthy Backdoor Attack in Self-Supervised Learning Vision Encoders for Large Vision Language Models
Self-supervised learning (SSL) vision encoders learn high-quality image representations and thus have become a vital part of developing vision modality of large vision language models (LVLMs). Due to the high cost of training such encoders, pre-trained encoders are widely shared and deployed into many LVLMs, which are security-critical or bear societal significance. Under this practical scenario, we reveal a new backdoor threat that significant visual hallucinations can be induced into these LVLMs by merely compromising vision encoders. Because of the sharing and reuse of these encoders, many downstream LVLMs may inherit backdoor behaviors from encoders, leading to widespread backdoors. In this work, we propose BadVision, the first method to exploit this vulnerability in SSL vision encoders for LVLMs with novel trigger optimization and backdoor learning techniques. We evaluate BadVision on two types of SSL encoders and LVLMs across eight benchmarks. We show that BadVision effectively drives the LVLMs to attacker-chosen hallucination with over 99% attack success rate, causing a 77.6% relative visual understanding error while maintaining the stealthiness. SoTA backdoor detection methods cannot detect our attack effectively.
♻ ☆ Hi-ALPS -- An Experimental Robustness Quantification of Six LiDAR-based Object Detection Systems for Autonomous Driving
Light Detection and Ranging (LiDAR) is an essential sensor technology for autonomous driving as it can capture high-resolution 3D data. As 3D object detection systems (OD) can interpret such point cloud data, they play a key role in the driving decisions of autonomous vehicles. Consequently, such 3D OD must be robust against all types of perturbations and must therefore be extensively tested. One approach is the use of adversarial examples, which are small, sometimes sophisticated perturbations in the input data that change, i.e., falsify, the prediction of the OD. These perturbations are carefully designed based on the weaknesses of the OD. The robustness of the OD cannot be quantified with adversarial examples in general, because if the OD is vulnerable to a given attack, it is unclear whether this is due to the robustness of the OD or whether the attack algorithm produces particularly strong adversarial examples. The contribution of this work is Hi-ALPS -- Hierarchical Adversarial-example-based LiDAR Perturbation Level System, where higher robustness of the OD is required to withstand the perturbations as the perturbation levels increase. In doing so, the Hi-ALPS levels successively implement a heuristic followed by established adversarial example approaches. In a series of comprehensive experiments using Hi-ALPS, we quantify the robustness of six state-of-the-art 3D OD under different types of perturbations. The results of the experiments show that none of the OD is robust against all Hi-ALPS levels; an important factor for the ranking is that human observers can still correctly recognize the perturbed objects, as the respective perturbations are small. To increase the robustness of the OD, we discuss the applicability of state-of-the-art countermeasures. In addition, we derive further suggestions for countermeasures based on our experimental results.
♻ ☆ COB-GS: Clear Object Boundaries in 3DGS Segmentation Based on Boundary-Adaptive Gaussian Splitting CVPR 2025
Accurate object segmentation is crucial for high-quality scene understanding in the 3D vision domain. However, 3D segmentation based on 3D Gaussian Splatting (3DGS) struggles with accurately delineating object boundaries, as Gaussian primitives often span across object edges due to their inherent volume and the lack of semantic guidance during training. In order to tackle these challenges, we introduce Clear Object Boundaries for 3DGS Segmentation (COB-GS), which aims to improve segmentation accuracy by clearly delineating blurry boundaries of interwoven Gaussian primitives within the scene. Unlike existing approaches that remove ambiguous Gaussians and sacrifice visual quality, COB-GS, as a 3DGS refinement method, jointly optimizes semantic and visual information, allowing the two different levels to cooperate with each other effectively. Specifically, for the semantic guidance, we introduce a boundary-adaptive Gaussian splitting technique that leverages semantic gradient statistics to identify and split ambiguous Gaussians, aligning them closely with object boundaries. For the visual optimization, we rectify the degraded suboptimal texture of the 3DGS scene, particularly along the refined boundary structures. Experimental results show that COB-GS substantially improves segmentation accuracy and robustness against inaccurate masks from pre-trained model, yielding clear boundaries while preserving high visual quality. Code is available at https://github.com/ZestfulJX/COB-GS.
comment: Accepted by CVPR 2025
♻ ☆ In the Blink of an Eye: Instant Game Map Editing using a Generative-AI Smart Brush
With video games steadily increasing in complexity, automated generation of game content has found widespread interest. However, the task of 3D gaming map art creation remains underexplored to date due to its unique complexity and domain-specific challenges. While recent works have addressed related topics such as retro-style level generation and procedural terrain creation, these works primarily focus on simpler data distributions. To the best of our knowledge, we are the first to demonstrate the application of modern AI techniques for high-resolution texture manipulation in complex, highly detailed AAA 3D game environments. We introduce a novel Smart Brush for map editing, designed to assist artists in seamlessly modifying selected areas of a game map with minimal effort. By leveraging generative adversarial networks and diffusion models we propose two variants of the brush that enable efficient and context-aware generation. Our hybrid workflow aims to enhance both artistic flexibility and production efficiency, enabling the refinement of environments without manually reworking every detail, thus helping to bridge the gap between automation and creative control in game development. A comparative evaluation of our two methods with adapted versions of several state-of-the art models shows that our GAN-based brush produces the sharpest and most detailed outputs while preserving image context while the evaluated state-of-the-art models tend towards blurrier results and exhibit difficulties in maintaining contextual consistency.
♻ ☆ VesselSAM: Leveraging SAM for Aortic Vessel Segmentation with LoRA and Atrous Attention
Medical image segmentation is crucial for clinical diagnosis and treatment planning, especially when dealing with complex anatomical structures such as vessels. However, accurately segmenting vessels remains challenging due to their small size, intricate edge structures, and susceptibility to artifacts and imaging noise. In this work, we propose VesselSAM, an enhanced version of the Segment Anything Model (SAM), specifically tailored for aortic vessel segmentation. VesselSAM incorporates AtrousLoRA, a novel module integrating Atrous Attention and Low-Rank Adaptation (LoRA), to enhance segmentation performance. Atrous Attention enables the model to capture multi-scale contextual information, preserving both fine-grained local details and broader global context. Additionally, LoRA facilitates efficient fine-tuning of the frozen SAM image encoder, reducing the number of trainable parameters and thereby enhancing computational efficiency. We evaluate VesselSAM using two challenging datasets: the Aortic Vessel Tree (AVT) dataset and the Type-B Aortic Dissection (TBAD) dataset. VesselSAM achieves state-of-the-art performance, attaining DSC scores of 93.50\%, 93.25\%, 93.02\%, and 93.26\% across multi-center datasets. Our results demonstrate that VesselSAM delivers high segmentation accuracy while significantly reducing computational overhead compared to existing large-scale models. This development paves the way for enhanced AI-based aortic vessel segmentation in clinical environments. The code and models will be released at https://github.com/Adnan-CAS/AtrousLora.
comment: Work in progress
♻ ☆ Benchmarking Large Vision-Language Models via Directed Scene Graph for Comprehensive Image Captioning CVPR2025
Generating detailed captions comprehending text-rich visual content in images has received growing attention for Large Vision-Language Models (LVLMs). However, few studies have developed benchmarks specifically tailored for detailed captions to measure their accuracy and comprehensiveness. In this paper, we introduce a detailed caption benchmark, termed as CompreCap, to evaluate the visual context from a directed scene graph view. Concretely, we first manually segment the image into semantically meaningful regions (i.e., semantic segmentation mask) according to common-object vocabulary, while also distinguishing attributes of objects within all those regions. Then directional relation labels of these objects are annotated to compose a directed scene graph that can well encode rich compositional information of the image. Based on our directed scene graph, we develop a pipeline to assess the generated detailed captions from LVLMs on multiple levels, including the object-level coverage, the accuracy of attribute descriptions, the score of key relationships, etc. Experimental results on the CompreCap dataset confirm that our evaluation method aligns closely with human evaluation scores across LVLMs.
comment: Accepted by CVPR2025. Code and Dataset: https://github.com/LuFan31/CompreCap
♻ ☆ Vision-based Multi-future Trajectory Prediction: A Survey
Vision-based trajectory prediction is an important task that supports safe and intelligent behaviours in autonomous systems. Many advanced approaches have been proposed over the years with improved spatial and temporal feature extraction. However, human behaviour is naturally diverse and uncertain. Given the past trajectory and surrounding environment information, an agent can have multiple plausible trajectories in the future. To tackle this problem, an essential task named multi-future trajectory prediction (MTP) has recently been studied. This task aims to generate a diverse, acceptable and explainable distribution of future predictions for each agent. In this paper, we present the first survey for MTP with our unique taxonomies and a comprehensive analysis of frameworks, datasets and evaluation metrics. We also compare models on existing MTP datasets and conduct experiments on the ForkingPath dataset. Finally, we discuss multiple future directions that can help researchers develop novel multi-future trajectory prediction systems and other diverse learning tasks similar to MTP.
comment: Accepted by TNNLS 2025
♻ ☆ Parameter-Efficient Fine-Tuning for Pre-Trained Vision Models: A Survey
Large-scale pre-trained vision models (PVMs) have shown great potential for adaptability across various downstream vision tasks. However, with state-of-the-art PVMs growing to billions or even trillions of parameters, the standard full fine-tuning paradigm is becoming unsustainable due to high computational and storage demands. In response, researchers are exploring parameter-efficient fine-tuning (PEFT), which seeks to exceed the performance of full fine-tuning with minimal parameter modifications. This survey provides a comprehensive overview and future directions for visual PEFT, offering a systematic review of the latest advancements. First, we provide a formal definition of PEFT and discuss model pre-training methods. We then categorize existing methods into three categories: addition-based, partial-based, and unified-based. Finally, we introduce the commonly used datasets and applications and suggest potential future research challenges. A comprehensive collection of resources is available at https://github.com/synbol/Awesome-Parameter-Efficient-Transfer-Learning.
comment: 9 pages, 3 figures, 2 tables
♻ ☆ GM-MoE: Low-Light Enhancement with Gated-Mechanism Mixture-of-Experts
Low-light enhancement has wide applications in autonomous driving, 3D reconstruction, remote sensing, surveillance, and so on, which can significantly improve information utilization. However, most existing methods lack generalization and are limited to specific tasks such as image recovery. To address these issues, we propose Gated-Mechanism Mixture-of-Experts (GM-MoE), the first framework to introduce a mixture-of-experts network for low-light image enhancement. GM-MoE comprises a dynamic gated weight conditioning network and three sub-expert networks, each specializing in a distinct enhancement task. Combining a self-designed gated mechanism that dynamically adjusts the weights of the sub-expert networks for different data domains. Additionally, we integrate local and global feature fusion within sub-expert networks to enhance image quality by capturing multi-scale features. Experimental results demonstrate that the GM-MoE achieves superior generalization with respect to 25 compared approaches, reaching state-of-the-art performance on PSNR on 5 benchmarks and SSIM on 4 benchmarks, respectively.
♻ ☆ EVT: Efficient View Transformation for Multi-Modal 3D Object Detection
Multi-modal sensor fusion in Bird's Eye View (BEV) representation has become the leading approach for 3D object detection. However, existing methods often rely on depth estimators or transformer encoders to transform image features into BEV space, which reduces robustness or introduces significant computational overhead. Moreover, the insufficient geometric guidance in view transformation results in ray-directional misalignments, limiting the effectiveness of BEV representations. To address these challenges, we propose Efficient View Transformation (EVT), a novel 3D object detection framework that constructs a well-structured BEV representation, improving both accuracy and efficiency. Our approach focuses on two key aspects. First, Adaptive Sampling and Adaptive Projection (ASAP), which utilizes LiDAR guidance to generate 3D sampling points and adaptive kernels, enables more effective transformation of image features into BEV space and a refined BEV representation. Second, an improved query-based detection framework, incorporating group-wise mixed query selection and geometry-aware cross-attention, effectively captures both the common properties and the geometric structure of objects in the transformer decoder. On the nuScenes test set, EVT achieves state-of-the-art performance of 75.3\% NDS with real-time inference speed.
♻ ☆ ST-VLM: Kinematic Instruction Tuning for Spatio-Temporal Reasoning in Vision-Language Models
Spatio-temporal reasoning is essential in understanding real-world environments in various fields, eg, autonomous driving and sports analytics. Recent advances have improved the spatial reasoning ability of Vision-Language Models (VLMs) by introducing large-scale data, but these models still struggle to analyze kinematic elements like traveled distance and speed of moving objects. To bridge this gap, we construct a spatio-temporal reasoning dataset and benchmark involving kinematic instruction tuning, referred to as STKit and STKit-Bench. They consist of real-world videos with 3D annotations, detailing object motion dynamics: traveled distance, speed, movement direction, inter-object distance comparisons, and relative movement direction. To further scale such data construction to videos without 3D labels, we propose an automatic pipeline to generate pseudo-labels using 4D reconstruction in real-world scale. With our kinematic instruction tuning data for spatio-temporal reasoning, we present ST-VLM, a VLM enhanced for spatio-temporal reasoning, which exhibits outstanding performance on STKit-Bench. Furthermore, we show that ST-VLM generalizes robustly across diverse domains and tasks, outperforming baselines on other spatio-temporal benchmarks (eg, ActivityNet, TVQA+). Finally, by integrating learned spatio-temporal reasoning with existing abilities, ST-VLM enables complex multi-step reasoning. Project page: https://ikodoh.github.io/ST-VLM.
♻ ☆ AvatarArtist: Open-Domain 4D Avatarization CVPR 2025
This work focuses on open-domain 4D avatarization, with the purpose of creating a 4D avatar from a portrait image in an arbitrary style. We select parametric triplanes as the intermediate 4D representation and propose a practical training paradigm that takes advantage of both generative adversarial networks (GANs) and diffusion models. Our design stems from the observation that 4D GANs excel at bridging images and triplanes without supervision yet usually face challenges in handling diverse data distributions. A robust 2D diffusion prior emerges as the solution, assisting the GAN in transferring its expertise across various domains. The synergy between these experts permits the construction of a multi-domain image-triplane dataset, which drives the development of a general 4D avatar creator. Extensive experiments suggest that our model, AvatarArtist, is capable of producing high-quality 4D avatars with strong robustness to various source image domains. The code, the data, and the models will be made publicly available to facilitate future studies.
comment: Accepted to CVPR 2025. Project page: https://kumapowerliu.github.io/AvatarArtist
♻ ☆ Human Motion Instruction Tuning CVPR 2025
This paper presents LLaMo (Large Language and Human Motion Assistant), a multimodal framework for human motion instruction tuning. In contrast to conventional instruction-tuning approaches that convert non-linguistic inputs, such as video or motion sequences, into language tokens, LLaMo retains motion in its native form for instruction tuning. This method preserves motion-specific details that are often diminished in tokenization, thereby improving the model's ability to interpret complex human behaviors. By processing both video and motion data alongside textual inputs, LLaMo enables a flexible, human-centric analysis. Experimental evaluations across high-complexity domains, including human behaviors and professional activities, indicate that LLaMo effectively captures domain-specific knowledge, enhancing comprehension and prediction in motion-intensive scenarios. We hope LLaMo offers a foundation for future multimodal AI systems with broad applications, from sports analytics to behavioral prediction. Our code and models are available on the project website: https://github.com/ILGLJ/LLaMo.
comment: Accepted by CVPR 2025
♻ ☆ MonoTAKD: Teaching Assistant Knowledge Distillation for Monocular 3D Object Detection CVPR 2025
Monocular 3D object detection (Mono3D) holds noteworthy promise for autonomous driving applications owing to the cost-effectiveness and rich visual context of monocular camera sensors. However, depth ambiguity poses a significant challenge, as it requires extracting precise 3D scene geometry from a single image, resulting in suboptimal performance when transferring knowledge from a LiDAR-based teacher model to a camera-based student model. To facilitate effective distillation, we introduce Monocular Teaching Assistant Knowledge Distillation (MonoTAKD), which proposes a camera-based teaching assistant (TA) model to transfer robust 3D visual knowledge to the student model, leveraging the smaller feature representation gap. Additionally, we define 3D spatial cues as residual features that capture the differences between the teacher and the TA models. We then leverage these cues to improve the student model's 3D perception capabilities. Experimental results show that our MonoTAKD achieves state-of-the-art performance on the KITTI3D dataset. Furthermore, we evaluate the performance on nuScenes and KITTI raw datasets to demonstrate the generalization of our model to multi-view 3D and unsupervised data settings. Our code is available at https://github.com/hoiliu-0801/MonoTAKD.
comment: Accepted by CVPR 2025. Our code is available at https://github.com/hoiliu-0801/MonoTAKD
♻ ☆ Towards Visual Discrimination and Reasoning of Real-World Physical Dynamics: Physics-Grounded Anomaly Detection CVPR25
Humans detect real-world object anomalies by perceiving, interacting, and reasoning based on object-conditioned physical knowledge. The long-term goal of Industrial Anomaly Detection (IAD) is to enable machines to autonomously replicate this skill. However, current IAD algorithms are largely developed and tested on static, semantically simple datasets, which diverge from real-world scenarios where physical understanding and reasoning are essential. To bridge this gap, we introduce the Physics Anomaly Detection (Phys-AD) dataset, the first large-scale, real-world, physics-grounded video dataset for industrial anomaly detection. Collected using a real robot arm and motor, Phys-AD provides a diverse set of dynamic, semantically rich scenarios. The dataset includes more than 6400 videos across 22 real-world object categories, interacting with robot arms and motors, and exhibits 47 types of anomalies. Anomaly detection in Phys-AD requires visual reasoning, combining both physical knowledge and video content to determine object abnormality. We benchmark state-of-the-art anomaly detection methods under three settings: unsupervised AD, weakly-supervised AD, and video-understanding AD, highlighting their limitations in handling physics-grounded anomalies. Additionally, we introduce the Physics Anomaly Explanation (PAEval) metric, designed to assess the ability of visual-language foundation models to not only detect anomalies but also provide accurate explanations for their underlying physical causes. Our project is available at https://guyao2023.github.io/Phys-AD/.
comment: Accepted by CVPR25
♻ ☆ LangBridge: Interpreting Image as a Combination of Language Embeddings
Recent years have witnessed remarkable advances in Large Vision-Language Models (LVLMs), which have achieved human-level performance across various complex vision-language tasks. Following LLaVA's paradigm, mainstream LVLMs typically employ a shallow MLP for visual-language alignment through a two-stage training process: pretraining for cross-modal alignment followed by instruction tuning. While this approach has proven effective, the underlying mechanisms of how MLPs bridge the modality gap remain poorly understood. Although some research has explored how LLMs process transformed visual tokens, few studies have investigated the fundamental alignment mechanism. Furthermore, the MLP adapter requires retraining whenever switching LLM backbones. To address these limitations, we first investigate the working principles of MLP adapters and discover that they learn to project visual embeddings into subspaces spanned by corresponding text embeddings progressively. Based on this insight, we propose LangBridge, a novel adapter that explicitly maps visual tokens to linear combinations of LLM vocabulary embeddings. This innovative design enables pretraining-free adapter transfer across different LLMs while maintaining performance. Our experimental results demonstrate that a LangBridge adapter pre-trained on Qwen2-0.5B can be directly applied to larger models such as LLaMA3-8B or Qwen2.5-14B while maintaining competitive performance. Overall, LangBridge enables interpretable vision-language alignment by grounding visual representations in LLM vocab embedding, while its plug-and-play design ensures efficient reuse across multiple LLMs with nearly no performance degradation. See our project page at https://jiaqiliao77.github.io/LangBridge.github.io/
comment: The code and weights will be open-sourced. Project page: https://jiaqiliao77.github.io/LangBridge.github.io/
♻ ☆ EfficientMT: Efficient Temporal Adaptation for Motion Transfer in Text-to-Video Diffusion Models
The progress on generative models has led to significant advances on text-to-video (T2V) generation, yet the motion controllability of generated videos remains limited. Existing motion transfer methods explored the motion representations of reference videos to guide generation. Nevertheless, these methods typically rely on sample-specific optimization strategy, resulting in high computational burdens. In this paper, we propose EfficientMT, a novel and efficient end-to-end framework for video motion transfer. By leveraging a small set of synthetic paired motion transfer samples, EfficientMT effectively adapts a pretrained T2V model into a general motion transfer framework that can accurately capture and reproduce diverse motion patterns. Specifically, we repurpose the backbone of the T2V model to extract temporal information from reference videos, and further propose a scaler module to distill motion-related information. Subsequently, we introduce a temporal integration mechanism that seamlessly incorporates reference motion features into the video generation process. After training on our self-collected synthetic paired samples, EfficientMT enables general video motion transfer without requiring test-time optimization. Extensive experiments demonstrate that our EfficientMT outperforms existing methods in efficiency while maintaining flexible motion controllability. Our code will be available https://github.com/PrototypeNx/EfficientMT.
♻ ☆ UniHDSA: A Unified Relation Prediction Approach for Hierarchical Document Structure Analysis
Document structure analysis, aka document layout analysis, is crucial for understanding both the physical layout and logical structure of documents, serving information retrieval, document summarization, knowledge extraction, etc. Hierarchical Document Structure Analysis (HDSA) specifically aims to restore the hierarchical structure of documents created using authoring software with hierarchical schemas. Previous research has primarily followed two approaches: one focuses on tackling specific subtasks of HDSA in isolation, such as table detection or reading order prediction, while the other adopts a unified framework that uses multiple branches or modules, each designed to address a distinct task. In this work, we propose a unified relation prediction approach for HDSA, called UniHDSA, which treats various HDSA sub-tasks as relation prediction problems and consolidates relation prediction labels into a unified label space. This allows a single relation prediction module to handle multiple tasks simultaneously, whether at a page-level or document-level structure analysis. To validate the effectiveness of UniHDSA, we develop a multimodal end-to-end system based on Transformer architectures. Extensive experimental results demonstrate that our approach achieves state-of-the-art performance on a hierarchical document structure analysis benchmark, Comp-HRDoc, and competitive results on a large-scale document layout analysis dataset, DocLayNet, effectively illustrating the superiority of our method across all sub-tasks. The Comp-HRDoc benchmark and UniHDSA's configurations are publicly available at https://github.com/microsoft/CompHRDoc.
comment: Accepted by Pattern Recognition. arXiv admin note: text overlap with arXiv:2405.11757
♻ ☆ Attention IoU: Examining Biases in CelebA using Attention Maps CVPR 2025
Computer vision models have been shown to exhibit and amplify biases across a wide array of datasets and tasks. Existing methods for quantifying bias in classification models primarily focus on dataset distribution and model performance on subgroups, overlooking the internal workings of a model. We introduce the Attention-IoU (Attention Intersection over Union) metric and related scores, which use attention maps to reveal biases within a model's internal representations and identify image features potentially causing the biases. First, we validate Attention-IoU on the synthetic Waterbirds dataset, showing that the metric accurately measures model bias. We then analyze the CelebA dataset, finding that Attention-IoU uncovers correlations beyond accuracy disparities. Through an investigation of individual attributes through the protected attribute of Male, we examine the distinct ways biases are represented in CelebA. Lastly, by subsampling the training set to change attribute correlations, we demonstrate that Attention-IoU reveals potential confounding variables not present in dataset labels.
comment: To appear in CVPR 2025. Code and data is available at https://github.com/aaronserianni/attention-iou . 15 pages, 14 figures, including appendix
♻ ☆ Three Cars Approaching within 100m! Enhancing Distant Geometry by Tri-Axis Voxel Scanning for Camera-based Semantic Scene Completion CVPR 2025
Camera-based Semantic Scene Completion (SSC) is gaining attentions in the 3D perception field. However, properties such as perspective and occlusion lead to the underestimation of the geometry in distant regions, posing a critical issue for safety-focused autonomous driving systems. To tackle this, we propose ScanSSC, a novel camera-based SSC model composed of a Scan Module and Scan Loss, both designed to enhance distant scenes by leveraging context from near-viewpoint scenes. The Scan Module uses axis-wise masked attention, where each axis employing a near-to-far cascade masking that enables distant voxels to capture relationships with preceding voxels. In addition, the Scan Loss computes the cross-entropy along each axis between cumulative logits and corresponding class distributions in a near-to-far direction, thereby propagating rich context-aware signals to distant voxels. Leveraging the synergy between these components, ScanSSC achieves state-of-the-art performance, with IoUs of 44.54 and 48.29, and mIoUs of 17.40 and 20.14 on the SemanticKITTI and SSCBench-KITTI-360 benchmarks.
comment: Accepted to CVPR 2025
♻ ☆ Uni$\textbf{F}^2$ace: Fine-grained Face Understanding and Generation with Unified Multimodal Models
Unified multimodal models (UMMs) have emerged as a powerful paradigm in foundational computer vision research, demonstrating significant potential in both image understanding and generation. However, existing research in the face domain primarily focuses on $\textbf{coarse}$ facial attribute understanding, with limited capacity to handle $\textbf{fine-grained}$ facial attributes and without addressing generation capabilities. To overcome these limitations, we propose Uni$\textbf{F}^2$ace, the first UMM tailored specifically for fine-grained face understanding and generation. In general, we train Uni$\textbf{F}^2$ace on a self-constructed, specialized dataset utilizing two mutually beneficial diffusion techniques and a two-level mixture-of-experts architecture. Specifically, we first build a large-scale facial dataset, Uni$\textbf{F}^2$ace-130K, which contains 130K image-text pairs with one million question-answering pairs that span a wide range of facial attributes. Second, we establish a theoretical connection between discrete diffusion score matching and masked generative models, optimizing both evidence lower bounds simultaneously, which significantly improves the model's ability to synthesize facial details. Finally, we introduce both token-level and sequence-level mixture-of-experts, enabling efficient fine-grained representation learning for both understanding and generation tasks. Extensive experiments on Uni$\textbf{F}^2$ace-130K demonstrate that Uni$\textbf{F}^2$ace outperforms existing UMMs and generative models, achieving superior performance across both understanding and generation tasks.
♻ ☆ Progress-Aware Video Frame Captioning CVPR 2025
While image captioning provides isolated descriptions for individual images, and video captioning offers one single narrative for an entire video clip, our work explores an important middle ground: progress-aware video captioning at the frame level. This novel task aims to generate temporally fine-grained captions that not only accurately describe each frame but also capture the subtle progression of actions throughout a video sequence. Despite the strong capabilities of existing leading vision language models, they often struggle to discern the nuances of frame-wise differences. To address this, we propose ProgressCaptioner, a captioning model designed to capture the fine-grained temporal dynamics within an action sequence. Alongside, we develop the FrameCap dataset to support training and the FrameCapEval benchmark to assess caption quality. The results demonstrate that ProgressCaptioner significantly surpasses leading captioning models, producing precise captions that accurately capture action progression and set a new standard for temporal precision in video captioning. Finally, we showcase practical applications of our approach, specifically in aiding keyframe selection and advancing video understanding, highlighting its broad utility.
comment: Accepted by CVPR 2025, Project website: https://vision.cs.utexas.edu/projects/ProgressCaptioner/
♻ ☆ CamSAM2: Segment Anything Accurately in Camouflaged Videos
Video camouflaged object segmentation (VCOS), aiming at segmenting camouflaged objects that seamlessly blend into their environment, is a fundamental vision task with various real-world applications. With the release of SAM2, video segmentation has witnessed significant progress. However, SAM2's capability of segmenting camouflaged videos is suboptimal, especially when given simple prompts such as point and box. To address the problem, we propose Camouflaged SAM2 (CamSAM2), which enhances SAM2's ability to handle camouflaged scenes without modifying SAM2's parameters. Specifically, we introduce a decamouflaged token to provide the flexibility of feature adjustment for VCOS. To make full use of fine-grained and high-resolution features from the current frame and previous frames, we propose implicit object-aware fusion (IOF) and explicit object-aware fusion (EOF) modules, respectively. Object prototype generation (OPG) is introduced to abstract and memorize object prototypes with informative details using high-quality features from previous frames. Extensive experiments are conducted to validate the effectiveness of our approach. While CamSAM2 only adds negligible learnable parameters to SAM2, it substantially outperforms SAM2 on three VCOS datasets, especially achieving 12.2 mDice gains with click prompt on MoCA-Mask and 19.6 mDice gains with mask prompt on SUN-SEG-Hard, with Hiera-T as the backbone. The code will be available at https://github.com/zhoustan/CamSAM2.
♻ ☆ HLV-1K: A Large-scale Hour-Long Video Benchmark for Time-Specific Long Video Understanding ICME 2025
Multimodal large language models have become a popular topic in deep visual understanding due to many promising real-world applications. However, hour-long video understanding, spanning over one hour and containing tens of thousands of visual frames, remains under-explored because of 1) challenging long-term video analyses, 2) inefficient large-model approaches, and 3) lack of large-scale benchmark datasets. Among them, in this paper, we focus on building a large-scale hour-long long video benchmark, HLV-1K, designed to evaluate long video understanding models. HLV-1K comprises 1009 hour-long videos with 14,847 high-quality question answering (QA) and multi-choice question asnwering (MCQA) pairs with time-aware query and diverse annotations, covering frame-level, within-event-level, cross-event-level, and long-term reasoning tasks. We evaluate our benchmark using existing state-of-the-art methods and demonstrate its value for testing deep long video understanding capabilities at different levels and for various tasks. This includes promoting future long video understanding tasks at a granular level, such as deep understanding of long live videos, meeting recordings, and movies.
comment: Accepted to ICME 2025
♻ ☆ Narrating the Video: Boosting Text-Video Retrieval via Comprehensive Utilization of Frame-Level Captions CVPR 2025
In recent text-video retrieval, the use of additional captions from vision-language models has shown promising effects on the performance. However, existing models using additional captions often have struggled to capture the rich semantics, including temporal changes, inherent in the video. In addition, incorrect information caused by generative models can lead to inaccurate retrieval. To address these issues, we propose a new framework, Narrating the Video (NarVid), which strategically leverages the comprehensive information available from frame-level captions, the narration. The proposed NarVid exploits narration in multiple ways: 1) feature enhancement through cross-modal interactions between narration and video, 2) query-aware adaptive filtering to suppress irrelevant or incorrect information, 3) dual-modal matching score by adding query-video similarity and query-narration similarity, and 4) hard-negative loss to learn discriminative features from multiple perspectives using the two similarities from different views. Experimental results demonstrate that NarVid achieves state-of-the-art performance on various benchmark datasets.
comment: Accepted at CVPR 2025
♻ ☆ Any2AnyTryon: Leveraging Adaptive Position Embeddings for Versatile Virtual Clothing Tasks
Image-based virtual try-on (VTON) aims to generate a virtual try-on result by transferring an input garment onto a target person's image. However, the scarcity of paired garment-model data makes it challenging for existing methods to achieve high generalization and quality in VTON. Also, it limits the ability to generate mask-free try-ons. To tackle the data scarcity problem, approaches such as Stable Garment and MMTryon use a synthetic data strategy, effectively increasing the amount of paired data on the model side. However, existing methods are typically limited to performing specific try-on tasks and lack user-friendliness. To enhance the generalization and controllability of VTON generation, we propose Any2AnyTryon, which can generate try-on results based on different textual instructions and model garment images to meet various needs, eliminating the reliance on masks, poses, or other conditions. Specifically, we first construct the virtual try-on dataset LAION-Garment, the largest known open-source garment try-on dataset. Then, we introduce adaptive position embedding, which enables the model to generate satisfactory outfitted model images or garment images based on input images of different sizes and categories, significantly enhancing the generalization and controllability of VTON generation. In our experiments, we demonstrate the effectiveness of our Any2AnyTryon and compare it with existing methods. The results show that Any2AnyTryon enables flexible, controllable, and high-quality image-based virtual try-on generation. https://logn-2024.github.io/Any2anyTryonProjectPage
comment: 13 pages,13 figures
♻ ☆ PAINT: Paying Attention to INformed Tokens to Mitigate Hallucination in Large Vision-Language Model
Large Vision Language Models (LVLMs) have demonstrated remarkable capabilities in understanding and describing visual content, achieving state-of-the-art performance across various vision-language tasks. However, these models often generate descriptions containing objects or details that are absent in the input image, a phenomenon commonly known as hallucination. Our work investigates the key reasons behind this issue by analyzing the pattern of self-attention in transformer layers. We find that hallucinations often arise from the progressive weakening of attention weight to visual tokens in the deeper layers of the LLM. Some previous works naively boost the attention of all visual tokens to mitigate this issue, resulting in suboptimal hallucination reduction. To address this, we identify two critical sets of visual tokens that facilitate the transfer of visual information from the vision encoder to the LLM. Local tokens encode grounded information about objects present in an image, while summary tokens capture the overall aggregated representation of the image. Importantly, these two sets of tokens require different levels of weight enhancement. To this end, we propose \textbf{PAINT} (\textbf{P}aying \textbf{A}ttention to \textbf{IN}formed \textbf{T}okens), a plug-and-play framework that intervenes in the self-attention mechanism of the LLM, selectively boosting the attention weights of local and summary tokens with experimentally learned margins. Evaluation on the MSCOCO image captioning dataset demonstrate that our approach reduces hallucination rates by up to 62.3\% compared to baseline models while maintaining accuracy. Code is available at \href{https://github.com/hasanar1f/PAINT}{https://github.com/hasanar1f/PAINT}
comment: 6 pages, 4 tables, 3 figures
♻ ☆ Explaining Deep Convolutional Neural Networks for Image Classification by Evolving Local Interpretable Model-agnostic Explanations
Deep convolutional neural networks have proven their effectiveness, and have been acknowledged as the most dominant method for image classification. However, a severe drawback of deep convolutional neural networks is poor explainability. Unfortunately, in many real-world applications, users need to understand the rationale behind the predictions of deep convolutional neural networks when determining whether they should trust the predictions or not. To resolve this issue, a novel genetic algorithm-based method is proposed for the first time to automatically evolve local explanations that can assist users to assess the rationality of the predictions. Furthermore, the proposed method is model-agnostic, i.e., it can be utilised to explain any deep convolutional neural network models. In the experiments, ResNet is used as an example model to be explained, and the ImageNet dataset is selected as the benchmark dataset. DenseNet and MobileNet are further explained to demonstrate the model-agnostic characteristic of the proposed method. The evolved local explanations on four images, randomly selected from ImageNet, are presented, which show that the evolved local explanations are straightforward to be recognised by humans. Moreover, the evolved explanations can explain the predictions of deep convolutional neural networks on all four images very well by successfully capturing meaningful interpretable features of the sample images. Further analysis based on the 30 runs of the experiments exhibits that the evolved local explanations can also improve the probabilities/confidences of the deep convolutional neural network models in making the predictions. The proposed method can obtain local explanations within one minute, which is more than ten times faster than LIME (the state-of-the-art method).
♻ ☆ CATD: Unified Representation Learning for EEG-to-fMRI Cross-Modal Generation
Multi-modal neuroimaging analysis is crucial for a comprehensive understanding of brain function and pathology, as it allows for the integration of different imaging techniques, thus overcoming the limitations of individual modalities. However, the high costs and limited availability of certain modalities pose significant challenges. To address these issues, this paper proposes the Condition-Aligned Temporal Diffusion (CATD) framework for end-to-end cross-modal synthesis of neuroimaging, enabling the generation of functional magnetic resonance imaging (fMRI)-detected Blood Oxygen Level Dependent (BOLD) signals from more accessible Electroencephalography (EEG) signals. By constructing Conditionally Aligned Block (CAB), heterogeneous neuroimages are aligned into a latent space, achieving a unified representation that provides the foundation for cross-modal transformation in neuroimaging. The combination with the constructed Dynamic Time-Frequency Segmentation (DTFS) module also enables the use of EEG signals to improve the temporal resolution of BOLD signals, thus augmenting the capture of the dynamic details of the brain. Experimental validation demonstrates that the framework improves the accuracy of brain activity state prediction by 9.13% (reaching 69.8%), enhances the diagnostic accuracy of brain disorders by 4.10% (reaching 99.55%), effectively identifies abnormal brain regions, enhancing the temporal resolution of BOLD signals. The proposed framework establishes a new paradigm for cross-modal synthesis of neuroimaging by unifying heterogeneous neuroimaging data into a latent representation space, showing promise in medical applications such as improving Parkinson's disease prediction and identifying abnormal brain regions.
comment: 11 pages, 9 figures, Accepted by IEEE Transactions on Medical Imaging
♻ ☆ Joint Learning for Scattered Point Cloud Understanding with Hierarchical Self-Distillation
Numerous point-cloud understanding techniques focus on whole entities and have succeeded in obtaining satisfactory results and limited sparsity tolerance. However, these methods are generally sensitive to incomplete point clouds that are scanned with flaws or large gaps. In this paper, we propose an end-to-end architecture that compensates for and identifies partial point clouds on the fly. First, we propose a cascaded solution that integrates both the upstream masked autoencoder (MAE) and downstream understanding networks simultaneously, allowing the task-oriented downstream to identify the points generated by the completion-oriented upstream. These two streams complement each other, resulting in improved performance for both completion and downstream-dependent tasks. Second, to explicitly understand the predicted points' pattern, we introduce hierarchical self-distillation (HSD), which can be applied to any hierarchy-based point cloud methods. HSD ensures that the deepest classifier with a larger perceptual field of local kernels and longer code length provides additional regularization to intermediate ones rather than simply aggregating the multi-scale features, and therefore maximizing the mutual information (MI) between a teacher and students. The proposed HSD strategy is particularly well-suited for tasks involving scattered point clouds, wherein a singular prediction may yield imprecise outcomes due to the inherently irregular and sparse nature of the geometric shape being reconstructed. We show the advantage of the self-distillation process in the hyperspaces based on the information bottleneck principle. Our method achieves state-of-the-art on both classification and part segmentation tasks.
comment: Published version (early-view) without bios
♻ ☆ MamBEV: Enabling State Space Models to Learn Birds-Eye-View Representations
3D visual perception tasks, such as 3D detection from multi-camera images, are essential components of autonomous driving and assistance systems. However, designing computationally efficient methods remains a significant challenge. In this paper, we propose a Mamba-based framework called MamBEV, which learns unified Bird's Eye View (BEV) representations using linear spatio-temporal SSM-based attention. This approach supports multiple 3D perception tasks with significantly improved computational and memory efficiency. Furthermore, we introduce SSM based cross-attention, analogous to standard cross attention, where BEV query representations can interact with relevant image features. Extensive experiments demonstrate MamBEV's promising performance across diverse visual perception metrics, highlighting its advantages in input scaling efficiency compared to existing benchmark models.
♻ ☆ A Multimodal Vision Foundation Model for Clinical Dermatology
Diagnosing and treating skin diseases require advanced visual skills across domains and the ability to synthesize information from multiple imaging modalities. While current deep learning models excel at specific tasks like skin cancer diagnosis from dermoscopic images, they struggle to meet the complex, multimodal requirements of clinical practice. Here, we introduce PanDerm, a multimodal dermatology foundation model pretrained through self-supervised learning on over 2 million real-world skin disease images from 11 clinical institutions across 4 imaging modalities. We evaluated PanDerm on 28 diverse benchmarks, including skin cancer screening, risk stratification, differential diagnosis of common and rare skin conditions, lesion segmentation, longitudinal monitoring, and metastasis prediction and prognosis. PanDerm achieved state-of-the-art performance across all evaluated tasks, often outperforming existing models when using only 10% of labeled data. We conducted three reader studies to assess PanDerm's potential clinical utility. PanDerm outperformed clinicians by 10.2% in early-stage melanoma detection through longitudinal analysis, improved clinicians' skin cancer diagnostic accuracy by 11% on dermoscopy images, and enhanced non-dermatologist healthcare providers' differential diagnosis by 16.5% across 128 skin conditions on clinical photographs. These results demonstrate PanDerm's potential to improve patient care across diverse clinical scenarios and serve as a model for developing multimodal foundation models in other medical specialties, potentially accelerating the integration of AI support in healthcare.
comment: 74 pages; Preprint
♻ ☆ Segmenting Bi-Atrial Structures Using ResNext Based Framework
Atrial fibrillation (AF) is the most common cardiac arrhythmia, significantly contributing to mortality, particularly in older populations. While pulmonary vein isolation is a standard treatment, its effectiveness is limited in patients with persistent AF. Recent research highlights the importance of targeting additional atrial regions, particularly fibrotic areas identified via late gadolinium-enhanced MRI (LGE-MRI). However, existing manual segmentation methods are time-consuming and prone to variability. Deep learning techniques, particularly convolutional neural networks (CNNs), have shown promise in automating segmentation. However, most studies focus solely on the left atrium (LA) and rely on small datasets, limiting generalizability. In this paper, we propose a novel two-stage framework incorporating ResNeXt encoders and a cyclic learning rate to segment both the right atrium (RA) and LA walls and cavities in LGE-MRIs. Our method aims to improve the segmentation of challenging small structures, such as atrial walls while maintaining high performance in larger regions like the atrial cavities. The results demonstrate that our approach offers superior segmentation accuracy and robustness compared to traditional architectures, particularly for imbalanced class structures.
♻ ☆ Grayscale to Hyperspectral at Any Resolution Using a Phase-Only Lens
We consider the problem of reconstructing a HxWx31 hyperspectral image from a HxW grayscale snapshot measurement that is captured using only a single diffractive optic and a filterless panchromatic photosensor. This problem is severely ill-posed, but we present the first model that produces high-quality results. We make efficient use of limited data by training a conditional denoising diffusion model that operates on small patches in a shift-invariant manner. During inference, we synchronize per-patch hyperspectral predictions using guidance derived from the optical point spread function. Surprisingly, our experiments reveal that patch sizes as small as the PSFs support achieve excellent results, and they show that local optical cues are sufficient to capture full spectral information. Moreover, by drawing multiple samples, our model provides per-pixel uncertainty estimates that strongly correlate with reconstruction error. Our work lays the foundation for a new class of high-resolution snapshot hyperspectral imagers that are compact and light-efficient.
♻ ☆ CLEAR: Cue Learning using Evolution for Accurate Recognition Applied to Sustainability Data Extraction
Large Language Model (LLM) image recognition is a powerful tool for extracting data from images, but accuracy depends on providing sufficient cues in the prompt - requiring a domain expert for specialized tasks. We introduce Cue Learning using Evolution for Accurate Recognition (CLEAR), which uses a combination of LLMs and evolutionary computation to generate and optimize cues such that recognition of specialized features in images is improved. It achieves this by auto-generating a novel domain-specific representation and then using it to optimize suitable textual cues with a genetic algorithm. We apply CLEAR to the real-world task of identifying sustainability data from interior and exterior images of buildings. We investigate the effects of using a variable-length representation compared to fixed-length and show how LLM consistency can be improved by refactoring from categorical to real-valued estimates. We show that CLEAR enables higher accuracy compared to expert human recognition and human-authored prompts in every task with error rates improved by up to two orders of magnitude and an ablation study evincing solution concision.
comment: 9 pages plus 2 pages of supplemental material
♻ ☆ Neural Light Spheres for Implicit Image Stitching and View Synthesis
Challenging to capture, and challenging to display on a cellphone screen, the panorama paradoxically remains both a staple and underused feature of modern mobile camera applications. In this work we address both of these challenges with a spherical neural light field model for implicit panoramic image stitching and re-rendering; able to accommodate for depth parallax, view-dependent lighting, and local scene motion and color changes during capture. Fit during test-time to an arbitrary path panoramic video capture -- vertical, horizontal, random-walk -- these neural light spheres jointly estimate the camera path and a high-resolution scene reconstruction to produce novel wide field-of-view projections of the environment. Our single-layer model avoids expensive volumetric sampling, and decomposes the scene into compact view-dependent ray offset and color components, with a total model size of 80 MB per scene, and real-time (50 FPS) rendering at 1080p resolution. We demonstrate improved reconstruction quality over traditional image stitching and radiance field methods, with significantly higher tolerance to scene motion and non-ideal capture settings.
comment: Project site: https://light.princeton.edu/publication/neuls/
♻ ☆ M-LLM Based Video Frame Selection for Efficient Video Understanding
Recent advances in Multi-Modal Large Language Models (M-LLMs) show promising results in video reasoning. Popular Multi-Modal Large Language Model (M-LLM) frameworks usually apply naive uniform sampling to reduce the number of video frames that are fed into an M-LLM, particularly for long context videos. However, it could lose crucial context in certain periods of a video, so that the downstream M-LLM may not have sufficient visual information to answer a question. To attack this pain point, we propose a light-weight M-LLM -based frame selection method that adaptively select frames that are more relevant to users' queries. In order to train the proposed frame selector, we introduce two supervision signals (i) Spatial signal, where single frame importance score by prompting a M-LLM; (ii) Temporal signal, in which multiple frames selection by prompting Large Language Model (LLM) using the captions of all frame candidates. The selected frames are then digested by a frozen downstream video M-LLM for visual reasoning and question answering. Empirical results show that the proposed M-LLM video frame selector improves the performances various downstream video Large Language Model (video-LLM) across medium (ActivityNet, NExT-QA) and long (EgoSchema, LongVideoBench) context video question answering benchmarks.
♻ ☆ JOG3R: Towards 3D-Consistent Video Generators
Emergent capabilities of image generators have led to many impactful zero- or few-shot applications. Inspired by this success, we investigate whether video generators similarly exhibit 3D-awareness. Using structure-from-motion as a 3D-aware task, we test if intermediate features of a video generator - OpenSora in our case - can support camera pose estimation. Surprisingly, at first, we only find a weak correlation between the two tasks. Deeper investigation reveals that although the video generator produces plausible video frames, the frames themselves are not truly 3D-consistent. Instead, we propose to jointly train for the two tasks, using photometric generation and 3D aware errors. Specifically, we find that SoTA video generation and camera pose estimation (i.e.,DUSt3R [79]) networks share common structures, and propose an architecture that unifies the two. The proposed unified model, named \nameMethod, produces camera pose estimates with competitive quality while producing 3D-consistent videos. In summary, we propose the first unified video generator that is 3D-consistent, generates realistic video frames, and can potentially be repurposed for other 3D-aware tasks.
♻ ☆ SimBEV: A Synthetic Multi-Task Multi-Sensor Driving Data Generation Tool and Dataset
Bird's-eye view (BEV) perception has garnered significant attention in autonomous driving in recent years, in part because BEV representation facilitates multi-modal sensor fusion. BEV representation enables a variety of perception tasks including BEV segmentation, a concise view of the environment useful for planning a vehicle's trajectory. However, this representation is not fully supported by existing datasets, and creation of new datasets for this purpose can be a time-consuming endeavor. To address this challenge, we introduce SimBEV. SimBEV is a randomized synthetic data generation tool that is extensively configurable and scalable, supports a wide array of sensors, incorporates information from multiple sources to capture accurate BEV ground truth, and enables a variety of perception tasks including BEV segmentation and 3D object detection. SimBEV is used to create the SimBEV dataset, a large collection of annotated perception data from diverse driving scenarios. SimBEV and the SimBEV dataset are open and available to the public.
♻ ☆ Evaluating Vision-Language Models as Evaluators in Path Planning
Despite their promise to perform complex reasoning, large language models (LLMs) have been shown to have limited effectiveness in end-to-end planning. This has inspired an intriguing question: if these models cannot plan well, can they still contribute to the planning framework as a helpful plan evaluator? In this work, we generalize this question to consider LLMs augmented with visual understanding, i.e., Vision-Language Models (VLMs). We introduce PathEval, a novel benchmark evaluating VLMs as plan evaluators in complex path-planning scenarios. Succeeding in the benchmark requires a VLM to be able to abstract traits of optimal paths from the scenario description, demonstrate precise low-level perception on each path, and integrate this information to decide the better path. Our analysis of state-of-the-art VLMs reveals that these models face significant challenges on the benchmark. We observe that the VLMs can precisely abstract given scenarios to identify the desired traits and exhibit mixed performance in integrating the provided information. Yet, their vision component presents a critical bottleneck, with models struggling to perceive low-level details about a path. Our experimental results show that this issue cannot be trivially addressed via end-to-end fine-tuning; rather, task-specific discriminative adaptation of these vision encoders is needed for these VLMs to become effective path evaluators.
♻ ☆ SparseGS: Real-Time 360° Sparse View Synthesis using Gaussian Splatting 3DV 2025
3D Gaussian Splatting (3DGS) has recently enabled real-time rendering of unbounded 3D scenes for novel view synthesis. However, this technique requires dense training views to accurately reconstruct 3D geometry. A limited number of input views will significantly degrade reconstruction quality, resulting in artifacts such as "floaters" and "background collapse" at unseen viewpoints. In this work, we introduce SparseGS, an efficient training pipeline designed to address the limitations of 3DGS in scenarios with sparse training views. SparseGS incorporates depth priors, novel depth rendering techniques, and a pruning heuristic to mitigate floater artifacts, alongside an Unseen Viewpoint Regularization module to alleviate background collapses. Our extensive evaluations on the Mip-NeRF360, LLFF, and DTU datasets demonstrate that SparseGS achieves high-quality reconstruction in both unbounded and forward-facing scenarios, with as few as 12 and 3 input images, respectively, while maintaining fast training and real-time rendering capabilities.
comment: Version accepted to 3DV 2025. Project page: https://github.com/ForMyCat/SparseGS
♻ ☆ MCBLT: Multi-Camera Multi-Object 3D Tracking in Long Videos
Object perception from multi-view cameras is crucial for intelligent systems, particularly in indoor environments, e.g., warehouses, retail stores, and hospitals. Most traditional multi-target multi-camera (MTMC) detection and tracking methods rely on 2D object detection, single-view multi-object tracking (MOT), and cross-view re-identification (ReID) techniques, without properly handling important 3D information by multi-view image aggregation. In this paper, we propose a 3D object detection and tracking framework, named MCBLT, which first aggregates multi-view images with necessary camera calibration parameters to obtain 3D object detections in bird's-eye view (BEV). Then, we introduce hierarchical graph neural networks (GNNs) to track these 3D detections in BEV for MTMC tracking results. Unlike existing methods, MCBLT has impressive generalizability across different scenes and diverse camera settings, with exceptional capability for long-term association handling. As a result, our proposed MCBLT establishes a new state-of-the-art on the AICity'24 dataset with $81.22$ HOTA, and on the WildTrack dataset with $95.6$ IDF1.
♻ ☆ Evaluating Pre-trained Convolutional Neural Networks and Foundation Models as Feature Extractors for Content-based Medical Image Retrieval
Medical image retrieval refers to the task of finding similar images for given query images in a database, with applications such as diagnosis support. While traditional medical image retrieval relied on clinical metadata, content-based medical image retrieval (CBMIR) depends on image features, which can be extracted automatically or semi-automatically. Many approaches have been proposed for CBMIR, and among them, using pre-trained convolutional neural networks (CNNs) is a widely utilized approach. However, considering the recent advances in the development of foundation models for various computer vision tasks, their application for CBMIR can also be investigated. In this study, we used several pre-trained feature extractors from well-known pre-trained CNNs and pre-trained foundation models and investigated the CBMIR performance on eight types of two-dimensional (2D) and three-dimensional (3D) medical images. Furthermore, we investigated the effect of image size on the CBMIR performance. Our results show that, overall, for the 2D datasets, foundation models deliver superior performance by a large margin compared to CNNs, with the general-purpose self-supervised model for computational pathology (UNI) providing the best overall performance across all datasets and image sizes. For 3D datasets, CNNs and foundation models deliver more competitive performance, with contrastive learning from captions for histopathology model (CONCH) achieving the best overall performance. Moreover, our findings confirm that while using larger image sizes (especially for 2D datasets) yields slightly better performance, competitive CBMIR performance can still be achieved even with smaller image sizes. Our codes to reproduce the results are available at: https://github.com/masih4/MedImageRetrieval.
comment: 37 pages
♻ ☆ Boltzmann Attention Sampling for Image Analysis with Small Objects
Detecting and segmenting small objects, such as lung nodules and tumor lesions, remains a critical challenge in image analysis. These objects often occupy less than 0.1% of an image, making traditional transformer architectures inefficient and prone to performance degradation due to redundant attention computations on irrelevant regions. Existing sparse attention mechanisms rely on rigid hierarchical structures, which are poorly suited for detecting small, variable, and uncertain object locations. In this paper, we propose BoltzFormer, a novel transformer-based architecture designed to address these challenges through dynamic sparse attention. BoltzFormer identifies and focuses attention on relevant areas by modeling uncertainty using a Boltzmann distribution with an annealing schedule. Initially, a higher temperature allows broader area sampling in early layers, when object location uncertainty is greatest. As the temperature decreases in later layers, attention becomes more focused, enhancing efficiency and accuracy. BoltzFormer seamlessly integrates into existing transformer architectures via a modular Boltzmann attention sampling mechanism. Comprehensive evaluations on benchmark datasets demonstrate that BoltzFormer significantly improves segmentation performance for small objects while reducing attention computation by an order of magnitude compared to previous state-of-the-art methods.
♻ ☆ WiLoR: End-to-end 3D Hand Localization and Reconstruction in-the-wild CVPR 2025
In recent years, 3D hand pose estimation methods have garnered significant attention due to their extensive applications in human-computer interaction, virtual reality, and robotics. In contrast, there has been a notable gap in hand detection pipelines, posing significant challenges in constructing effective real-world multi-hand reconstruction systems. In this work, we present a data-driven pipeline for efficient multi-hand reconstruction in the wild. The proposed pipeline is composed of two components: a real-time fully convolutional hand localization and a high-fidelity transformer-based 3D hand reconstruction model. To tackle the limitations of previous methods and build a robust and stable detection network, we introduce a large-scale dataset with over than 2M in-the-wild hand images with diverse lighting, illumination, and occlusion conditions. Our approach outperforms previous methods in both efficiency and accuracy on popular 2D and 3D benchmarks. Finally, we showcase the effectiveness of our pipeline to achieve smooth 3D hand tracking from monocular videos, without utilizing any temporal components. Code, models, and dataset are available https://rolpotamias.github.io/WiLoR.
comment: CVPR 2025, Project Page https://rolpotamias.github.io/WiLoR
♻ ☆ ILIAS: Instance-Level Image retrieval At Scale CVPR 2025
This work introduces ILIAS, a new test dataset for Instance-Level Image retrieval At Scale. It is designed to evaluate the ability of current and future foundation models and retrieval techniques to recognize particular objects. The key benefits over existing datasets include large scale, domain diversity, accurate ground truth, and a performance that is far from saturated. ILIAS includes query and positive images for 1,000 object instances, manually collected to capture challenging conditions and diverse domains. Large-scale retrieval is conducted against 100 million distractor images from YFCC100M. To avoid false negatives without extra annotation effort, we include only query objects confirmed to have emerged after 2014, i.e. the compilation date of YFCC100M. An extensive benchmarking is performed with the following observations: i) models fine-tuned on specific domains, such as landmarks or products, excel in that domain but fail on ILIAS ii) learning a linear adaptation layer using multi-domain class supervision results in performance improvements, especially for vision-language models iii) local descriptors in retrieval re-ranking are still a key ingredient, especially in the presence of severe background clutter iv) the text-to-image performance of the vision-language foundation models is surprisingly close to the corresponding image-to-image case. website: https://vrg.fel.cvut.cz/ilias/
comment: CVPR 2025
Machine Learning 173
☆ Understanding R1-Zero-Like Training: A Critical Perspective
DeepSeek-R1-Zero has shown that reinforcement learning (RL) at scale can directly enhance the reasoning capabilities of LLMs without supervised fine-tuning. In this work, we critically examine R1-Zero-like training by analyzing its two core components: base models and RL. We investigate a wide range of base models, including DeepSeek-V3-Base, to understand how pretraining characteristics influence RL performance. Our analysis reveals that DeepSeek-V3-Base already exhibit ''Aha moment'', while Qwen2.5 base models demonstrate strong reasoning capabilities even without prompt templates, suggesting potential pretraining biases. Additionally, we identify an optimization bias in Group Relative Policy Optimization (GRPO), which artificially increases response length (especially for incorrect outputs) during training. To address this, we introduce Dr. GRPO, an unbiased optimization method that improves token efficiency while maintaining reasoning performance. Leveraging these insights, we present a minimalist R1-Zero recipe that achieves 43.3% accuracy on AIME 2024 with a 7B base model, establishing a new state-of-the-art. Our code is available at https://github.com/sail-sg/understand-r1-zero.
☆ Zero-Shot Audio-Visual Editing via Cross-Modal Delta Denoising
In this paper, we introduce zero-shot audio-video editing, a novel task that requires transforming original audio-visual content to align with a specified textual prompt without additional model training. To evaluate this task, we curate a benchmark dataset, AvED-Bench, designed explicitly for zero-shot audio-video editing. AvED-Bench includes 110 videos, each with a 10-second duration, spanning 11 categories from VGGSound. It offers diverse prompts and scenarios that require precise alignment between auditory and visual elements, enabling robust evaluation. We identify limitations in existing zero-shot audio and video editing methods, particularly in synchronization and coherence between modalities, which often result in inconsistent outcomes. To address these challenges, we propose AvED, a zero-shot cross-modal delta denoising framework that leverages audio-video interactions to achieve synchronized and coherent edits. AvED demonstrates superior results on both AvED-Bench and the recent OAVE dataset to validate its generalization capabilities. Results are available at https://genjib.github.io/project_page/AVED/index.html
comment: Project page: https://genjib.github.io/project_page/AVED/index.html
☆ An Empirical Study of the Impact of Federated Learning on Machine Learning Model Accuracy
Federated Learning (FL) enables distributed ML model training on private user data at the global scale. Despite the potential of FL demonstrated in many domains, an in-depth view of its impact on model accuracy remains unclear. In this paper, we investigate, systematically, how this learning paradigm can affect the accuracy of state-of-the-art ML models for a variety of ML tasks. We present an empirical study that involves various data types: text, image, audio, and video, and FL configuration knobs: data distribution, FL scale, client sampling, and local and global computations. Our experiments are conducted in a unified FL framework to achieve high fidelity, with substantial human efforts and resource investments. Based on the results, we perform a quantitative analysis of the impact of FL, and highlight challenging scenarios where applying FL degrades the accuracy of the model drastically and identify cases where the impact is negligible. The detailed and extensive findings can benefit practical deployments and future development of FL.
☆ Reliable algorithm selection for machine learning-guided design
Algorithms for machine learning-guided design, or design algorithms, use machine learning-based predictions to propose novel objects with desired property values. Given a new design task -- for example, to design novel proteins with high binding affinity to a therapeutic target -- one must choose a design algorithm and specify any hyperparameters and predictive and/or generative models involved. How can these decisions be made such that the resulting designs are successful? This paper proposes a method for design algorithm selection, which aims to select design algorithms that will produce a distribution of design labels satisfying a user-specified success criterion -- for example, that at least ten percent of designs' labels exceed a threshold. It does so by combining designs' predicted property values with held-out labeled data to reliably forecast characteristics of the label distributions produced by different design algorithms, building upon techniques from prediction-powered inference. The method is guaranteed with high probability to return design algorithms that yield successful label distributions (or the null set if none exist), if the density ratios between the design and labeled data distributions are known. We demonstrate the method's effectiveness in simulated protein and RNA design tasks, in settings with either known or estimated density ratios.
comment: 25 pages, 7 figures
☆ ASGO: Adaptive Structured Gradient Optimization
Training deep neural networks (DNNs) is a structured optimization problem, because the parameters are naturally represented by matrices and tensors rather than simple vectors. Under this structural representation, it has been widely observed that gradients are low-rank and Hessians are approximately block-wise diagonal. These structured properties are crucial for designing efficient optimization algorithms but may not be utilized by current popular optimizers like Adam. In this paper, we present a novel optimization algorithm ASGO that capitalizes on these properties by employing a preconditioner that is adaptively updated using structured gradients. By fine-grained theoretical analysis, ASGO is proven to achieve superior convergence rates compared to existing structured gradient methods. Based on the convergence theory, we further demonstrate that ASGO can benefit from the low-rank and block-wise diagonal properties. We also discuss practical modifications of ASGO and empirically verify the effectiveness of the algorithm on language model tasks.
comment: 25 pages, 4 figures
☆ ADS-Edit: A Multimodal Knowledge Editing Dataset for Autonomous Driving Systems
Recent advancements in Large Multimodal Models (LMMs) have shown promise in Autonomous Driving Systems (ADS). However, their direct application to ADS is hindered by challenges such as misunderstanding of traffic knowledge, complex road conditions, and diverse states of vehicle. To address these challenges, we propose the use of Knowledge Editing, which enables targeted modifications to a model's behavior without the need for full retraining. Meanwhile, we introduce ADS-Edit, a multimodal knowledge editing dataset specifically designed for ADS, which includes various real-world scenarios, multiple data types, and comprehensive evaluation metrics. We conduct comprehensive experiments and derive several interesting conclusions. We hope that our work will contribute to the further advancement of knowledge editing applications in the field of autonomous driving. Code and data are available in https://github.com/zjunlp/EasyEdit.
comment: Work in progress
☆ Optimal Scaling Laws for Efficiency Gains in a Theoretical Transformer-Augmented Sectional MoE Framework
This paper introduces a theoretical framework for a Transformer-augmented, sectional Mixture-of-Experts (MoE) architecture that aims to enhance computational efficiency while preserving model scalability. Unlike conventional MoE models, which route entire token embeddings to selected experts, our approach portions the embedding dimension itself -- assigning segments of each token's representation to dedicated experts. To combat losses in token representation, we utilize a pre-expert transformer layer to recompute attention across tokens and reduce the sequence length dimensionality. We extend our theory by deriving optimal scaling laws that a non-linear relationship between the number of experts and factors such as model dimensionality, sequence length, and system overhead. These formulations yield closed-form and numerically-solvable expressions for identifying the optimal expert count under given architectural and hardware constraints. As a result, our framework not only provides theoretical bounds for computing efficiency with varying frameworks but also guides practical design choices for scaling large models effectively. While empirical validation is pending, we present a comprehensive experimental road map to evaluate the framework's efficiency, scalability, and practicality in future work.
☆ Quantum Neural Network Restatement of Markov Jump Process
Despite the many challenges in exploratory data analysis, artificial neural networks have motivated strong interests in scientists and researchers both in theoretical as well as practical applications. Among sources of such popularity of artificial neural networks the ability of modeling non-linear dynamical systems, generalization, and adaptation possibilities should be mentioned. Despite this, there is still significant debate about the role of various underlying stochastic processes in stabilizing a unique structure for data learning and prediction. One of such obstacles to the theoretical and numerical study of machine intelligent systems is the curse of dimensionality and the sampling from high-dimensional probability distributions. In general, this curse prevents efficient description of states, providing a significant complexity barrier for the system to be efficiently described and studied. In this strand of research, direct treatment and description of such abstract notions of learning theory in terms of quantum information be one of the most favorable candidates. Hence, the subject matter of these articles is devoted to problems of design, adaptation and the formulations of computationally hard problems in terms of quantum mechanical systems. In order to characterize the microscopic description of such dynamics in the language of inferential statistics, covariance matrix estimation of d-dimensional Gaussian densities and Bayesian interpretation of eigenvalue problem for dynamical systems is assessed.
☆ RecTable: Fast Modeling Tabular Data with Rectified Flow
Score-based or diffusion models generate high-quality tabular data, surpassing GAN-based and VAE-based models. However, these methods require substantial training time. In this paper, we introduce RecTable, which uses the rectified flow modeling, applied in such as text-to-image generation and text-to-video generation. RecTable features a simple architecture consisting of a few stacked gated linear unit blocks. Additionally, our training strategies are also simple, incorporating a mixed-type noise distribution and a logit-normal timestep distribution. Our experiments demonstrate that RecTable achieves competitive performance compared to the several state-of-the-art diffusion and score-based models while reducing the required training time. Our code is available at https://github.com/fmp453/rectable.
comment: 19 pages, 7 figures, 10 tables
☆ Benchmarking and optimizing organism wide single-cell RNA alignment methods ICLR 2025
Many methods have been proposed for removing batch effects and aligning single-cell RNA (scRNA) datasets. However, performance is typically evaluated based on multiple parameters and few datasets, creating challenges in assessing which method is best for aligning data at scale. Here, we introduce the K-Neighbors Intersection (KNI) score, a single score that both penalizes batch effects and measures accuracy at cross-dataset cell-type label prediction alongside carefully curated small (scMARK) and large (scREF) benchmarks comprising 11 and 46 human scRNA studies respectively, where we have standardized author labels. Using the KNI score, we evaluate and optimize approaches for cross-dataset single-cell RNA integration. We introduce Batch Adversarial single-cell Variational Inference (BA-scVI), as a new variant of scVI that uses adversarial training to penalize batch-effects in the encoder and decoder, and show this approach outperforms other methods. In the resulting aligned space, we find that the granularity of cell-type groupings is conserved, supporting the notion that whole-organism cell-type maps can be created by a single model without loss of information.
comment: Accepted to ICLR 2025 LMRL workshop (International Conference on Learning Representations, Learning Meaningful Representations of Life Workshop)
☆ Continual learning via probabilistic exchangeable sequence modelling
Continual learning (CL) refers to the ability to continuously learn and accumulate new knowledge while retaining useful information from past experiences. Although numerous CL methods have been proposed in recent years, it is not straightforward to deploy them directly to real-world decision-making problems due to their computational cost and lack of uncertainty quantification. To address these issues, we propose CL-BRUNO, a probabilistic, Neural Process-based CL model that performs scalable and tractable Bayesian update and prediction. Our proposed approach uses deep-generative models to create a unified probabilistic framework capable of handling different types of CL problems such as task- and class-incremental learning, allowing users to integrate information across different CL scenarios using a single model. Our approach is able to prevent catastrophic forgetting through distributional and functional regularisation without the need of retaining any previously seen samples, making it appealing to applications where data privacy or storage capacity is of concern. Experiments show that CL-BRUNO outperforms existing methods on both natural image and biomedical data sets, confirming its effectiveness in real-world applications.
☆ A weakly-supervised deep learning model for fast localisation and delineation of the skeleton, internal organs, and spinal canal on Whole-Body Diffusion-Weighted MRI (WB-DWI)
Background: Apparent Diffusion Coefficient (ADC) values and Total Diffusion Volume (TDV) from Whole-body diffusion-weighted MRI (WB-DWI) are recognized cancer imaging biomarkers. However, manual disease delineation for ADC and TDV measurements is unfeasible in clinical practice, demanding automation. As a first step, we propose an algorithm to generate fast and reproducible probability maps of the skeleton, adjacent internal organs (liver, spleen, urinary bladder, and kidneys), and spinal canal. Methods: We developed an automated deep-learning pipeline based on a 3D patch-based Residual U-Net architecture that localizes and delineates these anatomical structures on WB-DWI. The algorithm was trained using "soft-labels" (non-binary segmentations) derived from a computationally intensive atlas-based approach. For training and validation, we employed a multi-center WB-DWI dataset comprising 532 scans from patients with Advanced Prostate Cancer (APC) or Multiple Myeloma (MM), with testing on 45 patients. Results: Our weakly-supervised deep learning model achieved an average dice score/precision/recall of 0.66/0.6/0.73 for skeletal delineations, 0.8/0.79/0.81 for internal organs, and 0.85/0.79/0.94 for spinal canal, with surface distances consistently below 3 mm. Relative median ADC and log-transformed volume differences between automated and manual expert-defined full-body delineations were below 10% and 4%, respectively. The computational time for generating probability maps was 12x faster than the atlas-based registration algorithm (25 s vs. 5 min). An experienced radiologist rated the model's accuracy "good" or "excellent" on test datasets. Conclusion: Our model offers fast and reproducible probability maps for localizing and delineating body regions on WB-DWI, enabling ADC and TDV quantification, potentially supporting clinicians in disease staging and treatment response assessment.
☆ Learning Straight Flows by Learning Curved Interpolants ICLR 2025
Flow matching models typically use linear interpolants to define the forward/noise addition process. This, together with the independent coupling between noise and target distributions, yields a vector field which is often non-straight. Such curved fields lead to a slow inference/generation process. In this work, we propose to learn flexible (potentially curved) interpolants in order to learn straight vector fields to enable faster generation. We formulate this via a multi-level optimization problem and propose an efficient approximate procedure to solve it. Our framework provides an end-to-end and simulation-free optimization procedure, which can be leveraged to learn straight line generative trajectories.
comment: Delta workshop at ICLR 2025
☆ Demand Estimation with Text and Image Data
We propose a demand estimation method that leverages unstructured text and image data to infer substitution patterns. Using pre-trained deep learning models, we extract embeddings from product images and textual descriptions and incorporate them into a random coefficients logit model. This approach enables researchers to estimate demand even when they lack data on product attributes or when consumers value hard-to-quantify attributes, such as visual design or functional benefits. Using data from a choice experiment, we show that our approach outperforms standard attribute-based models in counterfactual predictions of consumers' second choices. We also apply it across 40 product categories on Amazon.com and consistently find that text and image data help identify close substitutes within each category.
☆ Semi-supervised Node Importance Estimation with Informative Distribution Modeling for Uncertainty Regularization
Node importance estimation, a classical problem in network analysis, underpins various web applications. Previous methods either exploit intrinsic topological characteristics, e.g., graph centrality, or leverage additional information, e.g., data heterogeneity, for node feature enhancement. However, these methods follow the supervised learning setting, overlooking the fact that ground-truth node-importance data are usually partially labeled in practice. In this work, we propose the first semi-supervised node importance estimation framework, i.e., EASING, to improve learning quality for unlabeled data in heterogeneous graphs. Different from previous approaches, EASING explicitly captures uncertainty to reflect the confidence of model predictions. To jointly estimate the importance values and uncertainties, EASING incorporates DJE, a deep encoder-decoder neural architecture. DJE introduces distribution modeling for graph nodes, where the distribution representations derive both importance and uncertainty estimates. Additionally, DJE facilitates effective pseudo-label generation for the unlabeled data to enrich the training samples. Based on labeled and pseudo-labeled data, EASING develops effective semi-supervised heteroscedastic learning with varying node uncertainty regularization. Extensive experiments on three real-world datasets highlight the superior performance of EASING compared to competing methods. Codes are available via https://github.com/yankai-chen/EASING.
☆ A Low-complexity Structured Neural Network Approach to Intelligently Realize Wideband Multi-beam Beamformers
True-time-delay (TTD) beamformers can produce wideband, squint-free beams in both analog and digital signal domains, unlike frequency-dependent FFT beams. Our previous work showed that TTD beamformers can be efficiently realized using the elements of delay Vandermonde matrix (DVM), answering the longstanding beam-squint problem. Thus, building on our work on classical algorithms based on DVM, we propose neural network (NN) architecture to realize wideband multi-beam beamformers using structure-imposed weight matrices and submatrices. The structure and sparsity of the weight matrices and submatrices are shown to reduce the space and computational complexities of the NN greatly. The proposed network architecture has O(pLM logM) complexity compared to a conventional fully connected L-layers network with O(M2L) complexity, where M is the number of nodes in each layer of the network, p is the number of submatrices per layer, and M >> p. We will show numerical simulations in the 24 GHz to 32 GHz range to demonstrate the numerical feasibility of realizing wideband multi-beam beamformers using the proposed neural architecture. We also show the complexity reduction of the proposed NN and compare that with fully connected NNs, to show the efficiency of the proposed architecture without sacrificing accuracy. The accuracy of the proposed NN architecture was shown using the mean squared error, which is based on an objective function of the weight matrices and beamformed signals of antenna arrays, while also normalizing nodes. The proposed NN architecture shows a low-complexity NN realizing wideband multi-beam beamformers in real-time for low-complexity intelligent systems.
comment: 10 pages, 3 figures
☆ Flip Learning: Weakly Supervised Erase to Segment Nodules in Breast Ultrasound
Accurate segmentation of nodules in both 2D breast ultrasound (BUS) and 3D automated breast ultrasound (ABUS) is crucial for clinical diagnosis and treatment planning. Therefore, developing an automated system for nodule segmentation can enhance user independence and expedite clinical analysis. Unlike fully-supervised learning, weakly-supervised segmentation (WSS) can streamline the laborious and intricate annotation process. However, current WSS methods face challenges in achieving precise nodule segmentation, as many of them depend on inaccurate activation maps or inefficient pseudo-mask generation algorithms. In this study, we introduce a novel multi-agent reinforcement learning-based WSS framework called Flip Learning, which relies solely on 2D/3D boxes for accurate segmentation. Specifically, multiple agents are employed to erase the target from the box to facilitate classification tag flipping, with the erased region serving as the predicted segmentation mask. The key contributions of this research are as follows: (1) Adoption of a superpixel/supervoxel-based approach to encode the standardized environment, capturing boundary priors and expediting the learning process. (2) Introduction of three meticulously designed rewards, comprising a classification score reward and two intensity distribution rewards, to steer the agents' erasing process precisely, thereby avoiding both under- and over-segmentation. (3) Implementation of a progressive curriculum learning strategy to enable agents to interact with the environment in a progressively challenging manner, thereby enhancing learning efficiency. Extensively validated on the large in-house BUS and ABUS datasets, our Flip Learning method outperforms state-of-the-art WSS methods and foundation models, and achieves comparable performance as fully-supervised learning algorithms.
comment: Accepted by Medical Image Analysis. 24 pages, 13 figures, 18 tabels
☆ Benchmarking Machine Learning Methods for Distributed Acoustic Sensing
Distributed acoustic sensing (DAS) technology represents an innovative fiber-optic-based sensing methodology that enables real-time acoustic signal monitoring through the detection of minute perturbations along optical fibers. This sensing approach offers compelling advantages, including extensive measurement ranges, exceptional spatial resolution, and an expansive dynamic measurement spectrum. The integration of machine learning (ML) paradigms presents transformative potential for DAS technology, encompassing critical domains such as data augmentation, sophisticated preprocessing techniques, and advanced acoustic event classification and recognition. By leveraging ML algorithms, DAS systems can transition from traditional data processing methodologies to more automated and intelligent analytical frameworks. The computational intelligence afforded by ML-enhanced DAS technologies facilitates unprecedented monitoring capabilities across diverse critical infrastructure sectors. Particularly noteworthy are the technology's applications in transportation infrastructure, energy management systems, and Natural disaster monitoring frameworks, where the precision of data acquisition and the reliability of intelligent decision-making mechanisms are paramount. This research critically examines the comparative performance characteristics of classical machine learning methodologies and state-of-the-art deep learning models in the context of DAS data recognition and interpretation, offering comprehensive insights into the evolving landscape of intelligent sensing technologies.
☆ Asset price movement prediction using empirical mode decomposition and Gaussian mixture models
We investigated the use of Empirical Mode Decomposition (EMD) combined with Gaussian Mixture Models (GMM), feature engineering and machine learning algorithms to optimize trading decisions. We used five, two, and one year samples of hourly candle data for GameStop, Tesla, and XRP (Ripple) markets respectively. Applying a 15 hour rolling window for each market, we collected several features based on a linear model and other classical features to predict the next hour's movement. Subsequently, a GMM filtering approach was used to identify clusters among these markets. For each cluster, we applied the EMD algorithm to extract high, medium, low and trend components from each feature collected. A simple thresholding algorithm was applied to classify market movements based on the percentage change in each market's close price. We then evaluated the performance of various machine learning models, including Random Forests (RF) and XGBoost, in classifying market movements. A naive random selection of trading decisions was used as a benchmark, which assumed equal probabilities for each outcome, and a temporal cross-validation approach was used to test models on 40%, 30%, and 20% of the dataset. Our results indicate that transforming selected features using EMD improves performance, particularly for ensemble learning algorithms like Random Forest and XGBoost, as measured by accumulated profit. Finally, GMM filtering expanded the range of learning algorithm and data source combinations that outperformed the top percentile of the random baseline.
comment: 21 pages
☆ Inductive Link Prediction on N-ary Relational Facts via Semantic Hypergraph Reasoning KDD
N-ary relational facts represent semantic correlations among more than two entities. While recent studies have developed link prediction (LP) methods to infer missing relations for knowledge graphs (KGs) containing n-ary relational facts, they are generally limited to transductive settings. Fully inductive settings, where predictions are made on previously unseen entities, remain a significant challenge. As existing methods are mainly entity embedding-based, they struggle to capture entity-independent logical rules. To fill in this gap, we propose an n-ary subgraph reasoning framework for fully inductive link prediction (ILP) on n-ary relational facts. This framework reasons over local subgraphs and has a strong inductive inference ability to capture n-ary patterns. Specifically, we introduce a novel graph structure, the n-ary semantic hypergraph, to facilitate subgraph extraction. Moreover, we develop a subgraph aggregating network, NS-HART, to effectively mine complex semantic correlations within subgraphs. Theoretically, we provide a thorough analysis from the score function optimization perspective to shed light on NS-HART's effectiveness for n-ary ILP tasks. Empirically, we conduct extensive experiments on a series of inductive benchmarks, including transfer reasoning (with and without entity features) and pairwise subgraph reasoning. The results highlight the superiority of the n-ary subgraph reasoning framework and the exceptional inductive ability of NS-HART. The source code of this paper has been made publicly available at https://github.com/yin-gz/Nary-Inductive-SubGraph.
comment: To be published in Proceedings of the 31st ACM SIGKDD Conference on Knowledge Discovery and Data Mining V.1 (KDD'25)
☆ AutoRad-Lung: A Radiomic-Guided Prompting Autoregressive Vision-Language Model for Lung Nodule Malignancy Prediction
Lung cancer remains one of the leading causes of cancer-related mortality worldwide. A crucial challenge for early diagnosis is differentiating uncertain cases with similar visual characteristics and closely annotation scores. In clinical practice, radiologists rely on quantitative, hand-crafted Radiomic features extracted from Computed Tomography (CT) images, while recent research has primarily focused on deep learning solutions. More recently, Vision-Language Models (VLMs), particularly Contrastive Language-Image Pre-Training (CLIP)-based models, have gained attention for their ability to integrate textual knowledge into lung cancer diagnosis. While CLIP-Lung models have shown promising results, we identified the following potential limitations: (a) dependence on radiologists' annotated attributes, which are inherently subjective and error-prone, (b) use of textual information only during training, limiting direct applicability at inference, and (c) Convolutional-based vision encoder with randomly initialized weights, which disregards prior knowledge. To address these limitations, we introduce AutoRad-Lung, which couples an autoregressively pre-trained VLM, with prompts generated from hand-crafted Radiomics. AutoRad-Lung uses the vision encoder of the Large-Scale Autoregressive Image Model (AIMv2), pre-trained using a multi-modal autoregressive objective. Given that lung tumors are typically small, irregularly shaped, and visually similar to healthy tissue, AutoRad-Lung offers significant advantages over its CLIP-based counterparts by capturing pixel-level differences. Additionally, we introduce conditional context optimization, which dynamically generates context-specific prompts based on input Radiomics, improving cross-modal alignment.
☆ DR-PETS: Learning-Based Control With Planning in Adversarial Environments
Ensuring robustness against epistemic, possibly adversarial, perturbations is essential for reliable real-world decision-making. While the Probabilistic Ensembles with Trajectory Sampling (PETS) algorithm inherently handles uncertainty via ensemble-based probabilistic models, it lacks guarantees against structured adversarial or worst-case uncertainty distributions. To address this, we propose DR-PETS, a distributionally robust extension of PETS that certifies robustness against adversarial perturbations. We formalize uncertainty via a p-Wasserstein ambiguity set, enabling worst-case-aware planning through a min-max optimization framework. While PETS passively accounts for stochasticity, DR-PETS actively optimizes robustness via a tractable convex approximation integrated into PETS planning loop. Experiments on pendulum stabilization and cart-pole balancing show that DR-PETS certifies robustness against adversarial parameter perturbations, achieving consistent performance in worst-case scenarios where PETS deteriorates.
comment: 6 pages, 2 figures, submitted to LCSS
☆ Probabilistic Forecasting for Network Resource Analysis in Integrated Terrestrial and Non-Terrestrial Networks
Efficient resource management is critical for Non-Terrestrial Networks (NTNs) to provide consistent, high-quality service in remote and under-served regions. While traditional single-point prediction methods, such as Long-Short Term Memory (LSTM), have been used in terrestrial networks, they often fall short in NTNs due to the complexity of satellite dynamics, signal latency and coverage variability. Probabilistic forecasting, which quantifies the uncertainties of the predictions, is a robust alternative. In this paper, we evaluate the application of probabilistic forecasting techniques, in particular SFF, to NTN resource allocation scenarios. Our results show their effectiveness in predicting bandwidth and capacity requirements in different NTN segments of probabilistic forecasting compared to single-point prediction techniques such as LSTM. The results show the potential of black probabilistic forecasting models to provide accurate and reliable predictions and to quantify their uncertainty, making them indispensable for optimizing NTN resource allocation. At the end of the paper, we also present application scenarios and a standardization roadmap for the use of probabilistic forecasting in integrated Terrestrial Network (TN)-NTN environments.
☆ PVLens: Enhancing Pharmacovigilance Through Automated Label Extraction
Reliable drug safety reference databases are essential for pharmacovigilance, yet existing resources like SIDER are outdated and static. We introduce PVLens, an automated system that extracts labeled safety information from FDA Structured Product Labels (SPLs) and maps terms to MedDRA. PVLens integrates automation with expert oversight through a web-based review tool. In validation against 97 drug labels, PVLens achieved an F1 score of 0.882, with high recall (0.983) and moderate precision (0.799). By offering a scalable, more accurate and continuously updated alternative to SIDER, PVLens enhances real-time pharamcovigilance with improved accuracy and contemporaneous insights.
☆ Enhancing Multi-modal Models with Heterogeneous MoE Adapters for Fine-tuning
Multi-modal models excel in cross-modal tasks but are computationally expensive due to their billions of parameters. Parameter-efficient fine-tuning (PEFT) offers a solution by adding small trainable components while freezing pre-trained parameters. However, existing methods primarily focus on uni-modal processing, overlooking the critical modal fusion needed for multi-modal tasks. To fill this gap, we propose heterogeneous mixture of experts adapters that extend the traditional PEFT framework to support multi-modal expert combinations and improve information interaction. Additionally, our approach modifies the affine linear expert design to enable efficient modal fusion in a low-rank space, achieving competitive performance with only 5-8\% of the parameters fine-tuned. Experiments across eight downstream tasks, including visual-audio and text-visual, demonstrate the superior performance of the approach.
comment: 6 pages,3 figures
☆ $β$-GNN: A Robust Ensemble Approach Against Graph Structure Perturbation
Graph Neural Networks (GNNs) are playing an increasingly important role in the efficient operation and security of computing systems, with applications in workload scheduling, anomaly detection, and resource management. However, their vulnerability to network perturbations poses a significant challenge. We propose $\beta$-GNN, a model enhancing GNN robustness without sacrificing clean data performance. $\beta$-GNN uses a weighted ensemble, combining any GNN with a multi-layer perceptron. A learned dynamic weight, $\beta$, modulates the GNN's contribution. This $\beta$ not only weights GNN influence but also indicates data perturbation levels, enabling proactive mitigation. Experimental results on diverse datasets show $\beta$-GNN's superior adversarial accuracy and attack severity quantification. Crucially, $\beta$-GNN avoids perturbation assumptions, preserving clean data structure and performance.
comment: This is the author's version of the paper accepted at EuroMLSys 2025
☆ ProFed: a Benchmark for Proximity-based non-IID Federated Learning
In recent years, cro:flFederated learning (FL) has gained significant attention within the machine learning community. Although various FL algorithms have been proposed in the literature, their performance often degrades when data across clients is non-independently and identically distributed (non-IID). This skewness in data distribution often emerges from geographic patterns, with notable examples including regional linguistic variations in text data or localized traffic patterns in urban environments. Such scenarios result in IID data within specific regions but non-IID data across regions. However, existing FL algorithms are typically evaluated by randomly splitting non-IID data across devices, disregarding their spatial distribution. To address this gap, we introduce ProFed, a benchmark that simulates data splits with varying degrees of skewness across different regions. We incorporate several skewness methods from the literature and apply them to well-known datasets, including MNIST, FashionMNIST, CIFAR-10, and CIFAR-100. Our goal is to provide researchers with a standardized framework to evaluate FL algorithms more effectively and consistently against established baselines.
☆ State-Aware Perturbation Optimization for Robust Deep Reinforcement Learning
Recently, deep reinforcement learning (DRL) has emerged as a promising approach for robotic control. However, the deployment of DRL in real-world robots is hindered by its sensitivity to environmental perturbations. While existing whitebox adversarial attacks rely on local gradient information and apply uniform perturbations across all states to evaluate DRL robustness, they fail to account for temporal dynamics and state-specific vulnerabilities. To combat the above challenge, we first conduct a theoretical analysis of white-box attacks in DRL by establishing the adversarial victim-dynamics Markov decision process (AVD-MDP), to derive the necessary and sufficient conditions for a successful attack. Based on this, we propose a selective state-aware reinforcement adversarial attack method, named STAR, to optimize perturbation stealthiness and state visitation dispersion. STAR first employs a soft mask-based state-targeting mechanism to minimize redundant perturbations, enhancing stealthiness and attack effectiveness. Then, it incorporates an information-theoretic optimization objective to maximize mutual information between perturbations, environmental states, and victim actions, ensuring a dispersed state-visitation distribution that steers the victim agent into vulnerable states for maximum return reduction. Extensive experiments demonstrate that STAR outperforms state-of-the-art benchmarks.
comment: 15 pages, 11 figures
Diffusion Counterfactuals for Image Regressors
Counterfactual explanations have been successfully applied to create human interpretable explanations for various black-box models. They are handy for tasks in the image domain, where the quality of the explanations benefits from recent advances in generative models. Although counterfactual explanations have been widely applied to classification models, their application to regression tasks remains underexplored. We present two methods to create counterfactual explanations for image regression tasks using diffusion-based generative models to address challenges in sparsity and quality: 1) one based on a Denoising Diffusion Probabilistic Model that operates directly in pixel-space and 2) another based on a Diffusion Autoencoder operating in latent space. Both produce realistic, semantic, and smooth counterfactuals on CelebA-HQ and a synthetic data set, providing easily interpretable insights into the decision-making process of the regression model and reveal spurious correlations. We find that for regression counterfactuals, changes in features depend on the region of the predicted value. Large semantic changes are needed for significant changes in predicted values, making it harder to find sparse counterfactuals than with classifiers. Moreover, pixel space counterfactuals are more sparse while latent space counterfactuals are of higher quality and allow bigger semantic changes.
comment: 24 Pages, 5 Figures, Accepted at 3rd World Conference on eXplainable Artificial Intelligence (xAI-2025), Code and reproduction instructions available on GitHub, see https://github.com/DevinTDHa/Diffusion-Counterfactuals-for-Image-Regressors
☆ Feature Statistics with Uncertainty Help Adversarial Robustness
Despite the remarkable success of deep neural networks (DNNs), the security threat of adversarial attacks poses a significant challenge to the reliability of DNNs. By introducing randomness into different parts of DNNs, stochastic methods can enable the model to learn some uncertainty, thereby improving model robustness efficiently. In this paper, we theoretically discover a universal phenomenon that adversarial attacks will shift the distributions of feature statistics. Motivated by this theoretical finding, we propose a robustness enhancement module called Feature Statistics with Uncertainty (FSU). It resamples channel-wise feature means and standard deviations of examples from multivariate Gaussian distributions, which helps to reconstruct the attacked examples and calibrate the shifted distributions. The calibration recovers some domain characteristics of the data for classification, thereby mitigating the influence of perturbations and weakening the ability of attacks to deceive models. The proposed FSU module has universal applicability in training, attacking, predicting and fine-tuning, demonstrating impressive robustness enhancement ability at trivial additional time cost. For example, against powerful optimization-based CW attacks, by incorporating FSU into attacking and predicting phases, it endows many collapsed state-of-the-art models with 50%-80% robust accuracy on CIFAR10, CIFAR100 and SVHN.
☆ Optimizing Case-Based Reasoning System for Functional Test Script Generation with Large Language Models
In this work, we explore the potential of large language models (LLMs) for generating functional test scripts, which necessitates understanding the dynamically evolving code structure of the target software. To achieve this, we propose a case-based reasoning (CBR) system utilizing a 4R cycle (i.e., retrieve, reuse, revise, and retain), which maintains and leverages a case bank of test intent descriptions and corresponding test scripts to facilitate LLMs for test script generation. To improve user experience further, we introduce Re4, an optimization method for the CBR system, comprising reranking-based retrieval finetuning and reinforced reuse finetuning. Specifically, we first identify positive examples with high semantic and script similarity, providing reliable pseudo-labels for finetuning the retriever model without costly labeling. Then, we apply supervised finetuning, followed by a reinforcement learning finetuning stage, to align LLMs with our production scenarios, ensuring the faithful reuse of retrieved cases. Extensive experimental results on two product development units from Huawei Datacom demonstrate the superiority of the proposed CBR+Re4. Notably, we also show that the proposed Re4 method can help alleviate the repetitive generation issues with LLMs.
☆ TerraTorch: The Geospatial Foundation Models Toolkit RSS 2025
TerraTorch is a fine-tuning and benchmarking toolkit for Geospatial Foundation Models built on PyTorch Lightning and tailored for satellite, weather, and climate data. It integrates domain-specific data modules, pre-defined tasks, and a modular model factory that pairs any backbone with diverse decoder heads. These components allow researchers and practitioners to fine-tune supported models in a no-code fashion by simply editing a training configuration. By consolidating best practices for model development and incorporating the automated hyperparameter optimization extension Iterate, TerraTorch reduces the expertise and time required to fine-tune or benchmark models on new Earth Observation use cases. Furthermore, TerraTorch directly integrates with GEO-Bench, allowing for systematic and reproducible benchmarking of Geospatial Foundation Models. TerraTorch is open sourced under Apache 2.0, available at https://github.com/IBM/terratorch, and can be installed via pip install terratorch.
comment: IGARSS 2025
☆ A Theoretical Framework for Prompt Engineering: Approximating Smooth Functions with Transformer Prompts
Prompt engineering has emerged as a powerful technique for guiding large language models (LLMs) toward desired responses, significantly enhancing their performance across diverse tasks. Beyond their role as static predictors, LLMs increasingly function as intelligent agents, capable of reasoning, decision-making, and adapting dynamically to complex environments. However, the theoretical underpinnings of prompt engineering remain largely unexplored. In this paper, we introduce a formal framework demonstrating that transformer models, when provided with carefully designed prompts, can act as a configurable computational system by emulating a ``virtual'' neural network during inference. Specifically, input prompts effectively translate into the corresponding network configuration, enabling LLMs to adjust their internal computations dynamically. Building on this construction, we establish an approximation theory for $\beta$-times differentiable functions, proving that transformers can approximate such functions with arbitrary precision when guided by appropriately structured prompts. Moreover, our framework provides theoretical justification for several empirically successful prompt engineering techniques, including the use of longer, structured prompts, filtering irrelevant information, enhancing prompt token diversity, and leveraging multi-agent interactions. By framing LLMs as adaptable agents rather than static models, our findings underscore their potential for autonomous reasoning and problem-solving, paving the way for more robust and theoretically grounded advancements in prompt engineering and AI agent design.
comment: 55 pages, 2 figures
☆ Injecting Adrenaline into LLM Serving: Boosting Resource Utilization and Throughput via Attention Disaggregation
In large language model (LLM) serving systems, executing each request consists of two phases: the compute-intensive prefill phase and the memory-intensive decoding phase. To prevent performance interference between the two phases, current LLM serving systems typically adopt prefill-decoding disaggregation, where the two phases are split across separate machines. However, we observe this approach leads to significant resource underutilization. Specifically, prefill instances that are compute-intensive suffer from low memory utilization, while decoding instances that are memory-intensive experience low compute utilization. To address this problem, this paper proposes Adrenaline, an attention disaggregation and offloading mechanism designed to enhance resource utilization and performance in LLM serving systems. Adrenaline's key innovation lies in disaggregating part of the attention computation in the decoding phase and offloading them to prefill instances. The memory-bound nature of decoding-phase attention computation inherently enables an effective offloading strategy, yielding two complementary advantages: 1) improved memory capacity and bandwidth utilization in prefill instances, and 2) increased decoding batch sizes that enhance compute utilization in decoding instances, collectively boosting overall system performance. Adrenaline achieves these gains through three key techniques: low-latency decoding synchronization, resource-efficient prefill colocation, and load-aware offloading scheduling. Experimental results show that Adrenaline achieves 2.28x higher memory capacity and 2.07x better memory bandwidth utilization in prefill instances, up to 1.67x improvements in compute utilization for decoding instances, and 1.68x higher overall inference throughput compared to state-of-the-art systems.
comment: 14 pages, 18 figures
☆ Regression-Based Estimation of Causal Effects in the Presence of Selection Bias and Confounding
We consider the problem of estimating the expected causal effect $E[Y|do(X)]$ for a target variable $Y$ when treatment $X$ is set by intervention, focusing on continuous random variables. In settings without selection bias or confounding, $E[Y|do(X)] = E[Y|X]$, which can be estimated using standard regression methods. However, regression fails when systematic missingness induced by selection bias, or confounding distorts the data. Boeken et al. [2023] show that when training data is subject to selection, proxy variables unaffected by this process can, under certain constraints, be used to correct for selection bias to estimate $E[Y|X]$, and hence $E[Y|do(X)]$, reliably. When data is additionally affected by confounding, however, this equality is no longer valid. Building on these results, we consider a more general setting and propose a framework that incorporates both selection bias and confounding. Specifically, we derive theoretical conditions ensuring identifiability and recoverability of causal effects under access to external data and proxy variables. We further introduce a two-step regression estimator (TSR), capable of exploiting proxy variables to adjust for selection bias while accounting for confounding. We show that TSR coincides with prior work if confounding is absent, but achieves a lower variance. Extensive simulation studies validate TSR's correctness for scenarios which may include both selection bias and confounding with proxy variables.
comment: 13 pages plus appendix
☆ Fast, Modular, and Differentiable Framework for Machine Learning-Enhanced Molecular Simulations
We present an end-to-end differentiable molecular simulation framework (DIMOS) for molecular dynamics and Monte Carlo simulations. DIMOS easily integrates machine-learning-based interatomic potentials and implements classical force fields including particle-mesh Ewald electrostatics. Thanks to its modularity, both classical and machine-learning-based approaches can be easily combined into a hybrid description of the system (ML/MM). By supporting key molecular dynamics features such as efficient neighborlists and constraint algorithms for larger time steps, the framework bridges the gap between hand-optimized simulation engines and the flexibility of a PyTorch implementation. The superior performance and the high versatility is probed in different benchmarks and applications, with speed-up factors of up to $170\times$. The advantage of differentiability is demonstrated by an end-to-end optimization of the proposal distribution in a Markov Chain Monte Carlo simulation based on Hamiltonian Monte Carlo. Using these optimized simulation parameters a $3\times$ acceleration is observed in comparison to ad-hoc chosen simulation parameters. The code is available at https://github.com/nec-research/DIMOS.
☆ Harmonia: A Multi-Agent Reinforcement Learning Approach to Data Placement and Migration in Hybrid Storage Systems
Hybrid storage systems (HSS) combine multiple storage devices with diverse characteristics to achieve high performance and capacity at low cost. The performance of an HSS highly depends on the effectiveness of two key policies: (1) the data-placement policy, which determines the best-fit storage device for incoming data, and (2) the data-migration policy, which rearranges stored data across the devices to sustain high HSS performance. Prior works focus on improving only data placement or only data migration in HSS, which leads to sub-optimal HSS performance. Unfortunately, no prior work tries to optimize both policies together. Our goal is to design a holistic data-management technique for HSS that optimizes both data-placement and data-migration policies to fully exploit the potential of an HSS. We propose Harmonia, a multi-agent reinforcement learning (RL)-based data-management technique that employs two light-weight autonomous RL agents, a data-placement agent and a data-migration agent, which adapt their policies for the current workload and HSS configuration, and coordinate with each other to improve overall HSS performance. We evaluate Harmonia on a real HSS with up to four heterogeneous storage devices with diverse characteristics. Our evaluation using 17 data-intensive workloads on performance-optimized (cost-optimized) HSS with two storage devices shows that, on average, Harmonia (1) outperforms the best-performing prior approach by 49.5% (31.7%), (2) bridges the performance gap between the best-performing prior work and Oracle by 64.2% (64.3%). On an HSS with three (four) devices, Harmonia outperforms the best-performing prior work by 37.0% (42.0%). Harmonia's performance benefits come with low latency (240ns for inference) and storage overheads (206 KiB for both RL agents together). We plan to open-source Harmonia's implementation to aid future research on HSS.
☆ Riemannian Optimization on Relaxed Indicator Matrix Manifold
The indicator matrix plays an important role in machine learning, but optimizing it is an NP-hard problem. We propose a new relaxation of the indicator matrix and prove that this relaxation forms a manifold, which we call the Relaxed Indicator Matrix Manifold (RIM manifold). Based on Riemannian geometry, we develop a Riemannian toolbox for optimization on the RIM manifold. Specifically, we provide several methods of Retraction, including a fast Retraction method to obtain geodesics. We point out that the RIM manifold is a generalization of the double stochastic manifold, and it is much faster than existing methods on the double stochastic manifold, which has a complexity of \( \mathcal{O}(n^3) \), while RIM manifold optimization is \( \mathcal{O}(n) \) and often yields better results. We conducted extensive experiments, including image denoising, with millions of variables to support our conclusion, and applied the RIM manifold to Ratio Cut, achieving clustering results that outperform the state-of-the-art methods. Our Code in \href{https://github.com/Yuan-Jinghui/Riemannian-Optimization-on-Relaxed-Indicator-Matrix-Manifold}{https://github.com/Yuan-Jinghui/Riemannian-Optimization-on-Relaxed-Indicator-Matrix-Manifold}.
☆ VPO: Aligning Text-to-Video Generation Models with Prompt Optimization
Video generation models have achieved remarkable progress in text-to-video tasks. These models are typically trained on text-video pairs with highly detailed and carefully crafted descriptions, while real-world user inputs during inference are often concise, vague, or poorly structured. This gap makes prompt optimization crucial for generating high-quality videos. Current methods often rely on large language models (LLMs) to refine prompts through in-context learning, but suffer from several limitations: they may distort user intent, omit critical details, or introduce safety risks. Moreover, they optimize prompts without considering the impact on the final video quality, which can lead to suboptimal results. To address these issues, we introduce VPO, a principled framework that optimizes prompts based on three core principles: harmlessness, accuracy, and helpfulness. The generated prompts faithfully preserve user intents and, more importantly, enhance the safety and quality of generated videos. To achieve this, VPO employs a two-stage optimization approach. First, we construct and refine a supervised fine-tuning (SFT) dataset based on principles of safety and alignment. Second, we introduce both text-level and video-level feedback to further optimize the SFT model with preference learning. Our extensive experiments demonstrate that VPO significantly improves safety, alignment, and video quality compared to baseline methods. Moreover, VPO shows strong generalization across video generation models. Furthermore, we demonstrate that VPO could outperform and be combined with RLHF methods on video generation models, underscoring the effectiveness of VPO in aligning video generation models. Our code and data are publicly available at https://github.com/thu-coai/VPO.
☆ Adaptive Local Clustering over Attributed Graphs ICDE2025
Given a graph $G$ and a seed node $v_s$, the objective of local graph clustering (LGC) is to identify a subgraph $C_s \in G$ (a.k.a. local cluster) surrounding $v_s$ in time roughly linear with the size of $C_s$. This approach yields personalized clusters without needing to access the entire graph, which makes it highly suitable for numerous applications involving large graphs. However, most existing solutions merely rely on the topological connectivity between nodes in $G$, rendering them vulnerable to missing or noisy links that are commonly present in real-world graphs. To address this issue, this paper resorts to leveraging the complementary nature of graph topology and node attributes to enhance local clustering quality. To effectively exploit the attribute information, we first formulate the LGC as an estimation of the bidirectional diffusion distribution (BDD), which is specialized for capturing the multi-hop affinity between nodes in the presence of attributes. Furthermore, we propose LACA, an efficient and effective approach for LGC that achieves superb empirical performance on multiple real datasets while maintaining strong locality. The core components of LACA include (i) a fast and theoretically-grounded preprocessing technique for node attributes, (ii) an adaptive algorithm for diffusing any vectors over $G$ with rigorous theoretical guarantees and expedited convergence, and (iii) an effective three-step scheme for BDD approximation. Extensive experiments, comparing 17 competitors on 8 real datasets, show that LACA outperforms all competitors in terms of result quality measured against ground truth local clusters, while also being up to orders of magnitude faster. The code is available at https://github.com/HaoranZ99/alac.
comment: Accepted by ICDE2025. The code is available at https://github.com/HaoranZ99/alac
☆ Dissecting and Mitigating Diffusion Bias via Mechanistic Interpretability CVPR 2025
Diffusion models have demonstrated impressive capabilities in synthesizing diverse content. However, despite their high-quality outputs, these models often perpetuate social biases, including those related to gender and race. These biases can potentially contribute to harmful real-world consequences, reinforcing stereotypes and exacerbating inequalities in various social contexts. While existing research on diffusion bias mitigation has predominantly focused on guiding content generation, it often neglects the intrinsic mechanisms within diffusion models that causally drive biased outputs. In this paper, we investigate the internal processes of diffusion models, identifying specific decision-making mechanisms, termed bias features, embedded within the model architecture. By directly manipulating these features, our method precisely isolates and adjusts the elements responsible for bias generation, permitting granular control over the bias levels in the generated content. Through experiments on both unconditional and conditional diffusion models across various social bias attributes, we demonstrate our method's efficacy in managing generation distribution while preserving image quality. We also dissect the discovered model mechanism, revealing different intrinsic features controlling fine-grained aspects of generation, boosting further research on mechanistic interpretability of diffusion models.
comment: CVPR 2025; Project Page: https://foundation-model-research.github.io/difflens
☆ Data-driven Seasonal Climate Predictions via Variational Inference and Transformers
Most operational climate services providers base their seasonal predictions on initialised general circulation models (GCMs) or statistical techniques that fit past observations. GCMs require substantial computational resources, which limits their capacity. In contrast, statistical methods often lack robustness due to short historical records. Recent works propose machine learning methods trained on climate model output, leveraging larger sample sizes and simulated scenarios. Yet, many of these studies focus on prediction tasks that might be restricted in spatial extent or temporal coverage, opening a gap with existing operational predictions. Thus, the present study evaluates the effectiveness of a methodology that combines variational inference with transformer models to predict fields of seasonal anomalies. The predictions cover all four seasons and are initialised one month before the start of each season. The model was trained on climate model output from CMIP6 and tested using ERA5 reanalysis data. We analyse the method's performance in predicting interannual anomalies beyond the climate change-induced trend. We also test the proposed methodology in a regional context with a use case focused on Europe. While climate change trends dominate the skill of temperature predictions, the method presents additional skill over the climatological forecast in regions influenced by known teleconnections. We reach similar conclusions based on the validation of precipitation predictions. Despite underperforming SEAS5 in most tropics, our model offers added value in numerous extratropical inland regions. This work demonstrates the effectiveness of training generative models on climate model output for seasonal predictions, providing skilful predictions beyond the induced climate change trend at time scales and lead times relevant for user applications.
☆ Lipschitz Constant Meets Condition Number: Learning Robust and Compact Deep Neural Networks
Recent research has revealed that high compression of Deep Neural Networks (DNNs), e.g., massive pruning of the weight matrix of a DNN, leads to a severe drop in accuracy and susceptibility to adversarial attacks. Integration of network pruning into an adversarial training framework has been proposed to promote adversarial robustness. It has been observed that a highly pruned weight matrix tends to be ill-conditioned, i.e., increasing the condition number of the weight matrix. This phenomenon aggravates the vulnerability of a DNN to input noise. Although a highly pruned weight matrix is considered to be able to lower the upper bound of the local Lipschitz constant to tolerate large distortion, the ill-conditionedness of such a weight matrix results in a non-robust DNN model. To overcome this challenge, this work develops novel joint constraints to adjust the weight distribution of networks, namely, the Transformed Sparse Constraint joint with Condition Number Constraint (TSCNC), which copes with smoothing distribution and differentiable constraint functions to reduce condition number and thus avoid the ill-conditionedness of weight matrices. Furthermore, our theoretical analyses unveil the relevance between the condition number and the local Lipschitz constant of the weight matrix, namely, the sharply increasing condition number becomes the dominant factor that restricts the robustness of over-sparsified models. Extensive experiments are conducted on several public datasets, and the results show that the proposed constraints significantly improve the robustness of a DNN with high pruning rates.
comment: 13 pages, 6 figures
☆ The Crucial Role of Problem Formulation in Real-World Reinforcement Learning
Reinforcement Learning (RL) offers promising solutions for control tasks in industrial cyber-physical systems (ICPSs), yet its real-world adoption remains limited. This paper demonstrates how seemingly small but well-designed modifications to the RL problem formulation can substantially improve performance, stability, and sample efficiency. We identify and investigate key elements of RL problem formulation and show that these enhance both learning speed and final policy quality. Our experiments use a one-degree-of-freedom (1-DoF) helicopter testbed, the Quanser Aero~2, which features non-linear dynamics representative of many industrial settings. In simulation, the proposed problem design principles yield more reliable and efficient training, and we further validate these results by training the agent directly on physical hardware. The encouraging real-world outcomes highlight the potential of RL for ICPS, especially when careful attention is paid to the design principles of problem formulation. Overall, our study underscores the crucial role of thoughtful problem formulation in bridging the gap between RL research and the demands of real-world industrial systems.
comment: Accepted at ICPS 2025
☆ TempTest: Local Normalization Distortion and the Detection of Machine-generated Text
Existing methods for the zero-shot detection of machine-generated text are dominated by three statistical quantities: log-likelihood, log-rank, and entropy. As language models mimic the distribution of human text ever closer, this will limit our ability to build effective detection algorithms. To combat this, we introduce a method for detecting machine-generated text that is entirely agnostic of the generating language model. This is achieved by targeting a defect in the way that decoding strategies, such as temperature or top-k sampling, normalize conditional probability measures. This method can be rigorously theoretically justified, is easily explainable, and is conceptually distinct from existing methods for detecting machine-generated text. We evaluate our detector in the white and black box settings across various language models, datasets, and passage lengths. We also study the effect of paraphrasing attacks on our detector and the extent to which it is biased against non-native speakers. In each of these settings, the performance of our test is at least comparable to that of other state-of-the-art text detectors, and in some cases, we strongly outperform these baselines.
☆ Active Data Sampling and Generation for Bias Remediation
Adequate sampling space coverage is the keystone to effectively train trustworthy Machine Learning models. Unfortunately, real data do carry several inherent risks due to the many potential biases they exhibit when gathered without a proper random sampling over the reference population, and most of the times this is way too expensive or time consuming to be a viable option. Depending on how training data have been gathered, unmitigated biases can lead to harmful or discriminatory consequences that ultimately hinders large scale applicability of pre-trained models and undermine their truthfulness or fairness expectations. In this paper, a mixed active sampling and data generation strategy -- called samplation -- is proposed as a mean to compensate during fine-tuning of a pre-trained classifer the unfair classifications it produces, assuming that the training data come from a non-probabilistic sampling schema. Given a pre-trained classifier, first a fairness metric is evaluated on a test set, then new reservoirs of labeled data are generated and finally a number of reversely-biased artificial samples are generated for the fine-tuning of the model. Using as case study Deep Models for visual semantic role labeling, the proposed method has been able to fully cure a simulated gender bias starting from a 90/10 imbalance, with only a small percentage of new data and with a minor effect on accuracy.
☆ Learning Data-Driven Uncertainty Set Partitions for Robust and Adaptive Energy Forecasting with Missing Data
Short-term forecasting models typically assume the availability of input data (features) when they are deployed and in use. However, equipment failures, disruptions, cyberattacks, may lead to missing features when such models are used operationally, which could negatively affect forecast accuracy, and result in suboptimal operational decisions. In this paper, we use adaptive robust optimization and adversarial machine learning to develop forecasting models that seamlessly handle missing data operationally. We propose linear- and neural network-based forecasting models with parameters that adapt to available features, combining linear adaptation with a novel algorithm for learning data-driven uncertainty set partitions. The proposed adaptive models do not rely on identifying historical missing data patterns and are suitable for real-time operations under stringent time constraints. Extensive numerical experiments on short-term wind power forecasting considering horizons from 15 minutes to 4 hours ahead illustrate that our proposed adaptive models are on par with imputation when data are missing for very short periods (e.g., when only the latest measurement is missing) whereas they significantly outperform imputation when data are missing for longer periods. We further provide insights by showcasing how linear adaptation and data-driven partitions (even with a few subsets) approach the performance of the optimal, yet impractical, method of retraining for every possible realization of missing data.
comment: Submitted to IEEE-TSG
☆ Comparative analysis and evaluation of ageing forecasting methods for semiconductor devices in online health monitoring
Semiconductor devices, especially MOSFETs (Metal-oxide-semiconductor field-effect transistor), are crucial in power electronics, but their reliability is affected by aging processes influenced by cycling and temperature. The primary aging mechanism in discrete semiconductors and power modules is the bond wire lift-off, caused by crack growth due to thermal fatigue. The process is empirically characterized by exponential growth and an abrupt end of life, making long-term aging forecasts challenging. This research presents a comprehensive comparative assessment of different forecasting methods for MOSFET failure forecasting applications. Classical tracking, statistical forecasting and Neural Network (NN) based forecasting models are implemented along with novel Temporal Fusion Transformers (TFTs). A comprehensive comparison is performed assessing their MOSFET ageing forecasting ability for different forecasting horizons. For short-term predictions, all algorithms result in acceptable results, with the best results produced by classical NN forecasting models at the expense of higher computations. For long-term forecasting, only the TFT is able to produce valid outcomes owing to the ability to integrate covariates from the expected future conditions. Additionally, TFT attention points identify key ageing turning points, which indicate new failure modes or accelerated ageing phases.
comment: 28 pages, 12 figures, published
☆ Multi-dataset and Transfer Learning Using Gene Expression Knowledge Graphs
Gene expression datasets offer insights into gene regulation mechanisms, biochemical pathways, and cellular functions. Additionally, comparing gene expression profiles between disease and control patients can deepen the understanding of disease pathology. Therefore, machine learning has been used to process gene expression data, with patient diagnosis emerging as one of the most popular applications. Although gene expression data can provide valuable insights, challenges arise because the number of patients in expression datasets is usually limited, and the data from different datasets with different gene expressions cannot be easily combined. This work proposes a novel methodology to address these challenges by integrating multiple gene expression datasets and domain-specific knowledge using knowledge graphs, a unique tool for biomedical data integration. Then, vector representations are produced using knowledge graph embedding techniques, which are used as inputs for a graph neural network and a multi-layer perceptron. We evaluate the efficacy of our methodology in three settings: single-dataset learning, multi-dataset learning, and transfer learning. The experimental results show that combining gene expression datasets and domain-specific knowledge improves patient diagnosis in all three settings.
comment: Accepted at the Extended Semantic Web Conference 2025
☆ Including local feature interactions in deep non-negative matrix factorization networks improves performance
The brain uses positive signals as a means of signaling. Forward interactions in the early visual cortex are also positive, realized by excitatory synapses. Only local interactions also include inhibition. Non-negative matrix factorization (NMF) captures the biological constraint of positive long-range interactions and can be implemented with stochastic spikes. While NMF can serve as an abstract formalization of early neural processing in the visual system, the performance of deep convolutional networks with NMF modules does not match that of CNNs of similar size. However, when the local NMF modules are each followed by a module that mixes the NMF's positive activities, the performances on the benchmark data exceed that of vanilla deep convolutional networks of similar size. This setting can be considered a biologically more plausible emulation of the processing in cortical (hyper-)columns with the potential to improve the performance of deep networks.
☆ FastFT: Accelerating Reinforced Feature Transformation via Advanced Exploration Strategies ICDE 2025
Feature Transformation is crucial for classic machine learning that aims to generate feature combinations to enhance the performance of downstream tasks from a data-centric perspective. Current methodologies, such as manual expert-driven processes, iterative-feedback techniques, and exploration-generative tactics, have shown promise in automating such data engineering workflow by minimizing human involvement. However, three challenges remain in those frameworks: (1) It predominantly depends on downstream task performance metrics, as assessment is time-consuming, especially for large datasets. (2) The diversity of feature combinations will hardly be guaranteed after random exploration ends. (3) Rare significant transformations lead to sparse valuable feedback that hinders the learning processes or leads to less effective results. In response to these challenges, we introduce FastFT, an innovative framework that leverages a trio of advanced strategies.We first decouple the feature transformation evaluation from the outcomes of the generated datasets via the performance predictor. To address the issue of reward sparsity, we developed a method to evaluate the novelty of generated transformation sequences. Incorporating this novelty into the reward function accelerates the model's exploration of effective transformations, thereby improving the search productivity. Additionally, we combine novelty and performance to create a prioritized memory buffer, ensuring that essential experiences are effectively revisited during exploration. Our extensive experimental evaluations validate the performance, efficiency, and traceability of our proposed framework, showcasing its superiority in handling complex feature transformation tasks.
comment: 14 pages, Accepted by ICDE 2025
☆ CNN+Transformer Based Anomaly Traffic Detection in UAV Networks for Emergency Rescue
The unmanned aerial vehicle (UAV) network has gained significant attentions in recent years due to its various applications. However, the traffic security becomes the key threatening public safety issue in an emergency rescue system due to the increasing vulnerability of UAVs to cyber attacks in environments with high heterogeneities. Hence, in this paper, we propose a novel anomaly traffic detection architecture for UAV networks based on the software-defined networking (SDN) framework and blockchain technology. Specifically, SDN separates the control and data plane to enhance the network manageability and security. Meanwhile, the blockchain provides decentralized identity authentication and data security records. Beisdes, a complete security architecture requires an effective mechanism to detect the time-series based abnormal traffic. Thus, an integrated algorithm combining convolutional neural networks (CNNs) and Transformer (CNN+Transformer) for anomaly traffic detection is developed, which is called CTranATD. Finally, the simulation results show that the proposed CTranATD algorithm is effective and outperforms the individual CNN, Transformer, and LSTM algorithms for detecting anomaly traffic.
☆ SURGEON: Memory-Adaptive Fully Test-Time Adaptation via Dynamic Activation Sparsity CVPR 2025
Despite the growing integration of deep models into mobile terminals, the accuracy of these models declines significantly due to various deployment interferences. Test-time adaptation (TTA) has emerged to improve the performance of deep models by adapting them to unlabeled target data online. Yet, the significant memory cost, particularly in resource-constrained terminals, impedes the effective deployment of most backward-propagation-based TTA methods. To tackle memory constraints, we introduce SURGEON, a method that substantially reduces memory cost while preserving comparable accuracy improvements during fully test-time adaptation (FTTA) without relying on specific network architectures or modifications to the original training procedure. Specifically, we propose a novel dynamic activation sparsity strategy that directly prunes activations at layer-specific dynamic ratios during adaptation, allowing for flexible control of learning ability and memory cost in a data-sensitive manner. Among this, two metrics, Gradient Importance and Layer Activation Memory, are considered to determine the layer-wise pruning ratios, reflecting accuracy contribution and memory efficiency, respectively. Experimentally, our method surpasses the baselines by not only reducing memory usage but also achieving superior accuracy, delivering SOTA performance across diverse datasets, architectures, and tasks.
comment: Accepted to CVPR 2025
☆ VideoGEM: Training-free Action Grounding in Videos
Vision-language foundation models have shown impressive capabilities across various zero-shot tasks, including training-free localization and grounding, primarily focusing on localizing objects in images. However, leveraging those capabilities to localize actions and events in videos is challenging, as actions have less physical outline and are usually described by higher-level concepts. In this work, we propose VideoGEM, the first training-free spatial action grounding method based on pretrained image- and video-language backbones. Namely, we adapt the self-self attention formulation of GEM to spatial activity grounding. We observe that high-level semantic concepts, such as actions, usually emerge in the higher layers of the image- and video-language models. We, therefore, propose a layer weighting in the self-attention path to prioritize higher layers. Additionally, we introduce a dynamic weighting method to automatically tune layer weights to capture each layer`s relevance to a specific prompt. Finally, we introduce a prompt decomposition, processing action, verb, and object prompts separately, resulting in a better spatial localization of actions. We evaluate the proposed approach on three image- and video-language backbones, CLIP, OpenCLIP, and ViCLIP, and on four video grounding datasets, V-HICO, DALY, YouCook-Interactions, and GroundingYouTube, showing that the proposed training-free approach is able to outperform current trained state-of-the-art approaches for spatial video grounding.
☆ Wasserstein Distributionally Robust Bayesian Optimization with Continuous Context
We address the challenge of sequential data-driven decision-making under context distributional uncertainty. This problem arises in numerous real-world scenarios where the learner optimizes black-box objective functions in the presence of uncontrollable contextual variables. We consider the setting where the context distribution is uncertain but known to lie within an ambiguity set defined as a ball in the Wasserstein distance. We propose a novel algorithm for Wasserstein Distributionally Robust Bayesian Optimization that can handle continuous context distributions while maintaining computational tractability. Our theoretical analysis combines recent results in self-normalized concentration in Hilbert spaces and finite-sample bounds for distributionally robust optimization to establish sublinear regret bounds that match state-of-the-art results. Through extensive comparisons with existing approaches on both synthetic and real-world problems, we demonstrate the simplicity, effectiveness, and practical applicability of our proposed method.
☆ Enabling Heterogeneous Adversarial Transferability via Feature Permutation Attacks PAKDD 2025
Adversarial attacks in black-box settings are highly practical, with transfer-based attacks being the most effective at generating adversarial examples (AEs) that transfer from surrogate models to unseen target models. However, their performance significantly degrades when transferring across heterogeneous architectures -- such as CNNs, MLPs, and Vision Transformers (ViTs) -- due to fundamental architectural differences. To address this, we propose Feature Permutation Attack (FPA), a zero-FLOP, parameter-free method that enhances adversarial transferability across diverse architectures. FPA introduces a novel feature permutation (FP) operation, which rearranges pixel values in selected feature maps to simulate long-range dependencies, effectively making CNNs behave more like ViTs and MLPs. This enhances feature diversity and improves transferability both across heterogeneous architectures and within homogeneous CNNs. Extensive evaluations on 14 state-of-the-art architectures show that FPA achieves maximum absolute gains in attack success rates of 7.68% on CNNs, 14.57% on ViTs, and 14.48% on MLPs, outperforming existing black-box attacks. Additionally, FPA is highly generalizable and can seamlessly integrate with other transfer-based attacks to further boost their performance. Our findings establish FPA as a robust, efficient, and computationally lightweight strategy for enhancing adversarial transferability across heterogeneous architectures.
comment: PAKDD 2025. Main Track
☆ CryoSAMU: Enhancing 3D Cryo-EM Density Maps of Protein Structures at Intermediate Resolution with Structure-Aware Multimodal U-Nets
Enhancing cryogenic electron microscopy (cryo-EM) 3D density maps at intermediate resolution (4-8 {\AA}) is crucial in protein structure determination. Recent advances in deep learning have led to the development of automated approaches for enhancing experimental cryo-EM density maps. Yet, these methods are not optimized for intermediate-resolution maps and rely on map density features alone. To address this, we propose CryoSAMU, a novel method designed to enhance 3D cryo-EM density maps of protein structures using structure-aware multimodal U-Nets and trained on curated intermediate-resolution density maps. We comprehensively evaluate CryoSAMU across various metrics and demonstrate its competitive performance compared to state-of-the-art methods. Notably, CryoSAMU achieves significantly faster processing speed, showing promise for future practical applications. Our code is available at https://github.com/chenwei-zhang/CryoSAMU.
comment: 18 pages, 6 main figures, 2 supplementary figures, 3 main tables, 4 supplementary tables
☆ Model-Based Offline Reinforcement Learning with Adversarial Data Augmentation
Model-based offline Reinforcement Learning (RL) constructs environment models from offline datasets to perform conservative policy optimization. Existing approaches focus on learning state transitions through ensemble models, rollouting conservative estimation to mitigate extrapolation errors. However, the static data makes it challenging to develop a robust policy, and offline agents cannot access the environment to gather new data. To address these challenges, we introduce Model-based Offline Reinforcement learning with AdversariaL data augmentation (MORAL). In MORAL, we replace the fixed horizon rollout by employing adversaria data augmentation to execute alternating sampling with ensemble models to enrich training data. Specifically, this adversarial process dynamically selects ensemble models against policy for biased sampling, mitigating the optimistic estimation of fixed models, thus robustly expanding the training data for policy optimization. Moreover, a differential factor is integrated into the adversarial process for regularization, ensuring error minimization in extrapolations. This data-augmented optimization adapts to diverse offline tasks without rollout horizon tuning, showing remarkable applicability. Extensive experiments on D4RL benchmark demonstrate that MORAL outperforms other model-based offline RL methods in terms of policy learning and sample efficiency.
☆ The cell as a token: high-dimensional geometry in language models and cell embeddings
Single-cell sequencing technology maps cells to a high-dimensional space encoding their internal activity. This process mirrors parallel developments in machine learning, where large language models ingest unstructured text by converting words into discrete tokens embedded within a high-dimensional vector space. This perspective explores how advances in understanding the structure of language embeddings can inform ongoing efforts to analyze and visualize single cell datasets. We discuss how the context of tokens influences the geometry of embedding space, and the role of low-dimensional manifolds in shaping this space's robustness and interpretability. We highlight new developments in language modeling, such as interpretability probes and in-context reasoning, that can inform future efforts to construct and consolidate cell atlases.
comment: 4 pages, 2 figures
☆ An $(ε,δ)$-accurate level set estimation with a stopping criterion
The level set estimation problem seeks to identify regions within a set of candidate points where an unknown and costly to evaluate function's value exceeds a specified threshold, providing an efficient alternative to exhaustive evaluations of function values. Traditional methods often use sequential optimization strategies to find $\epsilon$-accurate solutions, which permit a margin around the threshold contour but frequently lack effective stopping criteria, leading to excessive exploration and inefficiencies. This paper introduces an acquisition strategy for level set estimation that incorporates a stopping criterion, ensuring the algorithm halts when further exploration is unlikely to yield improvements, thereby reducing unnecessary function evaluations. We theoretically prove that our method satisfies $\epsilon$-accuracy with a confidence level of $1 - \delta$, addressing a key gap in existing approaches. Furthermore, we show that this also leads to guarantees on the lower bounds of performance metrics such as F-score. Numerical experiments demonstrate that the proposed acquisition function achieves comparable precision to existing methods while confirming that the stopping criterion effectively terminates the algorithm once adequate exploration is completed.
☆ Revisit Time Series Classification Benchmark: The Impact of Temporal Information for Classification PAKDD2025
Time series classification is usually regarded as a distinct task from tabular data classification due to the importance of temporal information. However, in this paper, by performing permutation tests that disrupt temporal information on the UCR time series classification archive, the most widely used benchmark for time series classification, we identify a significant proportion of datasets where temporal information has little to no impact on classification. Many of these datasets are tabular in nature or rely mainly on tabular features, leading to potentially biased evaluations of time series classifiers focused on temporal information. To address this, we propose UCR Augmented, a benchmark based on the UCR time series classification archive designed to evaluate classifiers' ability to extract and utilize temporal information. Testing classifiers from seven categories on this benchmark revealed notable shifts in performance rankings. Some previously overlooked approaches perform well, while others see their performance decline significantly when temporal information is crucial. UCR Augmented provides a more robust framework for assessing time series classifiers, ensuring fairer evaluations. Our code is available at https://github.com/YunruiZhang/Revisit-Time-Series-Classification-Benchmark.
comment: Accepted to PAKDD2025
☆ Incremental Object Keypoint Learning CVPR
Existing progress in object keypoint estimation primarily benefits from the conventional supervised learning paradigm based on numerous data labeled with pre-defined keypoints. However, these well-trained models can hardly detect the undefined new keypoints in test time, which largely hinders their feasibility for diverse downstream tasks. To handle this, various solutions are explored but still suffer from either limited generalizability or transferability. Therefore, in this paper, we explore a novel keypoint learning paradigm in that we only annotate new keypoints in the new data and incrementally train the model, without retaining any old data, called Incremental object Keypoint Learning (IKL). A two-stage learning scheme as a novel baseline tailored to IKL is developed. In the first Knowledge Association stage, given the data labeled with only new keypoints, an auxiliary KA-Net is trained to automatically associate the old keypoints to these new ones based on their spatial and intrinsic anatomical relations. In the second Mutual Promotion stage, based on a keypoint-oriented spatial distillation loss, we jointly leverage the auxiliary KA-Net and the old model for knowledge consolidation to mutually promote the estimation of all old and new keypoints. Owing to the investigation of the correlations between new and old keypoints, our proposed method can not just effectively mitigate the catastrophic forgetting of old keypoints, but may even further improve the estimation of the old ones and achieve a positive transfer beyond anti-forgetting. Such an observation has been solidly verified by extensive experiments on different keypoint datasets, where our method exhibits superiority in alleviating the forgetting issue and boosting performance while enjoying labeling efficiency even under the low-shot data regime.
comment: The IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) 2025
☆ TraNCE: Transformative Non-linear Concept Explainer for CNNs
Convolutional neural networks (CNNs) have succeeded remarkably in various computer vision tasks. However, they are not intrinsically explainable. While the feature-level understanding of CNNs reveals where the models looked, concept-based explainability methods provide insights into what the models saw. However, their assumption of linear reconstructability of image activations fails to capture the intricate relationships within these activations. Their Fidelity-only approach to evaluating global explanations also presents a new concern. For the first time, we address these limitations with the novel Transformative Nonlinear Concept Explainer (TraNCE) for CNNs. Unlike linear reconstruction assumptions made by existing methods, TraNCE captures the intricate relationships within the activations. This study presents three original contributions to the CNN explainability literature: (i) An automatic concept discovery mechanism based on variational autoencoders (VAEs). This transformative concept discovery process enhances the identification of meaningful concepts from image activations. (ii) A visualization module that leverages the Bessel function to create a smooth transition between prototypical image pixels, revealing not only what the CNN saw but also what the CNN avoided, thereby mitigating the challenges of concept duplication as documented in previous works. (iii) A new metric, the Faith score, integrates both Coherence and Fidelity for a comprehensive evaluation of explainer faithfulness and consistency.
☆ TeleLoRA: Teleporting Model-Specific Alignment Across LLMs
Mitigating Trojans in Large Language Models (LLMs) is one of many tasks where alignment data is LLM specific, as different LLMs have different Trojan triggers and trigger behaviors to be removed. In this paper, we introduce TeleLoRA (Teleporting Low-Rank Adaptation), a novel framework that synergizes model-specific alignment data across multiple LLMs to enable zero-shot Trojan mitigation on unseen LLMs without alignment data. TeleLoRA learns a unified generator of LoRA adapter weights by leveraging local activation information across multiple LLMs. This generator is designed to be permutation symmetric to generalize across models with different architectures and sizes. We optimize the model design for memory efficiency, making it feasible to learn with large-scale LLMs with minimal computational resources. Experiments on LLM Trojan mitigation benchmarks demonstrate that TeleLoRA effectively reduces attack success rates while preserving the benign performance of the models.
☆ Solving 2-D Helmholtz equation in the rectangular, circular, and elliptical domains using neural networks
Physics-informed neural networks offered an alternate way to solve several differential equations that govern complicated physics. However, their success in predicting the acoustic field is limited by the vanishing-gradient problem that occurs when solving the Helmholtz equation. In this paper, a formulation is presented that addresses this difficulty. The problem of solving the two-dimensional Helmholtz equation with the prescribed boundary conditions is posed as an unconstrained optimization problem using trial solution method. According to this method, a trial neural network that satisfies the given boundary conditions prior to the training process is constructed using the technique of transfinite interpolation and the theory of R-functions. This ansatz is initially applied to the rectangular domain and later extended to the circular and elliptical domains. The acoustic field predicted from the proposed formulation is compared with that obtained from the two-dimensional finite element methods. Good agreement is observed in all three domains considered. Minor limitations associated with the proposed formulation and their remedies are also discussed.
comment: 59 pages
☆ Learning Adaptive Dexterous Grasping from Single Demonstrations
How can robots learn dexterous grasping skills efficiently and apply them adaptively based on user instructions? This work tackles two key challenges: efficient skill acquisition from limited human demonstrations and context-driven skill selection. We introduce AdaDexGrasp, a framework that learns a library of grasping skills from a single human demonstration per skill and selects the most suitable one using a vision-language model (VLM). To improve sample efficiency, we propose a trajectory following reward that guides reinforcement learning (RL) toward states close to a human demonstration while allowing flexibility in exploration. To learn beyond the single demonstration, we employ curriculum learning, progressively increasing object pose variations to enhance robustness. At deployment, a VLM retrieves the appropriate skill based on user instructions, bridging low-level learned skills with high-level intent. We evaluate AdaDexGrasp in both simulation and real-world settings, showing that our approach significantly improves RL efficiency and enables learning human-like grasp strategies across varied object configurations. Finally, we demonstrate zero-shot transfer of our learned policies to a real-world PSYONIC Ability Hand, with a 90% success rate across objects, significantly outperforming the baseline.
☆ Generalized Phase Pressure Control Enhanced Reinforcement Learning for Traffic Signal Control
Appropriate traffic state representation is crucial for learning traffic signal control policies. However, most of the current traffic state representations are heuristically designed, with insufficient theoretical support. In this paper, we (1) develop a flexible, efficient, and theoretically grounded method, namely generalized phase pressure (G2P) control, which takes only simple lane features into consideration to decide which phase to be actuated; 2) extend the pressure control theory to a general form for multi-homogeneous-lane road networks based on queueing theory; (3) design a new traffic state representation based on the generalized phase state features from G2P control; and 4) develop a reinforcement learning (RL)-based algorithm template named G2P-XLight, and two RL algorithms, G2P-MPLight and G2P-CoLight, by combining the generalized phase state representation with MPLight and CoLight, two well-performed RL methods for learning traffic signal control policies. Extensive experiments conducted on multiple real-world datasets demonstrate that G2P control outperforms the state-of-the-art (SOTA) heuristic method in the transportation field and other recent human-designed heuristic methods; and that the newly proposed G2P-XLight significantly outperforms SOTA learning-based approaches. Our code is available online.
☆ Open Deep Search: Democratizing Search with Open-source Reasoning Agents
We introduce Open Deep Search (ODS) to close the increasing gap between the proprietary search AI solutions, such as Perplexity's Sonar Reasoning Pro and OpenAI's GPT-4o Search Preview, and their open-source counterparts. The main innovation introduced in ODS is to augment the reasoning capabilities of the latest open-source LLMs with reasoning agents that can judiciously use web search tools to answer queries. Concretely, ODS consists of two components that work with a base LLM chosen by the user: Open Search Tool and Open Reasoning Agent. Open Reasoning Agent interprets the given task and completes it by orchestrating a sequence of actions that includes calling tools, one of which is the Open Search Tool. Open Search Tool is a novel web search tool that outperforms proprietary counterparts. Together with powerful open-source reasoning LLMs, such as DeepSeek-R1, ODS nearly matches and sometimes surpasses the existing state-of-the-art baselines on two benchmarks: SimpleQA and FRAMES. For example, on the FRAMES evaluation benchmark, ODS improves the best existing baseline of the recently released GPT-4o Search Preview by 9.7% in accuracy. ODS is a general framework for seamlessly augmenting any LLMs -- for example, DeepSeek-R1 that achieves 82.4% on SimpleQA and 30.1% on FRAMES -- with search and reasoning capabilities to achieve state-of-the-art performance: 88.3% on SimpleQA and 75.3% on FRAMES.
comment: 27 pages, 8 figures, 4 tables
Assessing SAM for Tree Crown Instance Segmentation from Drone Imagery ICLR 2025
The potential of tree planting as a natural climate solution is often undermined by inadequate monitoring of tree planting projects. Current monitoring methods involve measuring trees by hand for each species, requiring extensive cost, time, and labour. Advances in drone remote sensing and computer vision offer great potential for mapping and characterizing trees from aerial imagery, and large pre-trained vision models, such as the Segment Anything Model (SAM), may be a particularly compelling choice given limited labeled data. In this work, we compare SAM methods for the task of automatic tree crown instance segmentation in high resolution drone imagery of young tree plantations. We explore the potential of SAM for this task, and find that methods using SAM out-of-the-box do not outperform a custom Mask R-CNN, even with well-designed prompts, but that there is potential for methods which tune SAM further. We also show that predictions can be improved by adding Digital Surface Model (DSM) information as an input.
comment: ICLR 2025 ML4RS workshop
☆ Maya: Optimizing Deep Learning Training Workloads using Emulated Virtual Accelerators
Training large foundation models costs hundreds of millions of dollars, making deployment optimization critical. Current approaches require machine learning engineers to manually craft training recipes through error-prone trial-and-error on expensive compute clusters. To enable efficient exploration of training configurations, researchers have developed performance modeling systems. However, these systems force users to translate their workloads into custom specification languages, introducing a fundamental semantic gap between the actual workload and its representation. This gap creates an inherent tradeoff: systems must either support a narrow set of workloads to maintain usability, require complex specifications that limit practical adoption, or compromise prediction accuracy with simplified models. We present Maya, a performance modeling system that eliminates these tradeoffs through transparent device emulation. By operating at the narrow interface between training frameworks and accelerator devices, Maya can capture complete workload behavior without requiring code modifications or translations. Maya intercepts device API calls from unmodified training code to directly observe low-level operations, enabling accurate performance prediction while maintaining both ease of use and generality. Our evaluation shows Maya achieves less than 5% prediction error across diverse models and optimization strategies, identifying configurations that reduce training costs by up to 56% compared to existing approaches.
☆ Network Inversion for Generating Confidently Classified Counterfeits
In machine learning, especially with vision classifiers, generating inputs that are confidently classified by the model is essential for understanding its decision boundaries and behavior. However, creating such samples that are confidently classified yet distinct from the training data distribution is a challenge. Traditional methods often modify existing inputs, but they don't always ensure confident classification. In this work, we extend network inversion techniques to generate Confidently Classified Counterfeits-synthetic samples that are confidently classified by the model despite being significantly different from the training data. We achieve this by modifying the generator's conditioning mechanism from soft vector conditioning to one-hot vector conditioning and applying Kullback-Leibler divergence (KLD) between the one-hot vectors and the classifier's output distribution. This encourages the generator to produce samples that are both plausible and confidently classified. Generating Confidently Classified Counterfeits is crucial for ensuring the safety and reliability of machine learning systems, particularly in safety-critical applications where models must exhibit confidence only on data within the training distribution. By generating such counterfeits, we challenge the assumption that high-confidence predictions are always indicative of in-distribution data, providing deeper insights into the model's limitations and decision-making process.
☆ Offline Reinforcement Learning with Discrete Diffusion Skills
Skills have been introduced to offline reinforcement learning (RL) as temporal abstractions to tackle complex, long-horizon tasks, promoting consistent behavior and enabling meaningful exploration. While skills in offline RL are predominantly modeled within a continuous latent space, the potential of discrete skill spaces remains largely underexplored. In this paper, we propose a compact discrete skill space for offline RL tasks supported by state-of-the-art transformer-based encoder and diffusion-based decoder. Coupled with a high-level policy trained via offline RL techniques, our method establishes a hierarchical RL framework where the trained diffusion decoder plays a pivotal role. Empirical evaluations show that the proposed algorithm, Discrete Diffusion Skill (DDS), is a powerful offline RL method. DDS performs competitively on Locomotion and Kitchen tasks and excels on long-horizon tasks, achieving at least a 12 percent improvement on AntMaze-v2 benchmarks compared to existing offline RL approaches. Furthermore, DDS offers improved interpretability, training stability, and online exploration compared to previous skill-based methods.
☆ AIGC-assisted Federated Learning for Edge Intelligence: Architecture Design, Research Challenges and Future Directions
Federated learning (FL) can fully leverage large-scale terminal data while ensuring privacy and security, and is considered as a distributed alternative for the centralized machine learning. However, the issue of data heterogeneity poses limitations on FL's performance. To address this challenge, artificial intelligence-generated content (AIGC) which is an innovative data synthesis technique emerges as one potential solution. In this article, we first provide an overview of the system architecture, performance metrics, and challenges associated with AIGC-assistant FL system design. We then propose the Generative federated learning (GenFL) architecture and present its workflow, including the design of aggregation and weight policy. Finally, using the CIFAR10 and CIFAR100 datasets, we employ diffusion models to generate dataset and improve FL performance. Experiments conducted under various non-independent and identically distributed (non-IID) data distributions demonstrate the effectiveness of GenFL on overcoming the bottlenecks in FL caused by data heterogeneity. Open research directions in the research of AIGC-assisted FL are also discussed.
☆ Emotion Detection in Twitter Messages Using Combination of Long Short-Term Memory and Convolutional Deep Neural Networks
One of the most significant issues as attended a lot in recent years is that of recognizing the sentiments and emotions in social media texts. The analysis of sentiments and emotions is intended to recognize the conceptual information such as the opinions, feelings, attitudes and emotions of people towards the products, services, organizations, people, topics, events and features in the written text. These indicate the greatness of the problem space. In the real world, businesses and organizations are always looking for tools to gather ideas, emotions, and directions of people about their products, services, or events related to their own. This article uses the Twitter social network, one of the most popular social networks with about 420 million active users, to extract data. Using this social network, users can share their information and opinions about personal issues, policies, products, events, etc. It can be used with appropriate classification of emotional states due to the availability of its data. In this study, supervised learning and deep neural network algorithms are used to classify the emotional states of Twitter users. The use of deep learning methods to increase the learning capacity of the model is an advantage due to the large amount of available data. Tweets collected on various topics are classified into four classes using a combination of two Bidirectional Long Short Term Memory network and a Convolutional network. The results obtained from this study with an average accuracy of 93%, show good results extracted from the proposed framework and improved accuracy compared to previous work.
☆ RxRx3-core: Benchmarking drug-target interactions in High-Content Microscopy ICLR 2025
High Content Screening (HCS) microscopy datasets have transformed the ability to profile cellular responses to genetic and chemical perturbations, enabling cell-based inference of drug-target interactions (DTI). However, the adoption of representation learning methods for HCS data has been hindered by the lack of accessible datasets and robust benchmarks. To address this gap, we present RxRx3-core, a curated and compressed subset of the RxRx3 dataset, and an associated DTI benchmarking task. At just 18GB, RxRx3-core significantly reduces the size barrier associated with large-scale HCS datasets while preserving critical data necessary for benchmarking representation learning models against a zero-shot DTI prediction task. RxRx3-core includes 222,601 microscopy images spanning 736 CRISPR knockouts and 1,674 compounds at 8 concentrations. RxRx3-core is available on HuggingFace and Polaris, along with pre-trained embeddings and benchmarking code, ensuring accessibility for the research community. By providing a compact dataset and robust benchmarks, we aim to accelerate innovation in representation learning methods for HCS data and support the discovery of novel biological insights.
comment: Published at LMRL Workshop at ICLR 2025
☆ Addressing Challenges in Time Series Forecasting: A Comprehensive Comparison of Machine Learning Techniques
The explosion of Time Series (TS) data, driven by advancements in technology, necessitates sophisticated analytical methods. Modern management systems increasingly rely on analyzing this data, highlighting the importance of effcient processing techniques. State-of-the-art Machine Learning (ML) approaches for TS analysis and forecasting are becoming prevalent. This paper briefly describes and compiles suitable algorithms for TS regression task. We compare these algorithms against each other and the classic ARIMA method using diverse datasets: complete data, data with outliers, and data with missing values. The focus is on forecasting accuracy, particularly for long-term predictions. This research aids in selecting the most appropriate algorithm based on forecasting needs and data characteristics.
☆ Physics-Informed Neural Networks with Unknown Partial Differential Equations: an Application in Multivariate Time Series
A significant advancement in Neural Network (NN) research is the integration of domain-specific knowledge through custom loss functions. This approach addresses a crucial challenge: how can models utilize physics or mathematical principles to enhance predictions when dealing with sparse, noisy, or incomplete data? Physics-Informed Neural Networks (PINNs) put this idea into practice by incorporating physical equations, such as Partial Differential Equations (PDEs), as soft constraints. This guidance helps the networks find solutions that align with established laws. Recently, researchers have expanded this framework to include Bayesian NNs (BNNs), which allow for uncertainty quantification while still adhering to physical principles. But what happens when the governing equations of a system are not known? In this work, we introduce methods to automatically extract PDEs from historical data. We then integrate these learned equations into three different modeling approaches: PINNs, Bayesian-PINNs (B-PINNs), and Bayesian Linear Regression (BLR). To assess these frameworks, we evaluate them on a real-world Multivariate Time Series (MTS) dataset. We compare their effectiveness in forecasting future states under different scenarios: with and without PDE constraints and accuracy considerations. This research aims to bridge the gap between data-driven discovery and physics-guided learning, providing valuable insights for practical applications.
☆ Look Before Leap: Look-Ahead Planning with Uncertainty in Reinforcement Learning
Model-based reinforcement learning (MBRL) has demonstrated superior sample efficiency compared to model-free reinforcement learning (MFRL). However, the presence of inaccurate models can introduce biases during policy learning, resulting in misleading trajectories. The challenge lies in obtaining accurate models due to limited diverse training data, particularly in regions with limited visits (uncertain regions). Existing approaches passively quantify uncertainty after sample generation, failing to actively collect uncertain samples that could enhance state coverage and improve model accuracy. Moreover, MBRL often faces difficulties in making accurate multi-step predictions, thereby impacting overall performance. To address these limitations, we propose a novel framework for uncertainty-aware policy optimization with model-based exploratory planning. In the model-based planning phase, we introduce an uncertainty-aware k-step lookahead planning approach to guide action selection at each step. This process involves a trade-off analysis between model uncertainty and value function approximation error, effectively enhancing policy performance. In the policy optimization phase, we leverage an uncertainty-driven exploratory policy to actively collect diverse training samples, resulting in improved model accuracy and overall performance of the RL agent. Our approach offers flexibility and applicability to tasks with varying state/action spaces and reward structures. We validate its effectiveness through experiments on challenging robotic manipulation tasks and Atari games, surpassing state-of-the-art methods with fewer interactions, thereby leading to significant performance improvements.
☆ Unlocking the Value of Decentralized Data: A Federated Dual Learning Approach for Model Aggregation
Artificial Intelligence (AI) technologies have revolutionized numerous fields, yet their applications often rely on costly and time-consuming data collection processes. Federated Learning (FL) offers a promising alternative by enabling AI models to be trained on decentralized data where data is scattered across clients (distributed nodes). However, existing FL approaches struggle to match the performance of centralized training due to challenges such as heterogeneous data distribution and communication delays, limiting their potential for breakthroughs. We observe that many real-world use cases involve hybrid data regimes, in which a server (center node) has access to some data while a large amount of data is distributed across associated clients. To improve the utilization of decentralized data under this regime, address data heterogeneity issue, and facilitate asynchronous communication between the server and clients, we propose a dual learning approach that leverages centralized data at the server to guide the merging of model updates from clients. Our method accommodates scenarios where server data is out-of-domain relative to decentralized client data, making it applicable to a wide range of use cases. We provide theoretical analysis demonstrating the faster convergence of our method compared to existing methods. Furthermore, experimental results across various scenarios show that our approach significantly outperforms existing technologies, highlighting its potential to unlock the value of large amounts of decentralized data.
☆ Innovative LSGTime Model for Crime Spatiotemporal Prediction Based on MindSpore Framework
With the acceleration of urbanization, the spatiotemporal characteristics of criminal activities have become increasingly complex. Accurate prediction of crime distribution is crucial for optimizing the allocation of police resources and preventing crime. This paper proposes LGSTime, a crime spatiotemporal prediction model that integrates Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), and the Multi-head Sparse Self-attention mechanism. LSTM and GRU capture long-term dependencies in crime time series, such as seasonality and periodicity, through their unique gating mechanisms. The Multi-head Sparse Self-attention mechanism, on the other hand, focuses on both temporal and spatial features of criminal events simultaneously through parallel processing and sparsification techniques, significantly improving computational efficiency and prediction accuracy. The integrated model leverages the strengths of each technique to better handle complex spatiotemporal data. Experimental findings demonstrate that the model attains optimal performance across four real - world crime datasets. In comparison to the CNN model, it exhibits performance enhancements of 2.8\%, 1.9\%, and 1.4\% in the Mean Squared Error (MSE), Mean Absolute Error (MAE), and Root Mean Squared Error (RMSE) metrics respectively. These results offer a valuable reference for tackling the challenges in crime prediction.
☆ On the Robustness of Kernel Ridge Regression Using the Cauchy Loss Function
Robust regression aims to develop methods for estimating an unknown regression function in the presence of outliers, heavy-tailed distributions, or contaminated data, which can severely impact performance. Most existing theoretical results in robust regression assume that the noise has a finite absolute mean, an assumption violated by certain distributions, such as Cauchy and some Pareto noise. In this paper, we introduce a generalized Cauchy noise framework that accommodates all noise distributions with finite moments of any order, even when the absolute mean is infinite. Within this framework, we study the \textit{kernel Cauchy ridge regressor} (\textit{KCRR}), which minimizes a regularized empirical Cauchy risk to achieve robustness. To derive the $L_2$-risk bound for KCRR, we establish a connection between the excess Cauchy risk and $L_2$-risk for sufficiently large scale parameters of the Cauchy loss, which reveals that these two risks are equivalent. Furthermore, under the assumption that the regression function satisfies H\"older smoothness, we derive excess Cauchy risk bounds for KCRR, showing improved performance as the scale parameter decreases. By considering the twofold effect of the scale parameter on the excess Cauchy risk and its equivalence with the $L_2$-risk, we establish the almost minimax-optimal convergence rate for KCRR in terms of $L_2$-risk, highlighting the robustness of the Cauchy loss in handling various types of noise. Finally, we validate the effectiveness of KCRR through experiments on both synthetic and real-world datasets under diverse noise corruption scenarios.
☆ Integrated utilization of equations and small dataset in the Koopman operator: applications to forward and inverse Problems
In recent years, there has been a growing interest in data-driven approaches in physics, such as extended dynamic mode decomposition (EDMD). The EDMD algorithm focuses on nonlinear time-evolution systems, and the constructed Koopman matrix yields the next-time prediction with only linear matrix-product operations. Note that data-driven approaches generally require a large dataset. However, assume that one has some prior knowledge, even if it may be ambiguous. Then, one could achieve sufficient learning from only a small dataset by taking advantage of the prior knowledge. This paper yields methods for incorporating ambiguous prior knowledge into the EDMD algorithm. The ambiguous prior knowledge in this paper corresponds to the underlying time-evolution equations with unknown parameters. First, we apply the proposed method to forward problems, i.e., prediction tasks. Second, we propose a scheme to apply the proposed method to inverse problems, i.e., parameter estimation tasks. We demonstrate the learning with only a small dataset using guiding examples, i.e., the Duffing and the van der Pol systems.
comment: 10 pages, 8 figures
World Model Agents with Change-Based Intrinsic Motivation
Sparse reward environments pose a significant challenge for reinforcement learning due to the scarcity of feedback. Intrinsic motivation and transfer learning have emerged as promising strategies to address this issue. Change Based Exploration Transfer (CBET), a technique that combines these two approaches for model-free algorithms, has shown potential in addressing sparse feedback but its effectiveness with modern algorithms remains understudied. This paper provides an adaptation of CBET for world model algorithms like DreamerV3 and compares the performance of DreamerV3 and IMPALA agents, both with and without CBET, in the sparse reward environments of Crafter and Minigrid. Our tabula rasa results highlight the possibility of CBET improving DreamerV3's returns in Crafter but the algorithm attains a suboptimal policy in Minigrid with CBET further reducing returns. In the same vein, our transfer learning experiments show that pre-training DreamerV3 with intrinsic rewards does not immediately lead to a policy that maximizes extrinsic rewards in Minigrid. Overall, our results suggest that CBET provides a positive impact on DreamerV3 in more complex environments like Crafter but may be detrimental in environments like Minigrid. In the latter case, the behaviours promoted by CBET in DreamerV3 may not align with the task objectives of the environment, leading to reduced returns and suboptimal policies.
comment: Submitted to Northern Lights Deep Learning Conference 2025
☆ Improving Speech Recognition Accuracy Using Custom Language Models with the Vosk Toolkit
Although speech recognition algorithms have developed quickly in recent years, achieving high transcription accuracy across diverse audio formats and acoustic environments remains a major challenge. This work explores how incorporating custom language models with the open-source Vosk Toolkit can improve speech-to-text accuracy in varied settings. Unlike many conventional systems limited to specific audio types, this approach supports multiple audio formats such as WAV, MP3, FLAC, and OGG by using Python modules for preprocessing and format conversion. A Python-based transcription pipeline was developed to process input audio, perform speech recognition using Vosk's KaldiRecognizer, and export the output to a DOCX file. Results showed that custom models reduced word error rates, especially in domain-specific scenarios involving technical terminology, varied accents, or background noise. This work presents a cost-effective, offline solution for high-accuracy transcription and opens up future opportunities for automation and real-time applications.
comment: 10 pages, 7 figures, includes workflow diagram, accuracy and WER comparisons, spectrograms, and model evaluation
☆ Data Mixture Optimization: A Multi-fidelity Multi-scale Bayesian Framework
Careful curation of data sources can significantly improve the performance of LLM pre-training, but predominant approaches rely heavily on intuition or costly trial-and-error, making them difficult to generalize across different data domains and downstream tasks. Although scaling laws can provide a principled and general approach for data curation, standard deterministic extrapolation from small-scale experiments to larger scales requires strong assumptions on the reliability of such extrapolation, whose brittleness has been highlighted in prior works. In this paper, we introduce a $\textit{probabilistic extrapolation framework}$ for data mixture optimization that avoids rigid assumptions and explicitly models the uncertainty in performance across decision variables. We formulate data curation as a sequential decision-making problem$\unicode{x2013}$multi-fidelity, multi-scale Bayesian optimization$\unicode{x2013}$where $\{$data mixtures, model scale, training steps$\}$ are adaptively selected to balance training cost and potential information gain. Our framework naturally gives rise to algorithm prototypes that leverage noisy information from inexpensive experiments to systematically inform costly training decisions. To accelerate methodological progress, we build a simulator based on 472 language model pre-training runs with varying data compositions from the SlimPajama dataset. We observe that even simple kernels and acquisition functions can enable principled decisions across training models from 20M to 1B parameters and achieve $\textbf{2.6x}$ and $\textbf{3.3x}$ speedups compared to multi-fidelity BO and random search baselines. Taken together, our framework underscores potential efficiency gains achievable by developing principled and transferable data mixture optimization methods.
☆ Offline Action-Free Learning of Ex-BMDPs by Comparing Diverse Datasets
While sequential decision-making environments often involve high-dimensional observations, not all features of these observations are relevant for control. In particular, the observation space may capture factors of the environment which are not controllable by the agent, but which add complexity to the observation space. The need to ignore these "noise" features in order to operate in a tractably-small state space poses a challenge for efficient policy learning. Due to the abundance of video data available in many such environments, task-independent representation learning from action-free offline data offers an attractive solution. However, recent work has highlighted theoretical limitations in action-free learning under the Exogenous Block MDP (Ex-BMDP) model, where temporally-correlated noise features are present in the observations. To address these limitations, we identify a realistic setting where representation learning in Ex-BMDPs becomes tractable: when action-free video data from multiple agents with differing policies are available. Concretely, this paper introduces CRAFT (Comparison-based Representations from Action-Free Trajectories), a sample-efficient algorithm leveraging differences in controllable feature dynamics across agents to learn representations. We provide theoretical guarantees for CRAFT's performance and demonstrate its feasibility on a toy example, offering a foundation for practical methods in similar settings.
☆ Improving User Behavior Prediction: Leveraging Annotator Metadata in Supervised Machine Learning Models SC
Supervised machine-learning models often underperform in predicting user behaviors from conversational text, hindered by poor crowdsourced label quality and low NLP task accuracy. We introduce the Metadata-Sensitive Weighted-Encoding Ensemble Model (MSWEEM), which integrates annotator meta-features like fatigue and speeding. First, our results show MSWEEM outperforms standard ensembles by 14\% on held-out data and 12\% on an alternative dataset. Second, we find that incorporating signals of annotator behavior, such as speed and fatigue, significantly boosts model performance. Third, we find that annotators with higher qualifications, such as Master's, deliver more consistent and faster annotations. Given the increasing uncertainty over annotation quality, our experiments show that understanding annotator patterns is crucial for enhancing model accuracy in user behavior prediction.
comment: Accepted at CSCW 2025
☆ Deep Learning for Forensic Identification of Source
We used contrastive neural networks to learn useful similarity scores between the 144 cartridge casings in the NBIDE dataset, under the common-but-unknown source paradigm. The common-but-unknown source problem is a problem archetype in forensics where the question is whether two objects share a common source (e.g. were two cartridge casings fired from the same firearm). Similarity scores are often used to interpret evidence under this paradigm. We directly compared our results to a state-of-the-art algorithm, Congruent Matching Cells (CMC). When trained on the E3 dataset of 2967 cartridge casings, contrastive learning achieved an ROC AUC of 0.892. The CMC algorithm achieved 0.867. We also conducted an ablation study where we varied the neural network architecture; specifically, the network's width or depth. The ablation study showed that contrastive network performance results are somewhat robust to the network architecture. This work was in part motivated by the use of similarity scores attained via contrastive learning for standard evidence interpretation methods such as score-based likelihood ratios.
☆ Reinforcement Learning for Efficient Toxicity Detection in Competitive Online Video Games
Online platforms take proactive measures to detect and address undesirable behavior, aiming to focus these resource-intensive efforts where such behavior is most prevalent. This article considers the problem of efficient sampling for toxicity detection in competitive online video games. To make optimal monitoring decisions, video game service operators need estimates of the likelihood of toxic behavior. If no model is available for these predictions, one must be estimated in real time. To close this gap, we propose a contextual bandit algorithm that makes monitoring decisions based on a small set of variables that, according to domain expertise, are associated with toxic behavior. This algorithm balances exploration and exploitation to optimize long-term outcomes and is deliberately designed for easy deployment in production. Using data from the popular first-person action game Call of Duty: Modern Warfare III, we show that our algorithm consistently outperforms baseline algorithms that rely solely on players' past behavior. This finding has substantive implications for the nature of toxicity. It also illustrates how domain expertise can be harnessed to help video game service operators identify and mitigate toxicity, ultimately fostering a safer and more enjoyable gaming experience.
☆ Multi-Modal Framing Analysis of News
Automated frame analysis of political communication is a popular task in computational social science that is used to study how authors select aspects of a topic to frame its reception. So far, such studies have been narrow, in that they use a fixed set of pre-defined frames and focus only on the text, ignoring the visual contexts in which those texts appear. Especially for framing in the news, this leaves out valuable information about editorial choices, which include not just the written article but also accompanying photographs. To overcome such limitations, we present a method for conducting multi-modal, multi-label framing analysis at scale using large (vision-)language models. Grounding our work in framing theory, we extract latent meaning embedded in images used to convey a certain point and contrast that to the text by comparing the respective frames used. We also identify highly partisan framing of topics with issue-specific frame analysis found in prior qualitative work. We demonstrate a method for doing scalable integrative framing analysis of both text and image in news, providing a more complete picture for understanding media bias.
☆ TS-Inverse: A Gradient Inversion Attack Tailored for Federated Time Series Forecasting Models
Federated learning (FL) for time series forecasting (TSF) enables clients with privacy-sensitive time series (TS) data to collaboratively learn accurate forecasting models, for example, in energy load prediction. Unfortunately, privacy risks in FL persist, as servers can potentially reconstruct clients' training data through gradient inversion attacks (GIA). Although GIA is demonstrated for image classification tasks, little is known about time series regression tasks. In this paper, we first conduct an extensive empirical study on inverting TS data across 4 TSF models and 4 datasets, identifying the unique challenges of reconstructing both observations and targets of TS data. We then propose TS-Inverse, a novel GIA that improves the inversion of TS data by (i) learning a gradient inversion model that outputs quantile predictions, (ii) a unique loss function that incorporates periodicity and trend regularization, and (iii) regularization according to the quantile predictions. Our evaluations demonstrate a remarkable performance of TS-Inverse, achieving at least a 2x-10x improvement in terms of the sMAPE metric over existing GIA methods on TS data. Code repository: https://github.com/Capsar/ts-inverse
☆ Global and Local Structure Learning for Sparse Tensor Completion
How can we accurately complete tensors by learning relationships of dimensions along each mode? Tensor completion, a widely studied problem, is to predict missing entries in incomplete tensors. Tensor decomposition methods, fundamental tensor analysis tools, have been actively developed to solve tensor completion tasks. However, standard tensor decomposition models have not been designed to learn relationships of dimensions along each mode, which limits to accurate tensor completion. Also, previously developed tensor decomposition models have required prior knowledge between relations within dimensions to model the relations, expensive to obtain. This paper proposes TGL (Tensor Decomposition Learning Global and Local Structures) to accurately predict missing entries in tensors. TGL reconstructs a tensor with factor matrices which learn local structures with GNN without prior knowledges. Extensive experiments are conducted to evaluate TGL with baselines and datasets.
☆ Prototype Guided Backdoor Defense
Deep learning models are susceptible to {\em backdoor attacks} involving malicious attackers perturbing a small subset of training data with a {\em trigger} to causes misclassifications. Various triggers have been used, including semantic triggers that are easily realizable without requiring the attacker to manipulate the image. The emergence of generative AI has eased the generation of varied poisoned samples. Robustness across types of triggers is crucial to effective defense. We propose Prototype Guided Backdoor Defense (PGBD), a robust post-hoc defense that scales across different trigger types, including previously unsolved semantic triggers. PGBD exploits displacements in the geometric spaces of activations to penalize movements toward the trigger. This is done using a novel sanitization loss of a post-hoc fine-tuning step. The geometric approach scales easily to all types of attacks. PGBD achieves better performance across all settings. We also present the first defense against a new semantic attack on celebrity face images. Project page: \hyperlink{https://venkatadithya9.github.io/pgbd.github.io/}{this https URL}.
☆ D4R -- Exploring and Querying Relational Graphs Using Natural Language and Large Language Models -- the Case of Historical Documents
D4R is a digital platform designed to assist non-technical users, particularly historians, in exploring textual documents through advanced graphical tools for text analysis and knowledge extraction. By leveraging a large language model, D4R translates natural language questions into Cypher queries, enabling the retrieval of data from a Neo4J database. A user-friendly graphical interface allows for intuitive interaction, enabling users to navigate and analyse complex relational data extracted from unstructured textual documents. Originally designed to bridge the gap between AI technologies and historical research, D4R's capabilities extend to various other domains. A demonstration video and a live software demo are available.
comment: 8 pages, 7 figures
☆ TransDiffSBDD: Causality-Aware Multi-Modal Structure-Based Drug Design
Structure-based drug design (SBDD) is a critical task in drug discovery, requiring the generation of molecular information across two distinct modalities: discrete molecular graphs and continuous 3D coordinates. However, existing SBDD methods often overlook two key challenges: (1) the multi-modal nature of this task and (2) the causal relationship between these modalities, limiting their plausibility and performance. To address both challenges, we propose TransDiffSBDD, an integrated framework combining autoregressive transformers and diffusion models for SBDD. Specifically, the autoregressive transformer models discrete molecular information, while the diffusion model samples continuous distributions, effectively resolving the first challenge. To address the second challenge, we design a hybrid-modal sequence for protein-ligand complexes that explicitly respects the causality between modalities. Experiments on the CrossDocked2020 benchmark demonstrate that TransDiffSBDD outperforms existing baselines.
☆ Assessing Generative Models for Structured Data
Synthetic tabular data generation has emerged as a promising method to address limited data availability and privacy concerns. With the sharp increase in the performance of large language models in recent years, researchers have been interested in applying these models to the generation of tabular data. However, little is known about the quality of the generated tabular data from large language models. The predominant method for assessing the quality of synthetic tabular data is the train-synthetic-test-real approach, where the artificial examples are compared to the original by how well machine learning models, trained separately on the real and synthetic sets, perform in some downstream tasks. This method does not directly measure how closely the distribution of generated data approximates that of the original. This paper introduces rigorous methods for directly assessing synthetic tabular data against real data by looking at inter-column dependencies within the data. We find that large language models (GPT-2), both when queried via few-shot prompting and when fine-tuned, and GAN (CTGAN) models do not produce data with dependencies that mirror the original real data. Results from this study can inform future practice in synthetic data generation to improve data quality.
☆ Quantum advantage for learning shallow neural networks with natural data distributions
The application of quantum computers to machine learning tasks is an exciting potential direction to explore in search of quantum advantage. In the absence of large quantum computers to empirically evaluate performance, theoretical frameworks such as the quantum probably approximately correct (PAC) and quantum statistical query (QSQ) models have been proposed to study quantum algorithms for learning classical functions. Despite numerous works investigating quantum advantage in these models, we nevertheless only understand it at two extremes: either exponential quantum advantages for uniform input distributions or no advantage for potentially adversarial distributions. In this work, we study the gap between these two regimes by designing an efficient quantum algorithm for learning periodic neurons in the QSQ model over a broad range of non-uniform distributions, which includes Gaussian, generalized Gaussian, and logistic distributions. To our knowledge, our work is also the first result in quantum learning theory for classical functions that explicitly considers real-valued functions. Recent advances in classical learning theory prove that learning periodic neurons is hard for any classical gradient-based algorithm, giving us an exponential quantum advantage over such algorithms, which are the standard workhorses of machine learning. Moreover, in some parameter regimes, the problem remains hard for classical statistical query algorithms and even general classical algorithms learning under small amounts of noise.
comment: 8 pages, 1 figure + 80-page appendix
☆ Unified Multimodal Discrete Diffusion
Multimodal generative models that can understand and generate across multiple modalities are dominated by autoregressive (AR) approaches, which process tokens sequentially from left to right, or top to bottom. These models jointly handle images, text, video, and audio for various tasks such as image captioning, question answering, and image generation. In this work, we explore discrete diffusion models as a unified generative formulation in the joint text and image domain, building upon their recent success in text generation. Discrete diffusion models offer several advantages over AR models, including improved control over quality versus diversity of generated samples, the ability to perform joint multimodal inpainting (across both text and image domains), and greater controllability in generation through guidance. Leveraging these benefits, we present the first Unified Multimodal Discrete Diffusion (UniDisc) model which is capable of jointly understanding and generating text and images for a variety of downstream tasks. We compare UniDisc to multimodal AR models, performing a scaling analysis and demonstrating that UniDisc outperforms them in terms of both performance and inference-time compute, enhanced controllability, editability, inpainting, and flexible trade-off between inference time and generation quality. Code and additional visualizations are available at https://unidisc.github.io.
comment: Project Website: https://unidisc.github.io
☆ Robust Deep Reinforcement Learning in Robotics via Adaptive Gradient-Masked Adversarial Attacks
Deep reinforcement learning (DRL) has emerged as a promising approach for robotic control, but its realworld deployment remains challenging due to its vulnerability to environmental perturbations. Existing white-box adversarial attack methods, adapted from supervised learning, fail to effectively target DRL agents as they overlook temporal dynamics and indiscriminately perturb all state dimensions, limiting their impact on long-term rewards. To address these challenges, we propose the Adaptive Gradient-Masked Reinforcement (AGMR) Attack, a white-box attack method that combines DRL with a gradient-based soft masking mechanism to dynamically identify critical state dimensions and optimize adversarial policies. AGMR selectively allocates perturbations to the most impactful state features and incorporates a dynamic adjustment mechanism to balance exploration and exploitation during training. Extensive experiments demonstrate that AGMR outperforms state-of-the-art adversarial attack methods in degrading the performance of the victim agent and enhances the victim agent's robustness through adversarial defense mechanisms.
comment: 9 pages, 6 figures
☆ TAR: Teacher-Aligned Representations via Contrastive Learning for Quadrupedal Locomotion IROS
Quadrupedal locomotion via Reinforcement Learning (RL) is commonly addressed using the teacher-student paradigm, where a privileged teacher guides a proprioceptive student policy. However, key challenges such as representation misalignment between the privileged teacher and the proprioceptive-only student, covariate shift due to behavioral cloning, and lack of deployable adaptation lead to poor generalization in real-world scenarios. We propose Teacher-Aligned Representations via Contrastive Learning (TAR), a framework that leverages privileged information with self-supervised contrastive learning to bridge this gap. By aligning representations to a privileged teacher in simulation via contrastive objectives, our student policy learns structured latent spaces and exhibits robust generalization to Out-of-Distribution (OOD) scenarios, surpassing the fully privileged "Teacher". Results showed accelerated training by 2x compared to state-of-the-art baselines to achieve peak performance. OOD scenarios showed better generalization by 40 percent on average compared to existing methods. Additionally, TAR transitions seamlessly into learning during deployment without requiring privileged states, setting a new benchmark in sample-efficient, adaptive locomotion and enabling continual fine-tuning in real-world scenarios. Open-source code and videos are available at https://ammousa.github.io/TARLoco/.
comment: This work has been submitted to the IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) 2025 for review
☆ Workshop Scientific HPC in the pre-Exascale era (part of ITADATA 2024) Proceedings
The proceedings of Workshop Scientific HPC in the pre-Exascale era (SHPC), held in Pisa, Italy, September 18, 2024, are part of 3rd Italian Conference on Big Data and Data Science (ITADATA2024) proceedings (arXiv: 2503.14937). The main objective of SHPC workshop was to discuss how the current most critical questions in HPC emerge in astrophysics, cosmology, and other scientific contexts and experiments. In particular, SHPC workshop focused on: $\bullet$ Scientific (mainly in astrophysical and medical fields) applications toward (pre-)Exascale computing $\bullet$ Performance portability $\bullet$ Green computing $\bullet$ Machine learning $\bullet$ Big Data management $\bullet$ Programming on heterogeneous architectures $\bullet$ Programming on accelerators $\bullet$ I/O techniques
♻ ☆ Control, Optimal Transport and Neural Differential Equations in Supervised Learning
From the perspective of control theory, neural differential equations (neural ODEs) have become an important tool for supervised learning. In the fundamental work of Ruiz-Balet and Zuazua (SIAM REVIEW 2023), the authors pose an open problem regarding the connection between control theory, optimal transport theory, and neural differential equations. More precisely, they inquire how one can quantify the closeness of the optimal flows in neural transport equations to the true dynamic optimal transport. In this work, we propose a construction of neural differential equations that converge to the true dynamic optimal transport in the limit, providing a significant step in solving the formerly mentioned open problem.
♻ ☆ A Pretraining-Finetuning Computational Framework for Material Homogenization
Homogenization is a fundamental tool for studying multiscale physical phenomena. Traditional numerical homogenization methods, heavily reliant on finite element analysis, demand significant computational resources, especially for complex geometries, materials, and high-resolution problems. To address these challenges, we propose PreFine-Homo, a novel numerical homogenization framework comprising two phases: pretraining and fine-tuning. In the pretraining phase, a Fourier Neural Operator (FNO) is trained on large datasets to learn the mapping from input geometries and material properties to displacement fields. In the fine-tuning phase, the pretrained predictions serve as initial solutions for iterative algorithms, drastically reducing the number of iterations needed for convergence. The pretraining phase of PreFine-Homo delivers homogenization results up to 1000 times faster than conventional methods, while the fine-tuning phase further enhances accuracy. Moreover, the fine-tuning phase grants PreFine-Homo unlimited generalization capabilities, enabling continuous learning and improvement as data availability increases. We validate PreFine-Homo by predicting the effective elastic tensor for 3D periodic materials, specifically Triply Periodic Minimal Surfaces (TPMS). The results demonstrate that PreFine-Homo achieves high precision, exceptional efficiency, robust learning capabilities, and strong extrapolation ability, establishing it as a powerful tool for multiscale homogenization tasks.
♻ ☆ Assessing Consistency and Reproducibility in the Outputs of Large Language Models: Evidence Across Diverse Finance and Accounting Tasks
This study provides the first comprehensive assessment of consistency and reproducibility in Large Language Model (LLM) outputs in finance and accounting research. We evaluate how consistently LLMs produce outputs given identical inputs through extensive experimentation with 50 independent runs across five common tasks: classification, sentiment analysis, summarization, text generation, and prediction. Using three OpenAI models (GPT-3.5-turbo, GPT-4o-mini, and GPT-4o), we generate over 3.4 million outputs from diverse financial source texts and data, covering MD&As, FOMC statements, finance news articles, earnings call transcripts, and financial statements. Our findings reveal substantial but task-dependent consistency, with binary classification and sentiment analysis achieving near-perfect reproducibility, while complex tasks show greater variability. More advanced models do not consistently demonstrate better consistency and reproducibility, with task-specific patterns emerging. LLMs significantly outperform expert human annotators in consistency and maintain high agreement even where human experts significantly disagree. We further find that simple aggregation strategies across 3-5 runs dramatically improve consistency. We also find that aggregation may come with an additional benefit of improved accuracy for sentiment analysis when using newer models. Simulation analysis reveals that despite measurable inconsistency in LLM outputs, downstream statistical inferences remain remarkably robust. These findings address concerns about what we term "G-hacking," the selective reporting of favorable outcomes from multiple Generative AI runs, by demonstrating that such risks are relatively low for finance and accounting tasks.
comment: 97 pages, 20 tables, 15 figures
♻ ☆ Graph-Instructed Neural Networks for Sparse Grid-Based Discontinuity Detectors
In this paper, we present a novel approach for detecting the discontinuity interfaces of a discontinuous function. This approach leverages Graph-Instructed Neural Networks (GINNs) and sparse grids to address discontinuity detection also in domains of dimension larger than 3. GINNs, trained to identify troubled points on sparse grids, exploit graph structures built on the grids to achieve efficient and accurate discontinuity detection performances. We also introduce a recursive algorithm for general sparse grid-based detectors, characterized by convergence properties and easy applicability. Numerical experiments on functions with dimensions n = 2 and n = 4 demonstrate the efficiency and robust generalization properties of GINNs in detecting discontinuity interfaces. Notably, the trained GINNs offer portability and versatility, allowing integration into various algorithms and sharing among users.
♻ ☆ Certified Robustness via Dynamic Margin Maximization and Improved Lipschitz Regularization NeurIPS 2023
To improve the robustness of deep classifiers against adversarial perturbations, many approaches have been proposed, such as designing new architectures with better robustness properties (e.g., Lipschitz-capped networks), or modifying the training process itself (e.g., min-max optimization, constrained learning, or regularization). These approaches, however, might not be effective at increasing the margin in the input (feature) space. As a result, there has been an increasing interest in developing training procedures that can directly manipulate the decision boundary in the input space. In this paper, we build upon recent developments in this category by developing a robust training algorithm whose objective is to increase the margin in the output (logit) space while regularizing the Lipschitz constant of the model along vulnerable directions. We show that these two objectives can directly promote larger margins in the input space. To this end, we develop a scalable method for calculating guaranteed differentiable upper bounds on the Lipschitz constant of neural networks accurately and efficiently. The relative accuracy of the bounds prevents excessive regularization and allows for more direct manipulation of the decision boundary. Furthermore, our Lipschitz bounding algorithm exploits the monotonicity and Lipschitz continuity of the activation layers, and the resulting bounds can be used to design new layers with controllable bounds on their Lipschitz constant. Experiments on the MNIST, CIFAR-10, and Tiny-ImageNet data sets verify that our proposed algorithm obtains competitively improved results compared to the state-of-the-art.
comment: 37th Conference on Neural Information Processing Systems (NeurIPS 2023)
♻ ☆ DexHandDiff: Interaction-aware Diffusion Planning for Adaptive Dexterous Manipulation CVPR 2025
Dexterous manipulation with contact-rich interactions is crucial for advanced robotics. While recent diffusion-based planning approaches show promise for simple manipulation tasks, they often produce unrealistic ghost states (e.g., the object automatically moves without hand contact) or lack adaptability when handling complex sequential interactions. In this work, we introduce DexHandDiff, an interaction-aware diffusion planning framework for adaptive dexterous manipulation. DexHandDiff models joint state-action dynamics through a dual-phase diffusion process which consists of pre-interaction contact alignment and post-contact goal-directed control, enabling goal-adaptive generalizable dexterous manipulation. Additionally, we incorporate dynamics model-based dual guidance and leverage large language models for automated guidance function generation, enhancing generalizability for physical interactions and facilitating diverse goal adaptation through language cues. Experiments on physical interaction tasks such as door opening, pen and block re-orientation, object relocation, and hammer striking demonstrate DexHandDiff's effectiveness on goals outside training distributions, achieving over twice the average success rate (59.2% vs. 29.5%) compared to existing methods. Our framework achieves an average of 70.7% success rate on goal adaptive dexterous tasks, highlighting its robustness and flexibility in contact-rich manipulation.
comment: Accepted by CVPR 2025. Camera ready version. Previous DexDiffuser. Project page: https://dexdiffuser.github.io/
♻ ☆ Text-to-Model: Text-Conditioned Neural Network Diffusion for Train-Once-for-All Personalization
Generative artificial intelligence (GenAI) has made significant progress in understanding world knowledge and generating content from human languages across various modalities, like text-to-text large language models, text-to-image stable diffusion, and text-to-video Sora. While in this paper, we investigate the capability of GenAI for text-to-model generation, to see whether GenAI can comprehend hyper-level knowledge embedded within AI itself parameters. Specifically, we study a practical scenario termed train-once-for-all personalization, aiming to generate personalized models for diverse end-users and tasks using text prompts. Inspired by the recent emergence of neural network diffusion, we present Tina, a text-conditioned neural network diffusion for train-once-for-all personalization. Tina leverages a diffusion transformer model conditioned on task descriptions embedded using a CLIP model. Despite the astronomical number of potential personalized tasks (e.g., $1.73\times10^{13}$), by our design, Tina demonstrates remarkable in-distribution and out-of-distribution generalization even trained on small datasets ($\sim 1000$). We further verify whether and how \Tina understands world knowledge by analyzing its capabilities under zero-shot/few-shot image prompts, different numbers of personalized classes, prompts of natural language descriptions, and predicting unseen entities.
comment: Preprint
♻ ☆ Sinkhorn Distributionally Robust Optimization
We study distributionally robust optimization with Sinkhorn distance -- a variant of Wasserstein distance based on entropic regularization. We derive a convex programming dual reformulation for general nominal distributions, transport costs, and loss functions. To solve the dual reformulation, we develop a stochastic mirror descent algorithm with biased subgradient estimators and derive its computational complexity guarantees. Finally, we provide numerical examples using synthetic and real data to demonstrate its superior performance.
comment: 55 pages, 15 figures
♻ ☆ Harmony: A Joint Self-Supervised and Weakly-Supervised Framework for Learning General Purpose Visual Representations
Vision-language contrastive learning frameworks like CLIP enable learning representations from natural language supervision, and provide strong zero-shot classification capabilities. However, due to the nature of the supervisory signal in these paradigms, they lack the ability to learn localized features, leading to degraded performance on dense prediction tasks like segmentation and detection. On the other hand, self-supervised learning methods have shown the ability to learn granular representations, complementing the high-level features in vision-language training. In this work, we present Harmony, a framework that combines vision-language training with discriminative and generative self-supervision to learn visual features that can be generalized across different vision downstream tasks. Our framework is specifically designed to work on web-scraped data by not relying on negative examples and addressing the one-to-one correspondence issue using soft CLIP targets generated by an EMA model. We comprehensively evaluate Harmony across various vision downstream tasks and find that it significantly outperforms the baseline CLIP and the previously leading joint self and weakly-supervised methods, MaskCLIP and SLIP. Specifically, when comparing against these methods, Harmony shows superior performance in fine-tuning and zero-shot classification on ImageNet-1k, semantic segmentation on ADE20K, and both object detection and instance segmentation on MS-COCO, when pre-training a ViT-B on CC3M. We also show that Harmony outperforms other self-supervised learning methods like iBOT and MAE across all tasks evaluated. Our code is publicly at https://github.com/MohammedSB/Harmony}{https://github.com/MohammedSB/Harmony available.
comment: 22 pages, 4 figures
♻ ☆ A General Framework for Interpretable Neural Learning based on Local Information-Theoretic Goal Functions
Despite the impressive performance of biological and artificial networks, an intuitive understanding of how their local learning dynamics contribute to network-level task solutions remains a challenge to this date. Efforts to bring learning to a more local scale indeed lead to valuable insights, however, a general constructive approach to describe local learning goals that is both interpretable and adaptable across diverse tasks is still missing. We have previously formulated a local information processing goal that is highly adaptable and interpretable for a model neuron with compartmental structure. Building on recent advances in Partial Information Decomposition (PID), we here derive a corresponding parametric local learning rule, which allows us to introduce 'infomorphic' neural networks. We demonstrate the versatility of these networks to perform tasks from supervised, unsupervised and memory learning. By leveraging the interpretable nature of the PID framework, infomorphic networks represent a valuable tool to advance our understanding of the intricate structure of local learning.
comment: 30 pages, 14 figures
♻ ☆ COSMOS: Cross-Modality Self-Distillation for Vision Language Pre-training CVPR 2025
Vision-Language Models (VLMs) trained with contrastive loss have achieved significant advancements in various vision and language tasks. However, the global nature of the contrastive loss makes VLMs focus predominantly on foreground objects, neglecting other crucial information in the image, which limits their effectiveness in downstream tasks. To address these challenges, we propose COSMOS: CrOSs-MOdality Self-distillation for vision-language pre-training that integrates a novel text-cropping strategy and cross-attention module into a self-supervised learning framework. We create global and local views of images and texts (i.e., multi-modal augmentations), which are essential for self-distillation in VLMs. We further introduce a cross-attention module, enabling COSMOS to learn comprehensive cross-modal representations optimized via a cross-modality self-distillation loss. COSMOS consistently outperforms previous strong baselines on various zero-shot downstream tasks, including retrieval, classification, and semantic segmentation. Additionally, it surpasses CLIP-based models trained on larger datasets in visual perception and contextual understanding tasks. Code is available at https://github.com/ExplainableML/cosmos.
comment: CVPR 2025
♻ ☆ The mathematics of adversarial attacks in AI -- Why deep learning is unstable despite the existence of stable neural networks
The unprecedented success of deep learning (DL) makes it unchallenged when it comes to classification problems. However, it is well established that the current DL methodology produces universally unstable neural networks (NNs). The instability problem has caused an enormous research effort -- with a vast literature on so-called adversarial attacks -- yet there has been no solution to the problem. Our paper addresses why there has been no solution to the problem, as we prove the following mathematical paradox: any training procedure based on training neural networks for classification problems with a fixed architecture will yield neural networks that are either inaccurate or unstable (if accurate) -- despite the provable existence of both accurate and stable neural networks for the same classification problems. The key is that the stable and accurate neural networks must have variable dimensions depending on the input, in particular, variable dimensions is a necessary condition for stability. Our result points towards the paradox that accurate and stable neural networks exist, however, modern algorithms do not compute them. This yields the question: if the existence of neural networks with desirable properties can be proven, can one also find algorithms that compute them? There are cases in mathematics where provable existence implies computability, but will this be the case for neural networks? The contrary is true, as we demonstrate how neural networks can provably exist as approximate minimisers to standard optimisation problems with standard cost functions, however, no randomised algorithm can compute them with probability better than 1/2.
comment: 31 pages, 1 figure. Revised to make minor changes to notation and references
♻ ☆ DeSplat: Decomposed Gaussian Splatting for Distractor-Free Rendering
Gaussian splatting enables fast novel view synthesis in static 3D environments. However, reconstructing real-world environments remains challenging as distractors or occluders break the multi-view consistency assumption required for accurate 3D reconstruction. Most existing methods rely on external semantic information from pre-trained models, introducing additional computational overhead as pre-processing steps or during optimization. In this work, we propose a novel method, DeSplat, that directly separates distractors and static scene elements purely based on volume rendering of Gaussian primitives. We initialize Gaussians within each camera view for reconstructing the view-specific distractors to separately model the static 3D scene and distractors in the alpha compositing stages. DeSplat yields an explicit scene separation of static elements and distractors, achieving comparable results to prior distractor-free approaches without sacrificing rendering speed. We demonstrate DeSplat's effectiveness on three benchmark data sets for distractor-free novel view synthesis. See the project website at https://aaltoml.github.io/desplat/.
♻ ☆ Fully Distributed Fog Load Balancing with Multi-Agent Reinforcement Learning
Real-time Internet of Things (IoT) applications require real-time support to handle the ever-growing demand for computing resources to process IoT workloads. Fog Computing provides high availability of such resources in a distributed manner. However, these resources must be efficiently managed to distribute unpredictable traffic demands among heterogeneous Fog resources. This paper proposes a fully distributed load-balancing solution with Multi-Agent Reinforcement Learning (MARL) that intelligently distributes IoT workloads to optimize the waiting time while providing fair resource utilization in the Fog network. These agents use transfer learning for life-long self-adaptation to dynamic changes in the environment. By leveraging distributed decision-making, MARL agents effectively minimize the waiting time compared to a single centralized agent solution and other baselines, enhancing end-to-end execution delay. Besides performance gain, a fully distributed solution allows for a global-scale implementation where agents can work independently in small collaboration regions, leveraging nearby local resources. Furthermore, we analyze the impact of a realistic frequency to observe the state of the environment, unlike the unrealistic common assumption in the literature of having observations readily available in real-time for every required action. The findings highlight the trade-off between realism and performance using an interval-based Gossip-based multi-casting protocol against assuming real-time observation availability for every generated workload.
comment: Submitted to IEEE TNSM with 14 pages, 11 figures, and 3 tables
♻ ☆ Contractive Dynamical Imitation Policies for Efficient Out-of-Sample Recovery
Imitation learning is a data-driven approach to learning policies from expert behavior, but it is prone to unreliable outcomes in out-of-sample (OOS) regions. While previous research relying on stable dynamical systems guarantees convergence to a desired state, it often overlooks transient behavior. We propose a framework for learning policies modeled by contractive dynamical systems, ensuring that all policy rollouts converge regardless of perturbations, and in turn, enable efficient OOS recovery. By leveraging recurrent equilibrium networks and coupling layers, the policy structure guarantees contractivity for any parameter choice, which facilitates unconstrained optimization. We also provide theoretical upper bounds for worst-case and expected loss to rigorously establish the reliability of our method in deployment. Empirically, we demonstrate substantial OOS performance improvements for simulated robotic manipulation and navigation tasks.
comment: International Conference on Learning Representations
♻ ☆ Personalized Federated Learning of Probabilistic Models: A PAC-Bayesian Approach
Federated Learning (FL) aims to infer a shared model from private and decentralized data stored by multiple clients. Personalized FL (PFL) enhances the model's fit for each client by adapting the global model to the clients. A significant level of personalization is required for highly heterogeneous clients but can be challenging to achieve, especially when clients' datasets are small. To address this issue, we introduce the PAC-PFL framework for PFL of probabilistic models. PAC-PFL infers a shared hyper-posterior and treats each client's posterior inference as the personalization step. Unlike previous PFL algorithms, PAC-PFL does not regularize all personalized models towards a single shared model, thereby greatly enhancing its personalization flexibility. By establishing and minimizing a PAC-Bayesian generalization bound on the average true loss of clients, PAC-PFL effectively mitigates overfitting even in data-poor scenarios. Additionally, PAC-PFL provides generalization bounds for new clients joining later. PAC-PFL achieves accurate and well-calibrated predictions, as supported by our experiments.
♻ ☆ Perception of Visual Content: Differences Between Humans and Foundation Models
Human-annotated content is often used to train machine learning (ML) models. However, recently, language and multi-modal foundational models have been used to replace and scale-up human annotator's efforts. This study compares human-generated and ML-generated annotations of images representing diverse socio-economic contexts. We aim to understand differences in perception and identify potential biases in content interpretation. Our dataset comprises images of people from various geographical regions and income levels, covering various daily activities and home environments. We compare human and ML-generated annotations semantically and evaluate their impact on predictive models. Our results show highest similarity between ML captions and human labels from a low-level perspective, i.e., types of words that appear and sentence structures, but all three annotations are alike in how similar or dissimilar they perceive images across different regions. Additionally, ML Captions resulted in best overall region classification performance, while ML Objects and ML Captions performed best overall for income regression. The varying performance of annotation sets highlights the notion that all annotations are important, and that human-generated annotations are yet to be replaceable.
comment: 12 pages, 5 figures, 5 tables; updated version for a Revise-and-Resubmit at ICWSM 2025. This version includes a larger and more diverse dataset, leading to updated results
♻ ☆ Flow Matching Beyond Kinematics: Generating Jets with Particle-ID and Trajectory Displacement Information
We introduce the first generative model trained on the JetClass dataset. Our model generates jets at the constituent level, and it is a permutation-equivariant continuous normalizing flow (CNF) trained with the flow matching technique. It is conditioned on the jet type, so that a single model can be used to generate the ten different jet types of JetClass. For the first time, we also introduce a generative model that goes beyond the kinematic features of jet constituents. The JetClass dataset includes more features, such as particle-ID and track impact parameter, and we demonstrate that our CNF can accurately model all of these additional features as well. Our generative model for JetClass expands on the versatility of existing jet generation techniques, enhancing their potential utility in high-energy physics research, and offering a more comprehensive understanding of the generated jets.
♻ ☆ Long-Sequence Recommendation Models Need Decoupled Embeddings ICLR 2025
Lifelong user behavior sequences are crucial for capturing user interests and predicting user responses in modern recommendation systems. A two-stage paradigm is typically adopted to handle these long sequences: a subset of relevant behaviors is first searched from the original long sequences via an attention mechanism in the first stage and then aggregated with the target item to construct a discriminative representation for prediction in the second stage. In this work, we identify and characterize, for the first time, a neglected deficiency in existing long-sequence recommendation models: a single set of embeddings struggles with learning both attention and representation, leading to interference between these two processes. Initial attempts to address this issue with some common methods (e.g., linear projections -- a technique borrowed from language processing) proved ineffective, shedding light on the unique challenges of recommendation models. To overcome this, we propose the Decoupled Attention and Representation Embeddings (DARE) model, where two distinct embedding tables are initialized and learned separately to fully decouple attention and representation. Extensive experiments and analysis demonstrate that DARE provides more accurate searches of correlated behaviors and outperforms baselines with AUC gains up to 0.9% on public datasets and notable improvements on Tencent's advertising platform. Furthermore, decoupling embedding spaces allows us to reduce the attention embedding dimension and accelerate the search procedure by 50% without significant performance impact, enabling more efficient, high-performance online serving. Code in PyTorch for experiments, including model analysis, is available at https://github.com/thuml/DARE.
comment: ICLR 2025. First three authors contributed equally. Code is available at https://github.com/thuml/DARE
♻ ☆ Development and Validation of a Deep-Learning Model for Differential Treatment Benefit Prediction for Adults with Major Depressive Disorder Deployed in the Artificial Intelligence in Depression Medication Enhancement (AIDME) Study
INTRODUCTION: The pharmacological treatment of Major Depressive Disorder (MDD) relies on a trial-and-error approach. We introduce an artificial intelligence (AI) model aiming to personalize treatment and improve outcomes, which was deployed in the Artificial Intelligence in Depression Medication Enhancement (AIDME) Study. OBJECTIVES: 1) Develop a model capable of predicting probabilities of remission across multiple pharmacological treatments for adults with at least moderate major depression. 2) Validate model predictions and examine them for amplification of harmful biases. METHODS: Data from previous clinical trials of antidepressant medications were standardized into a common framework and included 9,042 adults with moderate to severe major depression. Feature selection retained 25 clinical and demographic variables. Using Bayesian optimization, a deep learning model was trained on the training set, refined using the validation set, and tested once on the held-out test set. RESULTS: In the evaluation on the held-out test set, the model demonstrated achieved an AUC of 0.65. The model outperformed a null model on the test set (p = 0.01). The model demonstrated clinical utility, achieving an absolute improvement in population remission rate in hypothetical and actual improvement testing. While the model did identify one drug (escitalopram) as generally outperforming the other drugs (consistent with the input data), there was otherwise significant variation in drug rankings. On bias testing, the model did not amplify potentially harmful biases. CONCLUSIONS: We demonstrate the first model capable of predicting outcomes for 10 different treatment options for patients with MDD, intended to be used at or near the start of treatment to personalize treatment. The model was put into clinical practice during the AIDME randomized controlled trial whose results are reported separately.
♻ ☆ Bayesian Modeling of Zero-Shot Classifications for Urban Flood Detection
Street scene datasets, collected from Street View or dashboard cameras, offer a promising means of detecting urban objects and incidents like street flooding. However, a major challenge in using these datasets is their lack of reliable labels: there are myriad types of incidents, many types occur rarely, and ground-truth measures of where incidents occur are lacking. Here, we propose BayFlood, a two-stage approach which circumvents this difficulty. First, we perform zero-shot classification of where incidents occur using a pretrained vision-language model (VLM). Second, we fit a spatial Bayesian model on the VLM classifications. The zero-shot approach avoids the need to annotate large training sets, and the Bayesian model provides frequent desiderata in urban settings - principled measures of uncertainty, smoothing across locations, and incorporation of external data like stormwater accumulation zones. We comprehensively validate this two-stage approach, showing that VLMs provide strong zero-shot signal for floods across multiple cities and time periods, the Bayesian model improves out-of-sample prediction relative to baseline methods, and our inferred flood risk correlates with known external predictors of risk. Having validated our approach, we show it can be used to improve urban flood detection: our analysis reveals 113,738 people who are at high risk of flooding overlooked by current methods, identifies demographic biases in existing methods, and suggests locations for new flood sensors. More broadly, our results showcase how Bayesian modeling of zero-shot LM annotations represents a promising paradigm because it avoids the need to collect large labeled datasets and leverages the power of foundation models while providing the expressiveness and uncertainty quantification of Bayesian models.
comment: In review
♻ ☆ Fantastic Copyrighted Beasts and How (Not) to Generate Them
Recent studies show that image and video generation models can be prompted to reproduce copyrighted content from their training data, raising serious legal concerns about copyright infringement. Copyrighted characters (e.g., Mario, Batman) present a significant challenge: at least one lawsuit has already awarded damages based on the generation of such characters. Consequently, commercial services like DALL-E have started deploying interventions. However, little research has systematically examined these problems: (1) Can users easily prompt models to generate copyrighted characters, even if it is unintentional?; (2) How effective are the existing mitigation strategies? To address these questions, we introduce a novel evaluation framework with metrics that assess both the generated image's similarity to copyrighted characters and its consistency with user intent, grounded in a set of popular copyrighted characters from diverse studios and regions. We show that state-of-the-art image and video generation models can still generate characters even if characters' names are not explicitly mentioned, sometimes with only two generic keywords (e.g., prompting with "videogame, plumber" consistently generates Nintendo's Mario character). We also introduce semi-automatic techniques to identify such keywords or descriptions that trigger character generation. Using this framework, we evaluate mitigation strategies, including prompt rewriting and new approaches we propose. Our findings reveal that common methods, such as DALL-E's prompt rewriting, are insufficient alone and require supplementary strategies like negative prompting. Our work provides empirical grounding for discussions on copyright mitigation strategies and offers actionable insights for model deployers implementing these safeguards.
♻ ☆ Towards Real-World Test-Time Adaptation: Tri-Net Self-Training with Balanced Normalization AAAI 2024
Test-Time Adaptation aims to adapt source domain model to testing data at inference stage with success demonstrated in adapting to unseen corruptions. However, these attempts may fail under more challenging real-world scenarios. Existing works mainly consider real-world test-time adaptation under non-i.i.d. data stream and continual domain shift. In this work, we first complement the existing real-world TTA protocol with a globally class imbalanced testing set. We demonstrate that combining all settings together poses new challenges to existing methods. We argue the failure of state-of-the-art methods is first caused by indiscriminately adapting normalization layers to imbalanced testing data. To remedy this shortcoming, we propose a balanced batchnorm layer to swap out the regular batchnorm at inference stage. The new batchnorm layer is capable of adapting without biasing towards majority classes. We are further inspired by the success of self-training (ST) in learning from unlabeled data and adapt ST for test-time adaptation. However, ST alone is prone to over adaption which is responsible for the poor performance under continual domain shift. Hence, we propose to improve self-training under continual domain shift by regularizing model updates with an anchored loss. The final TTA model, termed as TRIBE, is built upon a tri-net architecture with balanced batchnorm layers. We evaluate TRIBE on four datasets representing real-world TTA settings. TRIBE consistently achieves the state-of-the-art performance across multiple evaluation protocols. The code is available at https://github.com/Gorilla-Lab-SCUT/TRIBE.
comment: Accepted by AAAI 2024. 19 pages, 7 figures and 22 tables
♻ ☆ Inference-Time Scaling for Flow Models via Stochastic Generation and Rollover Budget Forcing
We propose an inference-time scaling approach for pretrained flow models. Recently, inference-time scaling has gained significant attention in LLMs and diffusion models, improving sample quality or better aligning outputs with user preferences by leveraging additional computation. For diffusion models, particle sampling has allowed more efficient scaling due to the stochasticity at intermediate denoising steps. On the contrary, while flow models have gained popularity as an alternative to diffusion models--offering faster generation and high-quality outputs in state-of-the-art image and video generative models--efficient inference-time scaling methods used for diffusion models cannot be directly applied due to their deterministic generative process. To enable efficient inference-time scaling for flow models, we propose three key ideas: 1) SDE-based generation, enabling particle sampling in flow models, 2) Interpolant conversion, broadening the search space and enhancing sample diversity, and 3) Rollover Budget Forcing (RBF), an adaptive allocation of computational resources across timesteps to maximize budget utilization. Our experiments show that SDE-based generation, particularly variance-preserving (VP) interpolant-based generation, improves the performance of particle sampling methods for inference-time scaling in flow models. Additionally, we demonstrate that RBF with VP-SDE achieves the best performance, outperforming all previous inference-time scaling approaches.
comment: Project page: https://flow-inference-time-scaling.github.io/
♻ ☆ Valid Conformal Prediction for Dynamic GNNs
Dynamic graphs provide a flexible data abstraction for modelling many sorts of real-world systems, such as transport, trade, and social networks. Graph neural networks (GNNs) are powerful tools allowing for different kinds of prediction and inference on these systems, but getting a handle on uncertainty, especially in dynamic settings, is a challenging problem. In this work we propose to use a dynamic graph representation known in the tensor literature as the unfolding, to achieve valid prediction sets via conformal prediction. This representation, a simple graph, can be input to any standard GNN and does not require any modification to existing GNN architectures or conformal prediction routines. One of our key contributions is a careful mathematical consideration of the different inference scenarios which can arise in a dynamic graph modelling context. For a range of practically relevant cases, we obtain valid prediction sets with almost no assumptions, even dispensing with exchangeability. In a more challenging scenario, which we call the semi-inductive regime, we achieve valid prediction under stronger assumptions, akin to stationarity. We provide real data examples demonstrating validity, showing improved accuracy over baselines, and sign-posting different failure modes which can occur when those assumptions are violated.
comment: 25 pages, 6 figures
♻ ☆ Aligning Visual Contrastive learning models via Preference Optimization
Contrastive learning models have demonstrated impressive abilities to capture semantic similarities by aligning representations in the embedding space. However, their performance can be limited by the quality of the training data and its inherent biases. While Preference Optimization (PO) methods such as Reinforcement Learning from Human Feedback (RLHF) and Direct Preference Optimization (DPO) have been applied to align generative models with human preferences, their use in contrastive learning has yet to be explored. This paper introduces a novel method for training contrastive learning models using different PO methods to break down complex concepts. Our method systematically aligns model behavior with desired preferences, enhancing performance on the targeted task. In particular, we focus on enhancing model robustness against typographic attacks and inductive biases, commonly seen in contrastive vision-language models like CLIP. Our experiments demonstrate that models trained using PO outperform standard contrastive learning techniques while retaining their ability to handle adversarial challenges and maintain accuracy on other downstream tasks. This makes our method well-suited for tasks requiring fairness, robustness, and alignment with specific preferences. We evaluate our method for tackling typographic attacks on images and explore its ability to disentangle gender concepts and mitigate gender bias, showcasing the versatility of our approach.
♻ ☆ MMGDreamer: Mixed-Modality Graph for Geometry-Controllable 3D Indoor Scene Generation AAAI 2025
Controllable 3D scene generation has extensive applications in virtual reality and interior design, where the generated scenes should exhibit high levels of realism and controllability in terms of geometry. Scene graphs provide a suitable data representation that facilitates these applications. However, current graph-based methods for scene generation are constrained to text-based inputs and exhibit insufficient adaptability to flexible user inputs, hindering the ability to precisely control object geometry. To address this issue, we propose MMGDreamer, a dual-branch diffusion model for scene generation that incorporates a novel Mixed-Modality Graph, visual enhancement module, and relation predictor. The mixed-modality graph allows object nodes to integrate textual and visual modalities, with optional relationships between nodes. It enhances adaptability to flexible user inputs and enables meticulous control over the geometry of objects in the generated scenes. The visual enhancement module enriches the visual fidelity of text-only nodes by constructing visual representations using text embeddings. Furthermore, our relation predictor leverages node representations to infer absent relationships between nodes, resulting in more coherent scene layouts. Extensive experimental results demonstrate that MMGDreamer exhibits superior control of object geometry, achieving state-of-the-art scene generation performance. Project page: https://yangzhifeio.github.io/project/MMGDreamer.
comment: Accepted by AAAI 2025 Main Track
♻ ☆ MARVEL-40M+: Multi-Level Visual Elaboration for High-Fidelity Text-to-3D Content Creation
Generating high-fidelity 3D content from text prompts remains a significant challenge in computer vision due to the limited size, diversity, and annotation depth of the existing datasets. To address this, we introduce MARVEL-40M+, an extensive dataset with 40 million text annotations for over 8.9 million 3D assets aggregated from seven major 3D datasets. Our contribution is a novel multi-stage annotation pipeline that integrates open-source pretrained multi-view VLMs and LLMs to automatically produce multi-level descriptions, ranging from detailed (150-200 words) to concise semantic tags (10-20 words). This structure supports both fine-grained 3D reconstruction and rapid prototyping. Furthermore, we incorporate human metadata from source datasets into our annotation pipeline to add domain-specific information in our annotation and reduce VLM hallucinations. Additionally, we develop MARVEL-FX3D, a two-stage text-to-3D pipeline. We fine-tune Stable Diffusion with our annotations and use a pretrained image-to-3D network to generate 3D textured meshes within 15s. Extensive evaluations show that MARVEL-40M+ significantly outperforms existing datasets in annotation quality and linguistic diversity, achieving win rates of 72.41% by GPT-4 and 73.40% by human evaluators. Project page is available at https://sankalpsinha-cmos.github.io/MARVEL/.
♻ ☆ TopoBench: A Framework for Benchmarking Topological Deep Learning
This work introduces TopoBench, an open-source library designed to standardize benchmarking and accelerate research in topological deep learning (TDL). TopoBench decomposes TDL into a sequence of independent modules for data generation, loading, transforming and processing, as well as model training, optimization and evaluation. This modular organization provides flexibility for modifications and facilitates the adaptation and optimization of various TDL pipelines. A key feature of TopoBench is its support for transformations and lifting across topological domains. Mapping the topology and features of a graph to higher-order topological domains, such as simplicial and cell complexes, enables richer data representations and more fine-grained analyses. The applicability of TopoBench is demonstrated by benchmarking several TDL architectures across diverse tasks and datasets.
♻ ☆ Towards Efficient Training of Graph Neural Networks: A Multiscale Approach
Graph Neural Networks (GNNs) have emerged as a powerful tool for learning and inferring from graph-structured data, and are widely used in a variety of applications, often considering large amounts of data and large graphs. However, training on such data requires large memory and extensive computations. In this paper, we introduce a novel framework for efficient multiscale training of GNNs, designed to integrate information across multiscale representations of a graph. Our approach leverages a hierarchical graph representation, taking advantage of coarse graph scales in the training process, where each coarse scale graph has fewer nodes and edges. Based on this approach, we propose a suite of GNN training methods: such as coarse-to-fine, sub-to-full, and multiscale gradient computation. We demonstrate the effectiveness of our methods on various datasets and learning tasks.
♻ ☆ NLPrompt: Noise-Label Prompt Learning for Vision-Language Models
The emergence of vision-language foundation models, such as CLIP, has revolutionized image-text representation, enabling a broad range of applications via prompt learning. Despite its promise, real-world datasets often contain noisy labels that can degrade prompt learning performance. In this paper, we demonstrate that using mean absolute error (MAE) loss in prompt learning, named PromptMAE, significantly enhances robustness against noisy labels while maintaining high accuracy. Though MAE is straightforward and recognized for its robustness, it is rarely used in noisy-label learning due to its slow convergence and poor performance outside prompt learning scenarios. To elucidate the robustness of PromptMAE, we leverage feature learning theory to show that MAE can suppress the influence of noisy samples, thereby improving the signal-to-noise ratio and enhancing overall robustness. Additionally, we introduce PromptOT, a prompt-based optimal transport data purification method to enhance the robustness further. PromptOT employs text features in vision-language models as prototypes to construct an optimal transportation matrix. This matrix effectively partitions datasets into clean and noisy subsets, allowing for the application of cross-entropy loss to the clean subset and MAE loss to the noisy subset. Our Noise-Label Prompt Learning method, named NLPrompt, offers a simple and efficient approach that leverages the expressive representations and precise alignment capabilities of vision-language models for robust prompt learning. We validate NLPrompt through extensive experiments across various noise settings, demonstrating significant performance improvements.
♻ ☆ Integrating Biological-Informed Recurrent Neural Networks for Glucose-Insulin Dynamics Modeling
Type 1 Diabetes (T1D) management is a complex task due to many variability factors. Artificial Pancreas (AP) systems have alleviated patient burden by automating insulin delivery through advanced control algorithms. However, the effectiveness of these systems depends on accurate modeling of glucose-insulin dynamics, which traditional mathematical models often fail to capture due to their inability to adapt to patient-specific variations. This study introduces a Biological-Informed Recurrent Neural Network (BIRNN) framework to address these limitations. The BIRNN leverages a Gated Recurrent Units (GRU) architecture augmented with physics-informed loss functions that embed physiological constraints, ensuring a balance between predictive accuracy and consistency with biological principles. The framework is validated using the commercial UVA/Padova simulator, outperforming traditional linear models in glucose prediction accuracy and reconstruction of unmeasured states, even under circadian variations in insulin sensitivity. The results demonstrate the potential of BIRNN for personalized glucose regulation and future adaptive control strategies in AP systems.
comment: Accepted for pubblication in the proceedings of the Engineering Diabetes Technologies (EDT 2025). 7 pages, 2 figures and 1 table
♻ ☆ Contextual Similarity Distillation: Ensemble Uncertainties with a Single Model
Uncertainty quantification is a critical aspect of reinforcement learning and deep learning, with numerous applications ranging from efficient exploration and stable offline reinforcement learning to outlier detection in medical diagnostics. The scale of modern neural networks, however, complicates the use of many theoretically well-motivated approaches such as full Bayesian inference. Approximate methods like deep ensembles can provide reliable uncertainty estimates but still remain computationally expensive. In this work, we propose contextual similarity distillation, a novel approach that explicitly estimates the variance of an ensemble of deep neural networks with a single model, without ever learning or evaluating such an ensemble in the first place. Our method builds on the predictable learning dynamics of wide neural networks, governed by the neural tangent kernel, to derive an efficient approximation of the predictive variance of an infinite ensemble. Specifically, we reinterpret the computation of ensemble variance as a supervised regression problem with kernel similarities as regression targets. The resulting model can estimate predictive variance at inference time with a single forward pass, and can make use of unlabeled target-domain data or data augmentations to refine its uncertainty estimates. We empirically validate our method across a variety of out-of-distribution detection benchmarks and sparse-reward reinforcement learning environments. We find that our single-model method performs competitively and sometimes superior to ensemble-based baselines and serves as a reliable signal for efficient exploration. These results, we believe, position contextual similarity distillation as a principled and scalable alternative for uncertainty quantification in reinforcement learning and general deep learning.
♻ ☆ T-Graphormer: Using Transformers for Spatiotemporal Forecasting
Spatiotemporal data is ubiquitous, and forecasting it has important applications in many domains. However, its complex cross-component dependencies and non-linear temporal dynamics can be challenging for traditional techniques. Existing methods address this by learning the two dimensions separately. Here, we introduce Temporal Graphormer (T-Graphormer), a Transformer-based approach capable of modelling spatiotemporal correlations simultaneously. By adding temporal encodings in the Graphormer architecture, each node attends to all other tokens within the graph sequence, enabling the model to learn rich spacetime patterns with minimal predefined inductive biases. We show the effectiveness of T-Graphormer on real-world traffic prediction benchmark datasets. Compared to state-of-the-art methods, T-Graphormer reduces root mean squared error (RMSE) and mean absolute percentage error (MAPE) by up to 20% and 10%.
♻ ☆ One Framework to Rule Them All: Unifying RL-Based and RL-Free Methods in RLHF
In this article, we primarily examine a variety of RL-based and RL-free methods designed to address Reinforcement Learning from Human Feedback (RLHF) and Large Reasoning Models (LRMs). We begin with a concise overview of the typical steps involved in RLHF and LRMs. Next, we reinterpret several RL-based and RL-free algorithms through the perspective of neural structured bandit prediction, providing a clear conceptual framework that uncovers a deeper connection between these seemingly distinct approaches. Following this, we briefly review some core principles of reinforcement learning, drawing attention to an often-overlooked aspect in existing RLHF studies. This leads to a detailed derivation of the standard RLHF objective within a full RL context, demonstrating its equivalence to neural structured bandit prediction. Finally, by reinvestigating the principles behind Proximal Policy Optimization (PPO), we pinpoint areas needing adjustment, which culminates in the introduction of the Generalized Reinforce Optimization (GRO) framework, seamlessly integrating RL-based and RL-free methods in RLHF. We look forward to the community's efforts to empirically validate GRO and invite constructive feedback.
♻ ☆ MMRL: Multi-Modal Representation Learning for Vision-Language Models CVPR 2025
Large-scale pre-trained Vision-Language Models (VLMs) have become essential for transfer learning across diverse tasks. However, adapting these models with limited few-shot data often leads to overfitting, diminishing their performance on new tasks. To tackle this issue, we propose a novel Multi-Modal Representation Learning (MMRL) framework that introduces a shared, learnable, and modality-agnostic representation space. MMRL projects the space tokens to text and image representation tokens, facilitating more effective multi-modal interactions. Unlike previous approaches that solely optimize class token features, MMRL integrates representation tokens at higher layers of the encoders--where dataset-specific features are more prominent--while preserving generalized knowledge in the lower layers. During training, both representation and class features are optimized, with trainable projection layer applied to the representation tokens, whereas the class token projection layer remains frozen to retain pre-trained knowledge. Furthermore, a regularization term is introduced to align the class features and text features with the zero-shot features from the frozen VLM, thereby safeguarding the model's generalization capacity. For inference, a decoupling strategy is employed, wherein both representation and class features are utilized for base classes, while only the class features, which retain more generalized knowledge, are used for new tasks. Extensive experiments across 15 datasets demonstrate that MMRL outperforms state-of-the-art methods, achieving a balanced trade-off between task-specific adaptation and generalization. Code is available at https://github.com/yunncheng/MMRL.
comment: Accepted by CVPR 2025
♻ ☆ Making AI Less "Thirsty": Uncovering and Addressing the Secret Water Footprint of AI Models
The growing carbon footprint of artificial intelligence (AI) has been undergoing public scrutiny. Nonetheless, the equally important water (withdrawal and consumption) footprint of AI has largely remained under the radar. For example, training the GPT-3 language model in Microsoft's state-of-the-art U.S. data centers can directly evaporate 700,000 liters of clean freshwater, but such information has been kept a secret. More critically, the global AI demand is projected to account for 4.2-6.6 billion cubic meters of water withdrawal in 2027, which is more than the total annual water withdrawal of 4-6 Denmark or half of the United Kingdom. This is concerning, as freshwater scarcity has become one of the most pressing challenges. To respond to the global water challenges, AI can, and also must, take social responsibility and lead by example by addressing its own water footprint. In this paper, we provide a principled methodology to estimate the water footprint of AI, and also discuss the unique spatial-temporal diversities of AI's runtime water efficiency. Finally, we highlight the necessity of holistically addressing water footprint along with carbon footprint to enable truly sustainable AI.
comment: Accepted by Communications of the ACM. Source codes available at: https://github.com/Ren-Research/Making-AI-Less-Thirsty
♻ ☆ Preference Optimization with Multi-Sample Comparisons
Recent advancements in generative models, particularly large language models (LLMs) and diffusion models, have been driven by extensive pretraining on large datasets followed by post-training. However, current post-training methods such as reinforcement learning from human feedback (RLHF) and direct alignment from preference methods (DAP) primarily utilize single-sample comparisons. These approaches often fail to capture critical characteristics such as generative diversity and bias, which are more accurately assessed through multiple samples. To address these limitations, we introduce a novel approach that extends post-training to include multi-sample comparisons. To achieve this, we propose Multi-sample Direct Preference Optimization (mDPO) and Multi-sample Identity Preference Optimization (mIPO). These methods improve traditional DAP methods by focusing on group-wise characteristics. Empirically, we demonstrate that multi-sample comparison is more effective in optimizing collective characteristics~(e.g., diversity and bias) for generative models than single-sample comparison. Additionally, our findings suggest that multi-sample comparisons provide a more robust optimization framework, particularly for dataset with label noise.
comment: Code is available at https://github.com/alecwangcq/multi-sample-alignment
♻ ☆ Graph-Level Label-Only Membership Inference Attack against Graph Neural Networks
Graph neural networks (GNNs) are widely used for graph-structured data but are vulnerable to membership inference attacks (MIAs) in graph classification tasks, which determine if a graph was part of the training dataset, potentially causing data leakage. Existing MIAs rely on prediction probability vectors, but they become ineffective when only prediction labels are available. We propose a Graph-level Label-Only Membership Inference Attack (GLO-MIA), which is based on the intuition that the target model's predictions on training data are more stable than those on testing data. GLO-MIA generates a set of perturbed graphs for target graph by adding perturbations to its effective features and queries the target model with the perturbed graphs to get their prediction labels, which are then used to calculate robustness score of the target graph. Finally, by comparing the robustness score with a predefined threshold, the membership of the target graph can be inferred correctly with high probability. Our evaluation on three datasets and four GNN models shows that GLO-MIA achieves an attack accuracy of up to 0.825, outperforming baseline work by 8.5% and closely matching the performance of probability-based MIAs, even with only prediction labels.
♻ ☆ Hi-ALPS -- An Experimental Robustness Quantification of Six LiDAR-based Object Detection Systems for Autonomous Driving
Light Detection and Ranging (LiDAR) is an essential sensor technology for autonomous driving as it can capture high-resolution 3D data. As 3D object detection systems (OD) can interpret such point cloud data, they play a key role in the driving decisions of autonomous vehicles. Consequently, such 3D OD must be robust against all types of perturbations and must therefore be extensively tested. One approach is the use of adversarial examples, which are small, sometimes sophisticated perturbations in the input data that change, i.e., falsify, the prediction of the OD. These perturbations are carefully designed based on the weaknesses of the OD. The robustness of the OD cannot be quantified with adversarial examples in general, because if the OD is vulnerable to a given attack, it is unclear whether this is due to the robustness of the OD or whether the attack algorithm produces particularly strong adversarial examples. The contribution of this work is Hi-ALPS -- Hierarchical Adversarial-example-based LiDAR Perturbation Level System, where higher robustness of the OD is required to withstand the perturbations as the perturbation levels increase. In doing so, the Hi-ALPS levels successively implement a heuristic followed by established adversarial example approaches. In a series of comprehensive experiments using Hi-ALPS, we quantify the robustness of six state-of-the-art 3D OD under different types of perturbations. The results of the experiments show that none of the OD is robust against all Hi-ALPS levels; an important factor for the ranking is that human observers can still correctly recognize the perturbed objects, as the respective perturbations are small. To increase the robustness of the OD, we discuss the applicability of state-of-the-art countermeasures. In addition, we derive further suggestions for countermeasures based on our experimental results.
♻ ☆ Learning Partial Graph Matching via Optimal Partial Transport
Partial graph matching extends traditional graph matching by allowing some nodes to remain unmatched, enabling applications in more complex scenarios. However, this flexibility introduces additional complexity, as both the subset of nodes to match and the optimal mapping must be determined. While recent studies have explored deep learning techniques for partial graph matching, a significant limitation remains: the absence of an optimization objective that fully captures the problem's intrinsic nature while enabling efficient solutions. In this paper, we propose a novel optimization framework for partial graph matching, inspired by optimal partial transport. Our approach formulates an objective that enables partial assignments while incorporating matching biases, using weighted total variation as the divergence function to guarantee optimal partial assignments. Our method can achieve efficient, exact solutions within cubic worst case time complexity. Our contributions are threefold: (i) we introduce a novel optimization objective that balances matched and unmatched nodes; (ii) we establish a connection between partial graph matching and linear sum assignment problem, enabling efficient solutions; (iii) we propose a deep graph matching architecture with a novel partial matching loss, providing an end-to-end solution. The empirical evaluations on standard graph matching benchmarks demonstrate the efficacy of the proposed approach.
♻ ☆ Vision-based Multi-future Trajectory Prediction: A Survey
Vision-based trajectory prediction is an important task that supports safe and intelligent behaviours in autonomous systems. Many advanced approaches have been proposed over the years with improved spatial and temporal feature extraction. However, human behaviour is naturally diverse and uncertain. Given the past trajectory and surrounding environment information, an agent can have multiple plausible trajectories in the future. To tackle this problem, an essential task named multi-future trajectory prediction (MTP) has recently been studied. This task aims to generate a diverse, acceptable and explainable distribution of future predictions for each agent. In this paper, we present the first survey for MTP with our unique taxonomies and a comprehensive analysis of frameworks, datasets and evaluation metrics. We also compare models on existing MTP datasets and conduct experiments on the ForkingPath dataset. Finally, we discuss multiple future directions that can help researchers develop novel multi-future trajectory prediction systems and other diverse learning tasks similar to MTP.
comment: Accepted by TNNLS 2025
♻ ☆ Bonsai: Gradient-free Graph Condensation for Node Classification
Graph condensation has emerged as a promising avenue to enable scalable training of GNNs by compressing the training dataset while preserving essential graph characteristics. Our study uncovers significant shortcomings in current graph condensation techniques. First, the majority of the algorithms paradoxically require training on the full dataset to perform condensation. Second, due to their gradient-emulating approach, these methods require fresh condensation for any change in hyperparameters or GNN architecture, limiting their flexibility and reusability. Finally, they fail to achieve substantial size reduction due to synthesizing fully-connected, edge-weighted graphs. To address these challenges, we present Bonsai, a novel graph condensation method empowered by the observation that \textit{computation trees} form the fundamental processing units of message-passing GNNs. Bonsai condenses datasets by encoding a careful selection of \textit{exemplar} trees that maximize the representation of all computation trees in the training set. This unique approach imparts Bonsai as the first linear-time, model-agnostic graph condensation algorithm for node classification that outperforms existing baselines across $7$ real-world datasets on accuracy, while being $22$ times faster on average. Bonsai is grounded in rigorous mathematical guarantees on the adopted approximation strategies making it robust to GNN architectures, datasets, and parameters.
♻ ☆ Parameter-Efficient Fine-Tuning for Pre-Trained Vision Models: A Survey
Large-scale pre-trained vision models (PVMs) have shown great potential for adaptability across various downstream vision tasks. However, with state-of-the-art PVMs growing to billions or even trillions of parameters, the standard full fine-tuning paradigm is becoming unsustainable due to high computational and storage demands. In response, researchers are exploring parameter-efficient fine-tuning (PEFT), which seeks to exceed the performance of full fine-tuning with minimal parameter modifications. This survey provides a comprehensive overview and future directions for visual PEFT, offering a systematic review of the latest advancements. First, we provide a formal definition of PEFT and discuss model pre-training methods. We then categorize existing methods into three categories: addition-based, partial-based, and unified-based. Finally, we introduce the commonly used datasets and applications and suggest potential future research challenges. A comprehensive collection of resources is available at https://github.com/synbol/Awesome-Parameter-Efficient-Transfer-Learning.
comment: 9 pages, 3 figures, 2 tables
♻ ☆ Emergence of the Primacy Effect in Structured State-Space Models
Human and animal memory for sequentially presented items is well-documented to be more accurate for those at the beginning and end of the sequence, phenomena known as the primacy and recency effects, respectively. By contrast, artificial neural network (ANN) models are typically designed with a memory that decays monotonically over time. Accordingly, ANNs are expected to show the recency effect but not the primacy effect. Contrary to this theoretical expectation, however, the present study reveals a counterintuitive finding: a recently developed ANN architecture, called structured state-space models, exhibits the primacy effect when trained and evaluated on a synthetic task that mirrors psychological memory experiments. Given that this model was originally designed for recovering neuronal activity patterns observed in biological brains, this result provides a novel perspective on the psychological primacy effect while also posing a non-trivial puzzle for the current theories in machine learning.
♻ ☆ TransPlace: Transferable Circuit Global Placement via Graph Neural Network KDD 2025
Global placement, a critical step in designing the physical layout of computer chips, is essential to optimize chip performance. Prior global placement methods optimize each circuit design individually from scratch. Their neglect of transferable knowledge limits solution efficiency and chip performance as circuit complexity drastically increases. This study presents TransPlace, a global placement framework that learns to place millions of mixed-size cells in continuous space. TransPlace introduces i) Netlist Graph to efficiently model netlist topology, ii) Cell-flow and relative position encoding to learn SE(2)-invariant representation, iii) a tailored graph neural network architecture for informed parameterization of placement knowledge, and iv) a two-stage strategy for coarse-to-fine placement. Compared to state-of-the-art placement methods, TransPlace-trained on a few high-quality placements-can place unseen circuits with 1.2x speedup while reducing congestion by 30%, timing by 9%, and wirelength by 5%.
comment: Accepted at KDD 2025
♻ ☆ Persistence of Backdoor-based Watermarks for Neural Networks: A Comprehensive Evaluation
Deep Neural Networks (DNNs) have gained considerable traction in recent years due to the unparalleled results they gathered. However, the cost behind training such sophisticated models is resource intensive, resulting in many to consider DNNs to be intellectual property (IP) to model owners. In this era of cloud computing, high-performance DNNs are often deployed all over the internet so that people can access them publicly. As such, DNN watermarking schemes, especially backdoor-based watermarks, have been actively developed in recent years to preserve proprietary rights. Nonetheless, there lies much uncertainty on the robustness of existing backdoor watermark schemes, towards both adversarial attacks and unintended means such as fine-tuning neural network models. One reason for this is that no complete guarantee of robustness can be assured in the context of backdoor-based watermark. In this paper, we extensively evaluate the persistence of recent backdoor-based watermarks within neural networks in the scenario of fine-tuning, we propose/develop a novel data-driven idea to restore watermark after fine-tuning without exposing the trigger set. Our empirical results show that by solely introducing training data after fine-tuning, the watermark can be restored if model parameters do not shift dramatically during fine-tuning. Depending on the types of trigger samples used, trigger accuracy can be reinstated to up to 100%. Our study further explores how the restoration process works using loss landscape visualization, as well as the idea of introducing training data in fine-tuning stage to alleviate watermark vanishing.
comment: Preprint. Under Review
♻ ☆ Attention IoU: Examining Biases in CelebA using Attention Maps CVPR 2025
Computer vision models have been shown to exhibit and amplify biases across a wide array of datasets and tasks. Existing methods for quantifying bias in classification models primarily focus on dataset distribution and model performance on subgroups, overlooking the internal workings of a model. We introduce the Attention-IoU (Attention Intersection over Union) metric and related scores, which use attention maps to reveal biases within a model's internal representations and identify image features potentially causing the biases. First, we validate Attention-IoU on the synthetic Waterbirds dataset, showing that the metric accurately measures model bias. We then analyze the CelebA dataset, finding that Attention-IoU uncovers correlations beyond accuracy disparities. Through an investigation of individual attributes through the protected attribute of Male, we examine the distinct ways biases are represented in CelebA. Lastly, by subsampling the training set to change attribute correlations, we demonstrate that Attention-IoU reveals potential confounding variables not present in dataset labels.
comment: To appear in CVPR 2025. Code and data is available at https://github.com/aaronserianni/attention-iou . 15 pages, 14 figures, including appendix
Inference-Time Policy Steering through Human Interactions ICRA 2025
Generative policies trained with human demonstrations can autonomously accomplish multimodal, long-horizon tasks. However, during inference, humans are often removed from the policy execution loop, limiting the ability to guide a pre-trained policy towards a specific sub-goal or trajectory shape among multiple predictions. Naive human intervention may inadvertently exacerbate distribution shift, leading to constraint violations or execution failures. To better align policy output with human intent without inducing out-of-distribution errors, we propose an Inference-Time Policy Steering (ITPS) framework that leverages human interactions to bias the generative sampling process, rather than fine-tuning the policy on interaction data. We evaluate ITPS across three simulated and real-world benchmarks, testing three forms of human interaction and associated alignment distance metrics. Among six sampling strategies, our proposed stochastic sampling with diffusion policy achieves the best trade-off between alignment and distribution shift. Videos are available at https://yanweiw.github.io/itps/.
comment: ICRA 2025
♻ ☆ Review and Prospect of Algebraic Research in Equivalent Framework between Statistical Mechanics and Machine Learning Theory
Mathematical equivalence between statistical mechanics and machine learning theory has been known since the 20th century, and research based on this equivalence has provided novel methodologies in both theoretical physics and statistical learning theory. It is well known that algebraic approaches in statistical mechanics such as operator algebra enable us to analyze phase transition phenomena mathematically. In this paper, we review and prospect algebraic research in machine learning theory for theoretical physicists who are interested in artificial intelligence. If a learning machine has a hierarchical structure or latent variables, then the random Hamiltonian cannot be expressed by any quadratic perturbation because it has singularities. To study an equilibrium state defined by such a singular random Hamiltonian, algebraic approaches are necessary to derive the asymptotic form of the free energy and the generalization error. We also introduce the most recent advance: the theoretical foundation for the alignment of artificial intelligence is now being constructed based on algebraic learning theory. This paper is devoted to the memory of Professor Huzihiro Araki who is a pioneering founder of algebraic research in both statistical mechanics and quantum field theory.
♻ ☆ Uni$\textbf{F}^2$ace: Fine-grained Face Understanding and Generation with Unified Multimodal Models
Unified multimodal models (UMMs) have emerged as a powerful paradigm in foundational computer vision research, demonstrating significant potential in both image understanding and generation. However, existing research in the face domain primarily focuses on $\textbf{coarse}$ facial attribute understanding, with limited capacity to handle $\textbf{fine-grained}$ facial attributes and without addressing generation capabilities. To overcome these limitations, we propose Uni$\textbf{F}^2$ace, the first UMM tailored specifically for fine-grained face understanding and generation. In general, we train Uni$\textbf{F}^2$ace on a self-constructed, specialized dataset utilizing two mutually beneficial diffusion techniques and a two-level mixture-of-experts architecture. Specifically, we first build a large-scale facial dataset, Uni$\textbf{F}^2$ace-130K, which contains 130K image-text pairs with one million question-answering pairs that span a wide range of facial attributes. Second, we establish a theoretical connection between discrete diffusion score matching and masked generative models, optimizing both evidence lower bounds simultaneously, which significantly improves the model's ability to synthesize facial details. Finally, we introduce both token-level and sequence-level mixture-of-experts, enabling efficient fine-grained representation learning for both understanding and generation tasks. Extensive experiments on Uni$\textbf{F}^2$ace-130K demonstrate that Uni$\textbf{F}^2$ace outperforms existing UMMs and generative models, achieving superior performance across both understanding and generation tasks.
♻ ☆ Medha: Efficiently Serving Multi-Million Context Length LLM Inference Requests Without Approximations
As large language models (LLMs) handle increasingly longer contexts, serving inference requests for context lengths in the range of millions of tokens presents unique challenges. While existing techniques are effective for training, they fail to address the unique challenges of inference, such as varying prefill and decode phases and their associated latency constraints -- like Time to First Token (TTFT) and Time per Output Token (TPOT). Furthermore, no long-context inference solutions address head-of-line blocking today. We present Medha, a system for efficient long-context LLM inference that introduces three key innovations: adaptive chunking with slack-aware scheduling to prevent head-ofline blocking, Sequence Pipeline Parallelism (SPP) to reduce TTFT, and KV Cache Parallelism (KVP) to minimize TPOT. By combining these into a novel 3D parallelism serving engine, Medha achieves unprecedented scale -- supporting contexts up to 10M tokens with production-grade latency. Our evaluation shows Medha reduces median latency by up to 30x compared to state-of-the-art systems when serving a mix of short and long requests, while improving throughput by upwards of 5x. This enables, for the first time, efficient long-context LLM inference at scale without compromising on shorter request latencies or system efficiency.
♻ ☆ Explaining Deep Convolutional Neural Networks for Image Classification by Evolving Local Interpretable Model-agnostic Explanations
Deep convolutional neural networks have proven their effectiveness, and have been acknowledged as the most dominant method for image classification. However, a severe drawback of deep convolutional neural networks is poor explainability. Unfortunately, in many real-world applications, users need to understand the rationale behind the predictions of deep convolutional neural networks when determining whether they should trust the predictions or not. To resolve this issue, a novel genetic algorithm-based method is proposed for the first time to automatically evolve local explanations that can assist users to assess the rationality of the predictions. Furthermore, the proposed method is model-agnostic, i.e., it can be utilised to explain any deep convolutional neural network models. In the experiments, ResNet is used as an example model to be explained, and the ImageNet dataset is selected as the benchmark dataset. DenseNet and MobileNet are further explained to demonstrate the model-agnostic characteristic of the proposed method. The evolved local explanations on four images, randomly selected from ImageNet, are presented, which show that the evolved local explanations are straightforward to be recognised by humans. Moreover, the evolved explanations can explain the predictions of deep convolutional neural networks on all four images very well by successfully capturing meaningful interpretable features of the sample images. Further analysis based on the 30 runs of the experiments exhibits that the evolved local explanations can also improve the probabilities/confidences of the deep convolutional neural network models in making the predictions. The proposed method can obtain local explanations within one minute, which is more than ten times faster than LIME (the state-of-the-art method).
♻ ☆ MamBEV: Enabling State Space Models to Learn Birds-Eye-View Representations
3D visual perception tasks, such as 3D detection from multi-camera images, are essential components of autonomous driving and assistance systems. However, designing computationally efficient methods remains a significant challenge. In this paper, we propose a Mamba-based framework called MamBEV, which learns unified Bird's Eye View (BEV) representations using linear spatio-temporal SSM-based attention. This approach supports multiple 3D perception tasks with significantly improved computational and memory efficiency. Furthermore, we introduce SSM based cross-attention, analogous to standard cross attention, where BEV query representations can interact with relevant image features. Extensive experiments demonstrate MamBEV's promising performance across diverse visual perception metrics, highlighting its advantages in input scaling efficiency compared to existing benchmark models.
♻ ☆ Clustered Switchback Designs for Experimentation Under Spatio-temporal Interference
We consider experimentation in the presence of non-stationarity, inter-unit (spatial) interference, and carry-over effects (temporal interference), where we wish to estimate the global average treatment effect (GATE), the difference between average outcomes having exposed all units at all times to treatment or to control. We suppose spatial interference is described by a graph, where a unit's outcome depends on its neighborhood's treatments, and that temporal interference is described by an MDP, where the transition kernel under either treatment (action) satisfies a rapid mixing condition. We propose a clustered switchback design, where units are grouped into clusters and time steps are grouped into blocks, and each whole cluster-block combination is assigned a single random treatment. Under this design, we show that for graphs that admit good clustering, a truncated Horvitz-Thompson estimator achieves a $\tilde O(1/NT)$ mean squared error (MSE), matching the lower bound up to logarithmic terms for sparse graphs. Our results simultaneously generalize the results from \citet{hu2022switchback,ugander2013graph} and \citet{leung2022rate}. Simulation studies validate the favorable performance of our approach.
♻ ☆ Precise Asymptotic Generalization for Multiclass Classification with Overparameterized Linear Models NeurIPS 2023
We study the asymptotic generalization of an overparameterized linear model for multiclass classification under the Gaussian covariates bi-level model introduced in Subramanian et al.~'22, where the number of data points, features, and classes all grow together. We fully resolve the conjecture posed in Subramanian et al.~'22, matching the predicted regimes for generalization. Furthermore, our new lower bounds are akin to an information-theoretic strong converse: they establish that the misclassification rate goes to 0 or 1 asymptotically. One surprising consequence of our tight results is that the min-norm interpolating classifier can be asymptotically suboptimal relative to noninterpolating classifiers in the regime where the min-norm interpolating regressor is known to be optimal. The key to our tight analysis is a new variant of the Hanson-Wright inequality which is broadly useful for multiclass problems with sparse labels. As an application, we show that the same type of analysis can be used to analyze the related multilabel classification problem under the same bi-level ensemble.
comment: NeurIPS 2023, 56 pages v3: fixed typos in sparse Hanson-Wright theorem statement
♻ ☆ GD-VAEs: Geometric Dynamic Variational Autoencoders for Learning Nonlinear Dynamics and Dimension Reductions
We develop data-driven methods incorporating geometric and topological information to learn parsimonious representations of nonlinear dynamics from observations. The approaches learn nonlinear state-space models of the dynamics for general manifold latent spaces using training strategies related to Variational Autoencoders (VAEs). Our methods are referred to as Geometric Dynamic (GD) Variational Autoencoders (GD-VAEs). We learn encoders and decoders for the system states and evolution based on deep neural network architectures that include general Multilayer Perceptrons (MLPs), Convolutional Neural Networks (CNNs), and other architectures. Motivated by problems arising in parameterized PDEs and physics, we investigate the performance of our methods on tasks for learning reduced dimensional representations of the nonlinear Burgers Equations, Constrained Mechanical Systems, and spatial fields of Reaction-Diffusion Systems. GD-VAEs provide methods that can be used to obtain representations in manifold latent spaces for diverse learning tasks involving dynamics.
comment: 15 figures, related to non-archival proceedings communication
♻ ☆ Do We Need to Verify Step by Step? Rethinking Process Supervision from a Theoretical Perspective
As large language models have evolved, it has become crucial to distinguish between process supervision and outcome supervision -- two key reinforcement learning approaches to complex reasoning tasks. While process supervision offers intuitive advantages for long-term credit assignment, the precise relationship between these paradigms has remained an open question. Conventional wisdom suggests that outcome supervision is fundamentally more challenging due to the trajectory-level coverage problem, leading to significant investment in collecting fine-grained process supervision data. In this paper, we take steps towards resolving this debate. Our main theorem shows that, under standard data coverage assumptions, reinforcement learning through outcome supervision is no more statistically difficult than through process supervision, up to polynomial factors in horizon. At the core of this result lies the novel Change of Trajectory Measure Lemma -- a technical tool that bridges return-based trajectory measure and step-level distribution shift. Furthermore, for settings with access to a verifier or a rollout capability, we prove that any policy's advantage function can serve as an optimal process reward model, providing a direct connection between outcome and process supervision. These findings suggest that the empirically observed performance gap -- if any -- between outcome and process supervision likely stems from algorithmic limitations rather than inherent statistical difficulties, potentially transforming how we approach data collection and algorithm design for reinforcement learning.
♻ ☆ iGAiVA: Integrated Generative AI and Visual Analytics in a Machine Learning Workflow for Text Classification
In developing machine learning (ML) models for text classification, one common challenge is that the collected data is often not ideally distributed, especially when new classes are introduced in response to changes of data and tasks. In this paper, we present a solution for using visual analytics (VA) to guide the generation of synthetic data using large language models. As VA enables model developers to identify data-related deficiency, data synthesis can be targeted to address such deficiency. We discuss different types of data deficiency, describe different VA techniques for supporting their identification, and demonstrate the effectiveness of targeted data synthesis in improving model accuracy. In addition, we present a software tool, iGAiVA, which maps four groups of ML tasks into four VA views, integrating generative AI and VA into an ML workflow for developing and improving text classification models.
♻ ☆ Learning state and proposal dynamics in state-space models using differentiable particle filters and neural networks
State-space models are a popular statistical framework for analysing sequential data. Within this framework, particle filters are often used to perform inference on non-linear state-space models. We introduce a new method, StateMixNN, that uses a pair of neural networks to learn the proposal distribution and transition distribution of a particle filter. Both distributions are approximated using multivariate Gaussian mixtures. The component means and covariances of these mixtures are learnt as outputs of learned functions. Our method is trained targeting the log-likelihood, thereby requiring only the observation series, and combines the interpretability of state-space models with the flexibility and approximation power of artificial neural networks. The proposed method significantly improves recovery of the hidden state in comparison with the state-of-the-art, showing greater improvement in highly non-linear scenarios.
comment: update to accepted version
♻ ☆ No Black Box Anymore: Demystifying Clinical Predictive Modeling with Temporal-Feature Cross Attention Mechanism
Despite the outstanding performance of deep learning models in clinical prediction tasks, explainability remains a significant challenge. Inspired by transformer architectures, we introduce the Temporal-Feature Cross Attention Mechanism (TFCAM), a novel deep learning framework designed to capture dynamic interactions among clinical features across time, enhancing both predictive accuracy and interpretability. In an experiment with 1,422 patients with Chronic Kidney Disease, predicting progression to End-Stage Renal Disease, TFCAM outperformed LSTM and RETAIN baselines, achieving an AUROC of 0.95 and an F1-score of 0.69. Beyond performance gains, TFCAM provides multi-level explainability by identifying critical temporal periods, ranking feature importance, and quantifying how features influence each other across time before affecting predictions. Our approach addresses the "black box" limitations of deep learning in healthcare, offering clinicians transparent insights into disease progression mechanisms while maintaining state-of-the-art predictive performance.
comment: 10 pages, 3 figures, submitted to AMIA 2025
♻ ☆ Follow-the-Regularized-Leader with Adversarial Constraints
Constrained Online Convex Optimization (COCO) can be seen as a generalization of the standard Online Convex Optimization (OCO) framework. At each round, a cost function and constraint function are revealed after a learner chooses an action. The goal is to minimize both the regret and cumulative constraint violation (CCV) against an adaptive adversary. We show for the first time that is possible to obtain the optimal $O(\sqrt{T})$ bound on both regret and CCV, improving the best known bounds of $O \left( \sqrt{T} \right)$ and $\~{O} \left( \sqrt{T} \right)$ for the regret and CCV, respectively.
♻ ☆ Long-context Protein Language Modeling Using Bidirectional Mamba with Shared Projection Layers
Self-supervised training of language models (LMs) has seen great success for protein sequences in learning meaningful representations and for generative drug design. Most protein LMs are based on the Transformer architecture trained on individual proteins with short context lengths. Such protein LMs cannot extrapolate to longer proteins and protein complexes well. They also fail to account for the underlying biological mechanisms carried out by biomolecular interactions and dynamics i.e., proteins often interact with other proteins, molecules, and pathways in complex biological systems. In this work, we propose LC-PLM based on an alternative protein LM architecture, BiMamba-S, built upon selective structured state-space models, to learn high-quality universal protein representations at the amino acid token level using masked language modeling. We also introduce its graph-contextual variant, LC-PLM-G, which contextualizes protein-protein interaction (PPI) graphs for a second stage of training. LC-PLM demonstrates favorable neural scaling laws, better length extrapolation capability, and a 7% to 34% improvement on protein downstream tasks than Transformer-based ESM-2. LC-PLM-G further trained within the context of PPI graphs shows promising results on protein structure and function prediction tasks. Our study demonstrates the benefit of increasing the context size with computationally efficient LM architecture (e.g. structured state space models) in learning universal protein representations and incorporating molecular interaction context contained in biological graphs.
comment: model weights open-sourced at https://github.com/amazon-science/LC-PLM
♻ ☆ Lightweight Online Adaption for Time Series Foundation Model Forecasts
Foundation models (FMs) have emerged as a promising approach for time series forecasting. While effective, FMs typically remain fixed during deployment due to the high computational costs of learning them online. Consequently, deployed FMs fail to adapt their forecasts to current data characteristics, despite the availability of online feedback from newly arriving data. This raises the question of whether FM performance can be enhanced by the efficient usage of this feedback. We propose AdapTS to answer this question. AdapTS is a lightweight mechanism for the online adaption of FM forecasts in response to online feedback. AdapTS consists of two parts: a) the AdapTS-Forecaster which is used to learn the current data distribution; and b) the AdapTS-Weighter which is used to combine the forecasts of the FM and the AdapTS-Forecaster. We evaluate the performance of AdapTS in conjunction with several recent FMs across a suite of standard time series datasets. In all of our experiments we find that using AdapTS improves performance. This work demonstrates how efficient usage of online feedback can be used to improve FM forecasts.
comment: 8 pages, Preprint
♻ ☆ Multi-Agent Inverse Reinforcement Learning in Real World Unstructured Pedestrian Crowds
Social robot navigation in crowded public spaces such as university campuses, restaurants, grocery stores, and hospitals, is an increasingly important area of research. One of the core strategies for achieving this goal is to understand humans' intent--underlying psychological factors that govern their motion--by learning their reward functions, typically via inverse reinforcement learning (IRL). Despite significant progress in IRL, learning reward functions of multiple agents simultaneously in dense unstructured pedestrian crowds has remained intractable due to the nature of the tightly coupled social interactions that occur in these scenarios \textit{e.g.} passing, intersections, swerving, weaving, etc. In this paper, we present a new multi-agent maximum entropy inverse reinforcement learning algorithm for real world unstructured pedestrian crowds. Key to our approach is a simple, but effective, mathematical trick which we name the so-called tractability-rationality trade-off trick that achieves tractability at the cost of a slight reduction in accuracy. We compare our approach to the classical single-agent MaxEnt IRL as well as state-of-the-art trajectory prediction methods on several datasets including the ETH, UCY, SCAND, JRDB, and a new dataset, called Speedway, collected at a busy intersection on a University campus focusing on dense, complex agent interactions. Our key findings show that, on the dense Speedway dataset, our approach ranks 1st among top 7 baselines with >2X improvement over single-agent IRL, and is competitive with state-of-the-art large transformer-based encoder-decoder models on sparser datasets such as ETH/UCY (ranks 3rd among top 7 baselines).
♻ ☆ Policy Learning with a Language Bottleneck
Modern AI systems such as self-driving cars and game-playing agents achieve superhuman performance, but often lack human-like generalization, interpretability, and inter-operability with human users. Inspired by the rich interactions between language and decision-making in humans, we introduce Policy Learning with a Language Bottleneck (PLLB), a framework enabling AI agents to generate linguistic rules that capture the high-level strategies underlying rewarding behaviors. PLLB alternates between a *rule generation* step guided by language models, and an *update* step where agents learn new policies guided by rules, even when a rule is insufficient to describe an entire complex policy. Across five diverse tasks, including a two-player signaling game, maze navigation, image reconstruction, and robot grasp planning, we show that PLLB agents are not only able to learn more interpretable and generalizable behaviors, but can also share the learned rules with human users, enabling more effective human-AI coordination. We provide source code for our experiments at https://github.com/meghabyte/bottleneck .
comment: 21 pages, 15 figures, updated with robot manipulation task
♻ ☆ SimBEV: A Synthetic Multi-Task Multi-Sensor Driving Data Generation Tool and Dataset
Bird's-eye view (BEV) perception has garnered significant attention in autonomous driving in recent years, in part because BEV representation facilitates multi-modal sensor fusion. BEV representation enables a variety of perception tasks including BEV segmentation, a concise view of the environment useful for planning a vehicle's trajectory. However, this representation is not fully supported by existing datasets, and creation of new datasets for this purpose can be a time-consuming endeavor. To address this challenge, we introduce SimBEV. SimBEV is a randomized synthetic data generation tool that is extensively configurable and scalable, supports a wide array of sensors, incorporates information from multiple sources to capture accurate BEV ground truth, and enables a variety of perception tasks including BEV segmentation and 3D object detection. SimBEV is used to create the SimBEV dataset, a large collection of annotated perception data from diverse driving scenarios. SimBEV and the SimBEV dataset are open and available to the public.
♻ ☆ SparseGS: Real-Time 360° Sparse View Synthesis using Gaussian Splatting 3DV 2025
3D Gaussian Splatting (3DGS) has recently enabled real-time rendering of unbounded 3D scenes for novel view synthesis. However, this technique requires dense training views to accurately reconstruct 3D geometry. A limited number of input views will significantly degrade reconstruction quality, resulting in artifacts such as "floaters" and "background collapse" at unseen viewpoints. In this work, we introduce SparseGS, an efficient training pipeline designed to address the limitations of 3DGS in scenarios with sparse training views. SparseGS incorporates depth priors, novel depth rendering techniques, and a pruning heuristic to mitigate floater artifacts, alongside an Unseen Viewpoint Regularization module to alleviate background collapses. Our extensive evaluations on the Mip-NeRF360, LLFF, and DTU datasets demonstrate that SparseGS achieves high-quality reconstruction in both unbounded and forward-facing scenarios, with as few as 12 and 3 input images, respectively, while maintaining fast training and real-time rendering capabilities.
comment: Version accepted to 3DV 2025. Project page: https://github.com/ForMyCat/SparseGS
♻ ☆ Theory on Score-Mismatched Diffusion Models and Zero-Shot Conditional Samplers
The denoising diffusion model has recently emerged as a powerful generative technique, capable of transforming noise into meaningful data. While theoretical convergence guarantees for diffusion models are well established when the target distribution aligns with the training distribution, practical scenarios often present mismatches. One common case is in the zero-shot conditional diffusion sampling, where the target conditional distribution is different from the (unconditional) training distribution. These score-mismatched diffusion models remain largely unexplored from a theoretical perspective. In this paper, we present the first performance guarantee with explicit dimensional dependencies for general score-mismatched diffusion samplers, focusing on target distributions with finite second moments. We show that score mismatches result in an asymptotic distributional bias between the target and sampling distributions, proportional to the accumulated mismatch between the target and training distributions. This result can be directly applied to zero-shot conditional samplers for any conditional model, irrespective of measurement noise. Interestingly, the derived convergence upper bound offers useful guidance for designing a novel bias-optimal zero-shot sampler in linear conditional models that minimizes the asymptotic bias. For such bias-optimal samplers, we further establish convergence guarantees with explicit dependencies on dimension and conditioning, applied to several interesting target distributions, including those with bounded support and Gaussian mixtures. Our findings are supported by numerical studies.
♻ ☆ Efficient Training of Neural Stochastic Differential Equations by Matching Finite Dimensional Distributions
Neural Stochastic Differential Equations (Neural SDEs) have emerged as powerful mesh-free generative models for continuous stochastic processes, with critical applications in fields such as finance, physics, and biology. Previous state-of-the-art methods have relied on adversarial training, such as GANs, or on minimizing distance measures between processes using signature kernels. However, GANs suffer from issues like instability, mode collapse, and the need for specialized training techniques, while signature kernel-based methods require solving linear PDEs and backpropagating gradients through the solver, whose computational complexity scales quadratically with the discretization steps. In this paper, we identify a novel class of strictly proper scoring rules for comparing continuous Markov processes. This theoretical finding naturally leads to a novel approach called Finite Dimensional Matching (FDM) for training Neural SDEs. Our method leverages the Markov property of SDEs to provide a computationally efficient training objective. This scoring rule allows us to bypass the computational overhead associated with signature kernels and reduces the training complexity from $O(D^2)$ to $O(D)$ per epoch, where $D$ represents the number of discretization steps of the process. We demonstrate that FDM achieves superior performance, consistently outperforming existing methods in terms of both computational efficiency and generative quality.
♻ ☆ Broadening Target Distributions for Accelerated Diffusion Models via a Novel Analysis Approach
Accelerated diffusion models hold the potential to significantly enhance the efficiency of standard diffusion processes. Theoretically, these models have been shown to achieve faster convergence rates than the standard $\mathcal O(1/\epsilon^2)$ rate of vanilla diffusion models, where $\epsilon$ denotes the target accuracy. However, current theoretical studies have established the acceleration advantage only for restrictive target distribution classes, such as those with smoothness conditions imposed along the entire sampling path or with bounded support. In this work, we significantly broaden the target distribution classes with a new accelerated stochastic DDPM sampler. In particular, we show that it achieves accelerated performance for three broad distribution classes not considered before. Our first class relies on the smoothness condition posed only to the target density $q_0$, which is far more relaxed than the existing smoothness conditions posed to all $q_t$ along the entire sampling path. Our second class requires only a finite second moment condition, allowing for a much wider class of target distributions than the existing finite-support condition. Our third class is Gaussian mixture, for which our result establishes the first acceleration guarantee. Moreover, among accelerated DDPM type samplers, our results specialized for bounded-support distributions show an improved dependency on the data dimension $d$. Our analysis introduces a novel technique for establishing performance guarantees via constructing a tilting factor representation of the convergence error and utilizing Tweedie's formula to handle Taylor expansion terms. This new analytical framework may be of independent interest.
♻ ☆ Physics-based Deep Learning
This document is a hands-on, comprehensive guide to deep learning in the realm of physical simulations. Rather than just theory, we emphasize practical application: every concept is paired with interactive Jupyter notebooks to get you up and running quickly. Beyond traditional supervised learning, we dive into physical loss-constraints, differentiable simulations, diffusion-based approaches for probabilistic generative AI, as well as reinforcement learning and advanced neural network architectures. These foundations are paving the way for the next generation of scientific foundation models. We are living in an era of rapid transformation. These methods have the potential to redefine what's possible in computational science.
comment: PBDL v0.3, online version: https://www.physicsbaseddeeplearning.org/
Robotics 51
☆ Direct Post-Training Preference Alignment for Multi-Agent Motion Generation Models Using Implicit Feedback from Pre-training Demonstrations ICLR 2025
Recent advancements in LLMs have revolutionized motion generation models in embodied applications. While LLM-type auto-regressive motion generation models benefit from training scalability, there remains a discrepancy between their token prediction objectives and human preferences. As a result, models pre-trained solely with token-prediction objectives often generate behaviors that deviate from what humans would prefer, making post-training preference alignment crucial for producing human-preferred motions. Unfortunately, post-training alignment requires extensive preference rankings of motions generated by the pre-trained model, which are costly to annotate, especially in multi-agent settings. Recently, there has been growing interest in leveraging pre-training demonstrations to scalably generate preference data for post-training alignment. However, these methods often adopt an adversarial assumption, treating all pre-trained model-generated samples as unpreferred examples. This adversarial approach overlooks the valuable signal provided by preference rankings among the model's own generations, ultimately reducing alignment effectiveness and potentially leading to misaligned behaviors. In this work, instead of treating all generated samples as equally bad, we leverage implicit preferences encoded in pre-training demonstrations to construct preference rankings among the pre-trained model's generations, offering more nuanced preference alignment guidance with zero human cost. We apply our approach to large-scale traffic simulation and demonstrate its effectiveness in improving the realism of pre-trained model's generated behaviors, making a lightweight 1M motion generation model comparable to SOTA large imitation-based models by relying solely on implicit feedback from pre-training demonstrations, without additional post-training human preference annotations or high computational costs.
comment: ICLR 2025 Spotlight
Extendable Long-Horizon Planning via Hierarchical Multiscale Diffusion
This paper tackles a novel problem, extendable long-horizon planning-enabling agents to plan trajectories longer than those in training data without compounding errors. To tackle this, we propose the Hierarchical Multiscale Diffuser (HM-Diffuser) and Progressive Trajectory Extension (PTE), an augmentation method that iteratively generates longer trajectories by stitching shorter ones. HM-Diffuser trains on these extended trajectories using a hierarchical structure, efficiently handling tasks across multiple temporal scales. Additionally, we introduce Adaptive Plan Pondering and the Recursive HM-Diffuser, which consolidate hierarchical layers into a single model to process temporal scales recursively. Experimental results demonstrate the effectiveness of our approach, advancing diffusion-based planners for scalable long-horizon planning.
comment: First two authors contributed equally
☆ Learning Scene-Level Signed Directional Distance Function with Ellipsoidal Priors and Neural Residuals
Dense geometric environment representations are critical for autonomous mobile robot navigation and exploration. Recent work shows that implicit continuous representations of occupancy, signed distance, or radiance learned using neural networks offer advantages in reconstruction fidelity, efficiency, and differentiability over explicit discrete representations based on meshes, point clouds, and voxels. In this work, we explore a directional formulation of signed distance, called signed directional distance function (SDDF). Unlike signed distance function (SDF) and similar to neural radiance fields (NeRF), SDDF has a position and viewing direction as input. Like SDF and unlike NeRF, SDDF directly provides distance to the observed surface along the direction, rather than integrating along the view ray, allowing efficient view synthesis. To learn and predict scene-level SDDF efficiently, we develop a differentiable hybrid representation that combines explicit ellipsoid priors and implicit neural residuals. This approach allows the model to effectively handle large distance discontinuities around obstacle boundaries while preserving the ability for dense high-fidelity prediction. We show that SDDF is competitive with the state-of-the-art neural implicit scene models in terms of reconstruction accuracy and rendering efficiency, while allowing differentiable view prediction for robot trajectory optimization.
Gemini Robotics: Bringing AI into the Physical World
Recent advancements in large multimodal models have led to the emergence of remarkable generalist capabilities in digital domains, yet their translation to physical agents such as robots remains a significant challenge. This report introduces a new family of AI models purposefully designed for robotics and built upon the foundation of Gemini 2.0. We present Gemini Robotics, an advanced Vision-Language-Action (VLA) generalist model capable of directly controlling robots. Gemini Robotics executes smooth and reactive movements to tackle a wide range of complex manipulation tasks while also being robust to variations in object types and positions, handling unseen environments as well as following diverse, open vocabulary instructions. We show that with additional fine-tuning, Gemini Robotics can be specialized to new capabilities including solving long-horizon, highly dexterous tasks, learning new short-horizon tasks from as few as 100 demonstrations and adapting to completely novel robot embodiments. This is made possible because Gemini Robotics builds on top of the Gemini Robotics-ER model, the second model we introduce in this work. Gemini Robotics-ER (Embodied Reasoning) extends Gemini's multimodal reasoning capabilities into the physical world, with enhanced spatial and temporal understanding. This enables capabilities relevant to robotics including object detection, pointing, trajectory and grasp prediction, as well as multi-view correspondence and 3D bounding box predictions. We show how this novel combination can support a variety of robotics applications. We also discuss and address important safety considerations related to this new class of robotics foundation models. The Gemini Robotics family marks a substantial step towards developing general-purpose robots that realizes AI's potential in the physical world.
☆ Hyperdimensional Uncertainty Quantification for Multimodal Uncertainty Fusion in Autonomous Vehicles Perception CVPR 2025
Uncertainty Quantification (UQ) is crucial for ensuring the reliability of machine learning models deployed in real-world autonomous systems. However, existing approaches typically quantify task-level output prediction uncertainty without considering epistemic uncertainty at the multimodal feature fusion level, leading to sub-optimal outcomes. Additionally, popular uncertainty quantification methods, e.g., Bayesian approximations, remain challenging to deploy in practice due to high computational costs in training and inference. In this paper, we propose HyperDUM, a novel deterministic uncertainty method (DUM) that efficiently quantifies feature-level epistemic uncertainty by leveraging hyperdimensional computing. Our method captures the channel and spatial uncertainties through channel and patch -wise projection and bundling techniques respectively. Multimodal sensor features are then adaptively weighted to mitigate uncertainty propagation and improve feature fusion. Our evaluations show that HyperDUM on average outperforms the state-of-the-art (SOTA) algorithms by up to 2.01%/1.27% in 3D Object Detection and up to 1.29% improvement over baselines in semantic segmentation tasks under various types of uncertainties. Notably, HyperDUM requires 2.36x less Floating Point Operations and up to 38.30x less parameters than SOTA methods, providing an efficient solution for real-world autonomous systems.
comment: Accepted at CVPR 2025
☆ Hybrid Magnetically and Electrically Powered Metallo-Dielectric Janus Microrobots: Enhanced Motion Control and Operation Beyond Planar Limits
This study introduces the integration of hybrid magnetic and electric actuation mechanisms to achieve advanced motion capabilities for Janus particle (JP) microrobots. We demonstrate enhanced in-plane motion control through versatile control strategies and present the concepts of interplanar transitions and 2.5-dimensional (2.5D) trajectories, enabled by magnetic levitation and electrostatic trapping. These innovations expand the mobility of JPs into 3D space, allowing dynamic operation beyond the limitations of traditional surface-bound motion. Key functionalities include obstacle crossing, transitions to elevated surfaces, and discrete surface patterning enabling highly localized interventions. Using this set of tools, we also showcase the controlled out-of-plane transport of both synthetic and biological cargo. Together, these advancements lay the groundwork for novel microrobot-related applications in microfluidic systems and biomedical research.
☆ EventFly: Event Camera Perception from Ground to the Sky CVPR 2025
Cross-platform adaptation in event-based dense perception is crucial for deploying event cameras across diverse settings, such as vehicles, drones, and quadrupeds, each with unique motion dynamics, viewpoints, and class distributions. In this work, we introduce EventFly, a framework for robust cross-platform adaptation in event camera perception. Our approach comprises three key components: i) Event Activation Prior (EAP), which identifies high-activation regions in the target domain to minimize prediction entropy, fostering confident, domain-adaptive predictions; ii) EventBlend, a data-mixing strategy that integrates source and target event voxel grids based on EAP-driven similarity and density maps, enhancing feature alignment; and iii) EventMatch, a dual-discriminator technique that aligns features from source, target, and blended domains for better domain-invariant learning. To holistically assess cross-platform adaptation abilities, we introduce EXPo, a large-scale benchmark with diverse samples across vehicle, drone, and quadruped platforms. Extensive experiments validate our effectiveness, demonstrating substantial gains over popular adaptation methods. We hope this work can pave the way for more adaptive, high-performing event perception across diverse and complex environments.
comment: CVPR 2025; 30 pages, 8 figures, 16 tables; Project Page at https://event-fly.github.io/
☆ SuperFlow++: Enhanced Spatiotemporal Consistency for Cross-Modal Data Pretraining
LiDAR representation learning has emerged as a promising approach to reducing reliance on costly and labor-intensive human annotations. While existing methods primarily focus on spatial alignment between LiDAR and camera sensors, they often overlook the temporal dynamics critical for capturing motion and scene continuity in driving scenarios. To address this limitation, we propose SuperFlow++, a novel framework that integrates spatiotemporal cues in both pretraining and downstream tasks using consecutive LiDAR-camera pairs. SuperFlow++ introduces four key components: (1) a view consistency alignment module to unify semantic information across camera views, (2) a dense-to-sparse consistency regularization mechanism to enhance feature robustness across varying point cloud densities, (3) a flow-based contrastive learning approach that models temporal relationships for improved scene understanding, and (4) a temporal voting strategy that propagates semantic information across LiDAR scans to improve prediction consistency. Extensive evaluations on 11 heterogeneous LiDAR datasets demonstrate that SuperFlow++ outperforms state-of-the-art methods across diverse tasks and driving conditions. Furthermore, by scaling both 2D and 3D backbones during pretraining, we uncover emergent properties that provide deeper insights into developing scalable 3D foundation models. With strong generalizability and computational efficiency, SuperFlow++ establishes a new benchmark for data-efficient LiDAR-based perception in autonomous driving. The code is publicly available at https://github.com/Xiangxu-0103/SuperFlow
comment: Preprint; 15 pages, 6 figures, 10 tables; Code at https://github.com/Xiangxu-0103/SuperFlow
☆ Visuo-Tactile Object Pose Estimation for a Multi-Finger Robot Hand with Low-Resolution In-Hand Tactile Sensing ICRA
Accurate 3D pose estimation of grasped objects is an important prerequisite for robots to perform assembly or in-hand manipulation tasks, but object occlusion by the robot's own hand greatly increases the difficulty of this perceptual task. Here, we propose that combining visual information and proprioception with binary, low-resolution tactile contact measurements from across the interior surface of an articulated robotic hand can mitigate this issue. The visuo-tactile object-pose-estimation problem is formulated probabilistically in a factor graph. The pose of the object is optimized to align with the three kinds of measurements using a robust cost function to reduce the influence of visual or tactile outlier readings. The advantages of the proposed approach are first demonstrated in simulation: a custom 15-DoF robot hand with one binary tactile sensor per link grasps 17 YCB objects while observed by an RGB-D camera. This low-resolution in-hand tactile sensing significantly improves object-pose estimates under high occlusion and also high visual noise. We also show these benefits through grasping tests with a preliminary real version of our tactile hand, obtaining reasonable visuo-tactile estimates of object pose at approximately 13.3 Hz on average.
comment: Accepted for publication at the IEEE International Conference on Robotics and Automation (ICRA), 2025
☆ A Multi-Agent Framework Integrating Large Language Models and Generative AI for Accelerated Metamaterial Design
Metamaterials, renowned for their exceptional mechanical, electromagnetic, and thermal properties, hold transformative potential across diverse applications, yet their design remains constrained by labor-intensive trial-and-error methods and limited data interoperability. Here, we introduce CrossMatAgent--a novel multi-agent framework that synergistically integrates large language models with state-of-the-art generative AI to revolutionize metamaterial design. By orchestrating a hierarchical team of agents--each specializing in tasks such as pattern analysis, architectural synthesis, prompt engineering, and supervisory feedback--our system leverages the multimodal reasoning of GPT-4o alongside the generative precision of DALL-E 3 and a fine-tuned Stable Diffusion XL model. This integrated approach automates data augmentation, enhances design fidelity, and produces simulation- and 3D printing-ready metamaterial patterns. Comprehensive evaluations, including CLIP-based alignment, SHAP interpretability analyses, and mechanical simulations under varied load conditions, demonstrate the framework's ability to generate diverse, reproducible, and application-ready designs. CrossMatAgent thus establishes a scalable, AI-driven paradigm that bridges the gap between conceptual innovation and practical realization, paving the way for accelerated metamaterial development.
☆ OpenLex3D: A New Evaluation Benchmark for Open-Vocabulary 3D Scene Representations
3D scene understanding has been transformed by open-vocabulary language models that enable interaction via natural language. However, the evaluation of these representations is limited to closed-set semantics that do not capture the richness of language. This work presents OpenLex3D, a dedicated benchmark to evaluate 3D open-vocabulary scene representations. OpenLex3D provides entirely new label annotations for 23 scenes from Replica, ScanNet++, and HM3D, which capture real-world linguistic variability by introducing synonymical object categories and additional nuanced descriptions. By introducing an open-set 3D semantic segmentation task and an object retrieval task, we provide insights on feature precision, segmentation, and downstream capabilities. We evaluate various existing 3D open-vocabulary methods on OpenLex3D, showcasing failure cases, and avenues for improvement. The benchmark is publicly available at: https://openlex3d.github.io/.
☆ Dita: Scaling Diffusion Transformer for Generalist Vision-Language-Action Policy
While recent vision-language-action models trained on diverse robot datasets exhibit promising generalization capabilities with limited in-domain data, their reliance on compact action heads to predict discretized or continuous actions constrains adaptability to heterogeneous action spaces. We present Dita, a scalable framework that leverages Transformer architectures to directly denoise continuous action sequences through a unified multimodal diffusion process. Departing from prior methods that condition denoising on fused embeddings via shallow networks, Dita employs in-context conditioning -- enabling fine-grained alignment between denoised actions and raw visual tokens from historical observations. This design explicitly models action deltas and environmental nuances. By scaling the diffusion action denoiser alongside the Transformer's scalability, Dita effectively integrates cross-embodiment datasets across diverse camera perspectives, observation scenes, tasks, and action spaces. Such synergy enhances robustness against various variances and facilitates the successful execution of long-horizon tasks. Evaluations across extensive benchmarks demonstrate state-of-the-art or comparative performance in simulation. Notably, Dita achieves robust real-world adaptation to environmental variances and complex long-horizon tasks through 10-shot finetuning, using only third-person camera inputs. The architecture establishes a versatile, lightweight and open-source baseline for generalist robot policy learning. Project Page: https://robodita.github.io.
comment: Preprint; https://robodita.github.io;
☆ Semi-SD: Semi-Supervised Metric Depth Estimation via Surrounding Cameras for Autonomous Driving
In this paper, we introduce Semi-SD, a novel metric depth estimation framework tailored for surrounding cameras equipment in autonomous driving. In this work, the input data consists of adjacent surrounding frames and camera parameters. We propose a unified spatial-temporal-semantic fusion module to construct the visual fused features. Cross-attention components for surrounding cameras and adjacent frames are utilized to focus on metric scale information refinement and temporal feature matching. Building on this, we propose a pose estimation framework using surrounding cameras, their corresponding estimated depths, and extrinsic parameters, which effectively address the scale ambiguity in multi-camera setups. Moreover, semantic world model and monocular depth estimation world model are integrated to supervised the depth estimation, which improve the quality of depth estimation. We evaluate our algorithm on DDAD and nuScenes datasets, and the results demonstrate that our method achieves state-of-the-art performance in terms of surrounding camera based depth estimation quality. The source code will be available on https://github.com/xieyuser/Semi-SD.
☆ Leveraging Cognitive States for Adaptive Scaffolding of Understanding in Explanatory Tasks in HRI
Understanding how scaffolding strategies influence human understanding in human-robot interaction is important for developing effective assistive systems. This empirical study investigates linguistic scaffolding strategies based on negation as an important means that de-biases the user from potential errors but increases processing costs and hesitations as a means to ameliorate processing costs. In an adaptive strategy, the user state with respect to the current state of understanding and processing capacity was estimated via a scoring scheme based on task performance, prior scaffolding strategy, and current eye gaze behavior. In the study, the adaptive strategy of providing negations and hesitations was compared with a non-adaptive strategy of providing only affirmations. The adaptive scaffolding strategy was generated using the computational model SHIFT. Our findings indicate that using adaptive scaffolding strategies with SHIFT tends to (1) increased processing costs, as reflected in longer reaction times, but (2) improved task understanding, evidenced by a lower error rate of almost 23%. We assessed the efficiency of SHIFT's selected scaffolding strategies across different cognitive states, finding that in three out of five states, the error rate was lower compared to the baseline condition. We discuss how these results align with the assumptions of the SHIFT model and highlight areas for refinement. Moreover, we demonstrate how scaffolding strategies, such as negation and hesitation, contribute to more effective human-robot explanatory dialogues.
comment: 8 pages, 6 figures
☆ Risk-Aware Reinforcement Learning for Autonomous Driving: Improving Safety When Driving through Intersection
Applying reinforcement learning to autonomous driving has garnered widespread attention. However, classical reinforcement learning methods optimize policies by maximizing expected rewards but lack sufficient safety considerations, often putting agents in hazardous situations. This paper proposes a risk-aware reinforcement learning approach for autonomous driving to improve the safety performance when crossing the intersection. Safe critics are constructed to evaluate driving risk and work in conjunction with the reward critic to update the actor. Based on this, a Lagrangian relaxation method and cyclic gradient iteration are combined to project actions into a feasible safe region. Furthermore, a Multi-hop and Multi-layer perception (MLP) mixed Attention Mechanism (MMAM) is incorporated into the actor-critic network, enabling the policy to adapt to dynamic traffic and overcome permutation sensitivity challenges. This allows the policy to focus more effectively on surrounding potential risks while enhancing the identification of passing opportunities. Simulation tests are conducted on different tasks at unsignalized intersections. The results show that the proposed approach effectively reduces collision rates and improves crossing efficiency in comparison to baseline algorithms. Additionally, our ablation experiments demonstrate the benefits of incorporating risk-awareness and MMAM into RL.
comment: 11 pages, 10 figures
☆ Energy-aware Joint Orchestration of 5G and Robots: Experimental Testbed and Field Validation
5G mobile networks introduce a new dimension for connecting and operating mobile robots in outdoor environments, leveraging cloud-native and offloading features of 5G networks to enable fully flexible and collaborative cloud robot operations. However, the limited battery life of robots remains a significant obstacle to their effective adoption in real-world exploration scenarios. This paper explores, via field experiments, the potential energy-saving gains of OROS, a joint orchestration of 5G and Robot Operating System (ROS) that coordinates multiple 5G-connected robots both in terms of navigation and sensing, as well as optimizes their cloud-native service resource utilization while minimizing total resource and energy consumption on the robots based on real-time feedback. We designed, implemented and evaluated our proposed OROS in an experimental testbed composed of commercial off-the-shelf robots and a local 5G infrastructure deployed on a campus. The experimental results demonstrated that OROS significantly outperforms state-of-the-art approaches in terms of energy savings by offloading demanding computational tasks to the 5G edge infrastructure and dynamic energy management of on-board sensors (e.g., switching them off when they are not needed). This strategy achieves approximately 15% energy savings on the robots, thereby extending battery life, which in turn allows for longer operating times and better resource utilization.
comment: 14 pages, 15 figures, journal
☆ ZodiAq: An Isotropic Flagella-Inspired Soft Underwater Drone for Safe Marine Exploration
The inherent challenges of robotic underwater exploration, such as hydrodynamic effects, the complexity of dynamic coupling, and the necessity for sensitive interaction with marine life, call for the adoption of soft robotic approaches in marine exploration. To address this, we present a novel prototype, ZodiAq, a soft underwater drone inspired by prokaryotic bacterial flagella. ZodiAq's unique dodecahedral structure, equipped with 12 flagella-like arms, ensures design redundancy and compliance, ideal for navigating complex underwater terrains. The prototype features a central unit based on a Raspberry Pi, connected to a sensory system for inertial, depth, and vision detection, and an acoustic modem for communication. Combined with the implemented control law, it renders ZodiAq an intelligent system. This paper details the design and fabrication process of ZodiAq, highlighting design choices and prototype capabilities. Based on the strain-based modeling of Cosserat rods, we have developed a digital twin of the prototype within a simulation toolbox to ease analysis and control. To optimize its operation in dynamic aquatic conditions, a simplified model-based controller has been developed and implemented, facilitating intelligent and adaptive movement in the hydrodynamic environment. Extensive experimental demonstrations highlight the drone's potential, showcasing its design redundancy, embodied intelligence, crawling gait, and practical applications in diverse underwater settings. This research contributes significantly to the field of underwater soft robotics, offering a promising new avenue for safe, efficient, and environmentally conscious underwater exploration.
comment: 43 pages, including disclaimer page, pre-peer-review version of the manuscript, and supplementary material
☆ DataPlatter: Boosting Robotic Manipulation Generalization with Minimal Costly Data
The growing adoption of Vision-Language-Action (VLA) models in embodied AI intensifies the demand for diverse manipulation demonstrations. However, high costs associated with data collection often result in insufficient data coverage across all scenarios, which limits the performance of the models. It is observed that the spatial reasoning phase (SRP) in large workspace dominates the failure cases. Fortunately, this data can be collected with low cost, underscoring the potential of leveraging inexpensive data to improve model performance. In this paper, we introduce the DataPlatter method, a framework that decouples training trajectories into distinct task stages and leverages abundant easily collectible SRP data to enhance VLA model's generalization. Through analysis we demonstrate that sub-task-specific training with additional SRP data with proper proportion can act as a performance catalyst for robot manipulation, maximizing the utilization of costly physical interaction phase (PIP) data. Experiments show that through introducing large proportion of cost-effective SRP trajectories into a limited set of PIP data, we can achieve a maximum improvement of 41\% on success rate in zero-shot scenes, while with the ability to transfer manipulation skill to novel targets.
☆ RoboFlamingo-Plus: Fusion of Depth and RGB Perception with Vision-Language Models for Enhanced Robotic Manipulation
As robotic technologies advancing towards more complex multimodal interactions and manipulation tasks, the integration of advanced Vision-Language Models (VLMs) has become a key driver in the field. Despite progress with current methods, challenges persist in fusing depth and RGB information within 3D environments and executing tasks guided by linguistic instructions. In response to these challenges, we have enhanced the existing RoboFlamingo framework by introducing RoboFlamingo-Plus, which incorporates depth data into VLMs to significantly improve robotic manipulation performance. Our research achieves a nuanced fusion of RGB and depth information by integrating a pre-trained Vision Transformer (ViT) with a resampling technique, closely aligning this combined data with linguistic cues for superior multimodal understanding. The novelty of RoboFlamingo-Plus lies in its adaptation of inputs for depth data processing, leveraging a pre-trained resampler for depth feature extraction, and employing cross-attention mechanisms for optimal feature integration. These improvements allow RoboFlamingo-Plus to not only deeply understand 3D environments but also easily perform complex, language-guided tasks in challenging settings. Experimental results show that RoboFlamingo-Plus boosts robotic manipulation by 10-20% over current methods, marking a significant advancement. Codes and model weights are public at RoboFlamingo-Plus.
☆ MM-LINS: a Multi-Map LiDAR-Inertial System for Over-Degenerate Environments
SLAM plays a crucial role in automation tasks, such as warehouse logistics, healthcare robotics, and restaurant delivery. These scenes come with various challenges, including navigating around crowds of people, dealing with flying plastic bags that can temporarily blind sensors, and addressing reduced LiDAR density caused by cooking smoke. Such scenarios can result in over-degeneracy, causing the map to drift. To address this issue, this paper presents a multi-map LiDAR-inertial system (MM-LINS) for the first time. The front-end employs an iterated error state Kalman filter for state estimation and introduces a reliable evaluation strategy for degeneracy detection. If over-degeneracy is detected, the active map will be stored into sleeping maps. Subsequently, the system continuously attempts to construct new maps using a dynamic initialization method to ensure successful initialization upon leaving the over-degeneracy. Regarding the back-end, the Scan Context descriptor is utilized to detect inter-map similarity. Upon successful recognition of a sleeping map that shares a common region with the active map, the overlapping trajectory region is utilized to constrain the positional transformation near the edge of the prior map. In response to this, a constraint-enhanced map fusion strategy is proposed to achieve high-precision positional and mapping results. Experiments have been conducted separately on both public datasets that exhibited over-degenerate conditions and in real-world environments. These tests demonstrated the effectiveness of MM-LINS in over-degeneracy environment. Our codes are open-sourced on Github.
comment: Accepted by IEEE Transactions on Intelligent Vehicles
☆ Body Discovery of Embodied AI
In the pursuit of realizing artificial general intelligence (AGI), the importance of embodied artificial intelligence (AI) becomes increasingly apparent. Following this trend, research integrating robots with AGI has become prominent. As various kinds of embodiments have been designed, adaptability to diverse embodiments will become important to AGI. We introduce a new challenge, termed "Body Discovery of Embodied AI", focusing on tasks of recognizing embodiments and summarizing neural signal functionality. The challenge encompasses the precise definition of an AI body and the intricate task of identifying embodiments in dynamic environments, where conventional approaches often prove inadequate. To address these challenges, we apply causal inference method and evaluate it by developing a simulator tailored for testing algorithms with virtual environments. Finally, we validate the efficacy of our algorithms through empirical testing, demonstrating their robust performance in various scenarios based on virtual environments.
☆ G-DexGrasp: Generalizable Dexterous Grasping Synthesis Via Part-Aware Prior Retrieval and Prior-Assisted Generation
Recent advances in dexterous grasping synthesis have demonstrated significant progress in producing reasonable and plausible grasps for many task purposes. But it remains challenging to generalize to unseen object categories and diverse task instructions. In this paper, we propose G-DexGrasp, a retrieval-augmented generation approach that can produce high-quality dexterous hand configurations for unseen object categories and language-based task instructions. The key is to retrieve generalizable grasping priors, including the fine-grained contact part and the affordance-related distribution of relevant grasping instances, for the following synthesis pipeline. Specifically, the fine-grained contact part and affordance act as generalizable guidance to infer reasonable grasping configurations for unseen objects with a generative model, while the relevant grasping distribution plays as regularization to guarantee the plausibility of synthesized grasps during the subsequent refinement optimization. Our comparison experiments validate the effectiveness of our key designs for generalization and demonstrate the remarkable performance against the existing approaches. Project page: https://g-dexgrasp.github.io/
comment: 11 pages, 5 figures
☆ Quality-focused Active Adversarial Policy for Safe Grasping in Human-Robot Interaction
Vision-guided robot grasping methods based on Deep Neural Networks (DNNs) have achieved remarkable success in handling unknown objects, attributable to their powerful generalizability. However, these methods with this generalizability tend to recognize the human hand and its adjacent objects as graspable targets, compromising safety during Human-Robot Interaction (HRI). In this work, we propose the Quality-focused Active Adversarial Policy (QFAAP) to solve this problem. Specifically, the first part is the Adversarial Quality Patch (AQP), wherein we design the adversarial quality patch loss and leverage the grasp dataset to optimize a patch with high quality scores. Next, we construct the Projected Quality Gradient Descent (PQGD) and integrate it with the AQP, which contains only the hand region within each real-time frame, endowing the AQP with fast adaptability to the human hand shape. Through AQP and PQGD, the hand can be actively adversarial with the surrounding objects, lowering their quality scores. Therefore, further setting the quality score of the hand to zero will reduce the grasping priority of both the hand and its adjacent objects, enabling the robot to grasp other objects away from the hand without emergency stops. We conduct extensive experiments on the benchmark datasets and a cobot, showing the effectiveness of QFAAP. Our code and demo videos are available here: https://github.com/clee-jaist/QFAAP.
☆ MATT-GS: Masked Attention-based 3DGS for Robot Perception and Object Detection IROS
This paper presents a novel masked attention-based 3D Gaussian Splatting (3DGS) approach to enhance robotic perception and object detection in industrial and smart factory environments. U2-Net is employed for background removal to isolate target objects from raw images, thereby minimizing clutter and ensuring that the model processes only relevant data. Additionally, a Sobel filter-based attention mechanism is integrated into the 3DGS framework to enhance fine details - capturing critical features such as screws, wires, and intricate textures essential for high-precision tasks. We validate our approach using quantitative metrics, including L1 loss, SSIM, PSNR, comparing the performance of the background-removed and attention-incorporated 3DGS model against the ground truth images and the original 3DGS training baseline. The results demonstrate significant improves in visual fidelity and detail preservation, highlighting the effectiveness of our method in enhancing robotic vision for object recognition and manipulation in complex industrial settings.
comment: This work has been submitted to the 2025 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) for possible publication
☆ Towards Uncertainty Unification: A Case Study for Preference Learning
Learning human preferences is essential for human-robot interaction, as it enables robots to adapt their behaviors to align with human expectations and goals. However, the inherent uncertainties in both human behavior and robotic systems make preference learning a challenging task. While probabilistic robotics algorithms offer uncertainty quantification, the integration of human preference uncertainty remains underexplored. To bridge this gap, we introduce uncertainty unification and propose a novel framework, uncertainty-unified preference learning (UUPL), which enhances Gaussian Process (GP)-based preference learning by unifying human and robot uncertainties. Specifically, UUPL includes a human preference uncertainty model that improves GP posterior mean estimation, and an uncertainty-weighted Gaussian Mixture Model (GMM) that enhances GP predictive variance accuracy. Additionally, we design a user-specific calibration process to align uncertainty representations across users, ensuring consistency and reliability in the model performance. Comprehensive experiments and user studies demonstrate that UUPL achieves state-of-the-art performance in both prediction accuracy and user rating. An ablation study further validates the effectiveness of human uncertainty model and uncertainty-weighted GMM of UUPL.
☆ Observation Adaptation via Annealed Importance Resampling for Partially Observable Markov Decision Processes ICAPS 2025
Partially observable Markov decision processes (POMDPs) are a general mathematical model for sequential decision-making in stochastic environments under state uncertainty. POMDPs are often solved \textit{online}, which enables the algorithm to adapt to new information in real time. Online solvers typically use bootstrap particle filters based on importance resampling for updating the belief distribution. Since directly sampling from the ideal state distribution given the latest observation and previous state is infeasible, particle filters approximate the posterior belief distribution by propagating states and adjusting weights through prediction and resampling steps. However, in practice, the importance resampling technique often leads to particle degeneracy and sample impoverishment when the state transition model poorly aligns with the posterior belief distribution, especially when the received observation is highly informative. We propose an approach that constructs a sequence of bridge distributions between the state-transition and optimal distributions through iterative Monte Carlo steps, better accommodating noisy observations in online POMDP solvers. Our algorithm demonstrates significantly superior performance compared to state-of-the-art methods when evaluated across multiple challenging POMDP domains.
comment: Accepted as Oral Presentation to ICAPS 2025
☆ A Novel Underwater Vehicle With Orientation Adjustable Thrusters: Design and Adaptive Tracking Control
Autonomous underwater vehicles (AUVs) are essential for marine exploration and research. However, conventional designs often struggle with limited maneuverability in complex, dynamic underwater environments. This paper introduces an innovative orientation-adjustable thruster AUV (OATAUV), equipped with a redundant vector thruster configuration that enables full six-degree-of-freedom (6-DOF) motion and composite maneuvers. To overcome challenges associated with uncertain model parameters and environmental disturbances, a novel feedforward adaptive model predictive controller (FFAMPC) is proposed to ensure robust trajectory tracking, which integrates real-time state feedback with adaptive parameter updates. Extensive experiments, including closed-loop tracking and composite motion tests in a laboratory pool, validate the enhanced performance of the OAT-AUV. The results demonstrate that the OAT-AUV's redundant vector thruster configuration enables 23.8% cost reduction relative to common vehicles, while the FF-AMPC controller achieves 68.6% trajectory tracking improvement compared to PID controllers. Uniquely, the system executes composite helical/spiral trajectories unattainable by similar vehicles.
☆ CubeRobot: Grounding Language in Rubik's Cube Manipulation via Vision-Language Model
Proving Rubik's Cube theorems at the high level represents a notable milestone in human-level spatial imagination and logic thinking and reasoning. Traditional Rubik's Cube robots, relying on complex vision systems and fixed algorithms, often struggle to adapt to complex and dynamic scenarios. To overcome this limitation, we introduce CubeRobot, a novel vision-language model (VLM) tailored for solving 3x3 Rubik's Cubes, empowering embodied agents with multimodal understanding and execution capabilities. We used the CubeCoT image dataset, which contains multiple-level tasks (43 subtasks in total) that humans are unable to handle, encompassing various cube states. We incorporate a dual-loop VisionCoT architecture and Memory Stream, a paradigm for extracting task-related features from VLM-generated planning queries, thus enabling CubeRobot to independent planning, decision-making, reflection and separate management of high- and low-level Rubik's Cube tasks. Furthermore, in low-level Rubik's Cube restoration tasks, CubeRobot achieved a high accuracy rate of 100%, similar to 100% in medium-level tasks, and achieved an accuracy rate of 80% in high-level tasks.
☆ CoinFT: A Coin-Sized, Capacitive 6-Axis Force Torque Sensor for Robotic Applications
We introduce CoinFT, a capacitive 6-axis force/torque (F/T) sensor that is compact, light, low-cost, and robust with an average mean-squared error of 0.11N for force and 0.84mNm for moment when the input ranges from 0~10N and 0~4N in normal and shear directions, respectively. CoinFT is a stack of two rigid PCBs with comb-shaped electrodes connected by an array of silicone rubber pillars. The microcontroller interrogates the electrodes in different subsets in order to enhance sensitivity for measuring 6-axis F/T. The combination of desirable features of CoinFT enables various contact-rich robot interactions at a scale, across different embodiment domains including drones, robot end-effectors, and wearable haptic devices. We demonstrate the utility of CoinFT on drones by performing an attitude-based force control to perform tasks that require careful contact force modulation. The design, fabrication, and firmware of CoinFT are open-sourced at https://hojung-choi.github.io/coinft.github.io/.
☆ Benchmarking Multi-Object Grasping
In this work, we describe a multi-object grasping benchmark to evaluate the grasping and manipulation capabilities of robotic systems in both pile and surface scenarios. The benchmark introduces three robot multi-object grasping benchmarking protocols designed to challenge different aspects of robotic manipulation. These protocols are: 1) the Only-Pick-Once protocol, which assesses the robot's ability to efficiently pick multiple objects in a single attempt; 2) the Accurate pick-trnsferring protocol, which evaluates the robot's capacity to selectively grasp and transport a specific number of objects from a cluttered environment; and 3) the Pick-transferring-all protocol, which challenges the robot to clear an entire scene by sequentially grasping and transferring all available objects. These protocols are intended to be adopted by the broader robotics research community, providing a standardized method to assess and compare robotic systems' performance in multi-object grasping tasks. We establish baselines for these protocols using standard planning and perception algorithms on a Barrett hand, Robotiq parallel jar gripper, and the Pisa/IIT Softhand-2, which is a soft underactuated robotic hand. We discuss the results in relation to human performance in similar tasks we well.
comment: This paper contains 11 pages and 5 figures. This paper is under review of a robotics journal
♻ ☆ Aether: Geometric-Aware Unified World Modeling
The integration of geometric reconstruction and generative modeling remains a critical challenge in developing AI systems capable of human-like spatial reasoning. This paper proposes Aether, a unified framework that enables geometry-aware reasoning in world models by jointly optimizing three core capabilities: (1) 4D dynamic reconstruction, (2) action-conditioned video prediction, and (3) goal-conditioned visual planning. Through task-interleaved feature learning, Aether achieves synergistic knowledge sharing across reconstruction, prediction, and planning objectives. Building upon video generation models, our framework demonstrates unprecedented synthetic-to-real generalization despite never observing real-world data during training. Furthermore, our approach achieves zero-shot generalization in both action following and reconstruction tasks, thanks to its intrinsic geometric modeling. Remarkably, even without real-world data, its reconstruction performance is comparable with or even better than that of domain-specific models. Additionally, Aether employs camera trajectories as geometry-informed action spaces, enabling effective action-conditioned prediction and visual planning. We hope our work inspires the community to explore new frontiers in physically-reasonable world modeling and its applications.
comment: Project Page: https://aether-world.github.io/
♻ ☆ Any6D: Model-free 6D Pose Estimation of Novel Objects CVPR 2025
We introduce Any6D, a model-free framework for 6D object pose estimation that requires only a single RGB-D anchor image to estimate both the 6D pose and size of unknown objects in novel scenes. Unlike existing methods that rely on textured 3D models or multiple viewpoints, Any6D leverages a joint object alignment process to enhance 2D-3D alignment and metric scale estimation for improved pose accuracy. Our approach integrates a render-and-compare strategy to generate and refine pose hypotheses, enabling robust performance in scenarios with occlusions, non-overlapping views, diverse lighting conditions, and large cross-environment variations. We evaluate our method on five challenging datasets: REAL275, Toyota-Light, HO3D, YCBINEOAT, and LM-O, demonstrating its effectiveness in significantly outperforming state-of-the-art methods for novel object pose estimation. Project page: https://taeyeop.com/any6d
comment: CVPR 2025, Project Page: https://taeyeop.com/any6d
♻ ☆ Deep learning framework for action prediction reveals multi-timescale locomotor control
Modeling movement in real-world tasks is a fundamental goal for motor control, biomechanics, and rehabilitation engineering. However, widely used data-driven models of essential tasks like locomotion make simplifying assumptions such as linear and fixed timescale mappings between past inputs and future actions, which do not generalize to real-world contexts. Here, we develop a deep learning-based framework for action prediction with architecture-dependent trial embeddings, outperforming traditional models across contexts (walking and running, treadmill and overground, varying terrains) and input modalities (multiple body states, gaze). We find that neural network architectures with flexible input history-dependence like GRU and Transformer perform best overall. By quantifying the model's predictions relative to an autoregressive baseline, we identify context- and modality-dependent timescales. These analyses reveal that there is greater reliance on fast-timescale predictions in complex terrain, gaze predicts future foot placement before body states, and the full-body state predictions precede those by center-of-mass-relevant states. This deep learning framework for action prediction provides quantifiable insights into the control of real-world locomotion and can be extended to other actions, contexts, and populations.
♻ ☆ AO-Grasp: Articulated Object Grasp Generation
We introduce AO-Grasp, a grasp proposal method that generates 6 DoF grasps that enable robots to interact with articulated objects, such as opening and closing cabinets and appliances. AO-Grasp consists of two main contributions: the AO-Grasp Model and the AO-Grasp Dataset. Given a segmented partial point cloud of a single articulated object, the AO-Grasp Model predicts the best grasp points on the object with an Actionable Grasp Point Predictor. Then, it finds corresponding grasp orientations for each of these points, resulting in stable and actionable grasp proposals. We train the AO-Grasp Model on our new AO-Grasp Dataset, which contains 78K actionable parallel-jaw grasps on synthetic articulated objects. In simulation, AO-Grasp achieves a 45.0 % grasp success rate, whereas the highest performing baseline achieves a 35.0% success rate. Additionally, we evaluate AO-Grasp on 120 real-world scenes of objects with varied geometries, articulation axes, and joint states, where AO-Grasp produces successful grasps on 67.5% of scenes, while the baseline only produces successful grasps on 33.3% of scenes. To the best of our knowledge, AO-Grasp is the first method for generating 6 DoF grasps on articulated objects directly from partial point clouds without requiring part detection or hand-designed grasp heuristics. Project website: https://stanford-iprl-lab.github.io/ao-grasp
comment: Project website: https://stanford-iprl-lab.github.io/ao-grasp
♻ ☆ TwoStep: Multi-agent Task Planning using Classical Planners and Large Language Models
Classical planning formulations like the Planning Domain Definition Language (PDDL) admit action sequences guaranteed to achieve a goal state given an initial state if any are possible. However, reasoning problems defined in PDDL do not capture temporal aspects of action taking, such as concurrent actions between two agents when there are no conflicting conditions, without significant modification and definition to existing PDDL domains. A human expert aware of such constraints can decompose a goal into subgoals, each reachable through single agent planning, to take advantage of simultaneous actions. In contrast to classical planning, large language models (LLMs) directly used for inferring plan steps rarely guarantee execution success, but are capable of leveraging commonsense reasoning to assemble action sequences. We combine the strengths of both classical planning and LLMs by approximating human intuitions for multi-agent planning goal decomposition. We demonstrate that LLM-based goal decomposition leads to faster planning times than solving multi-agent PDDL problems directly while simultaneously achieving fewer plan execution steps than a single agent plan alone, as well as most multiagent plans, while guaranteeing execution success. Additionally, we find that LLM-based approximations of subgoals result in similar multi-agent execution lengths to those specified by human experts. Website and resources at https://glamor-usc.github.io/twostep
comment: 14 pages
♻ ☆ Reactive Collision Avoidance for Safe Agile Navigation
Reactive collision avoidance is essential for agile robots navigating complex and dynamic environments, enabling real-time obstacle response. However, this task is inherently challenging because it requires a tight integration of perception, planning, and control, which traditional methods often handle separately, resulting in compounded errors and delays. This paper introduces a novel approach that unifies these tasks into a single reactive framework using solely onboard sensing and computing. Our method combines nonlinear model predictive control with adaptive control barrier functions, directly linking perception-driven constraints to real-time planning and control. Constraints are determined by using a neural network to refine noisy RGB-D data, enhancing depth accuracy, and selecting points with the minimum time-to-collision to prioritize the most immediate threats. To maintain a balance between safety and agility, a heuristic dynamically adjusts the optimization process, preventing overconstraints in real time. Extensive experiments with an agile quadrotor demonstrate effective collision avoidance across diverse indoor and outdoor environments, without requiring environment-specific tuning or explicit mapping.
♻ ☆ Asymptotically-Optimal Multi-Query Path Planning for a Polygonal Robot ICRA 2025
Shortest-path roadmaps, also known as reduced visibility graphs, provides a highly efficient multi-query method for computing optimal paths in two-dimensional environments. Combined with Minkowski sum computations, shortest-path roadmaps can compute optimal paths for a translating robot in 2D. In this study, we explore the intuitive idea of stacking up a set of reduced visibility graphs at different orientations for a polygonal holonomic robot to support the fast computation of near-optimal paths, allowing simultaneous 2D translation and rotation. The resulting algorithm, rotation-stacked visibility graph (RVG), is shown to be resolution-complete and asymptotically optimal. Extensive computational experiments show RVG significantly outperforms state-of-the-art single- and multi-query sampling-based methods on both computation time and solution optimality fronts.
comment: ICRA 2025
♻ ☆ Perception of Emotions in Human and Robot Faces: Is the Eye Region Enough?
The increased interest in developing next-gen social robots has raised questions about the factors affecting the perception of robot emotions. This study investigates the impact of robot appearances (humanlike, mechanical) and face regions (full-face, eye-region) on human perception of robot emotions. A between-subjects user study (N = 305) was conducted where participants were asked to identify the emotions being displayed in videos of robot faces, as well as a human baseline. Our findings reveal three important insights for effective social robot face design in Human-Robot Interaction (HRI): Firstly, robots equipped with a back-projected, fully animated face - regardless of whether they are more human-like or more mechanical-looking - demonstrate a capacity for emotional expression comparable to that of humans. Secondly, the recognition accuracy of emotional expressions in both humans and robots declines when only the eye region is visible. Lastly, within the constraint of only the eye region being visible, robots with more human-like features significantly enhance emotion recognition.
comment: Accepted for publication at the 16th International Conference on Social Robotics, Odense, Denmark (ICSR 2024)
♻ ☆ Autoregressive Action Sequence Learning for Robotic Manipulation
Designing a universal policy architecture that performs well across diverse robots and task configurations remains a key challenge. In this work, we address this by representing robot actions as sequential data and generating actions through autoregressive sequence modeling. Existing autoregressive architectures generate end-effector waypoints sequentially as word tokens in language modeling, which are limited to low-frequency control tasks. Unlike language, robot actions are heterogeneous and often include continuous values -- such as joint positions, 2D pixel coordinates, and end-effector poses -- which are not easily suited for language-based modeling. Based on this insight, we introduce a straightforward enhancement: we extend causal transformers' single-token prediction to support predicting a variable number of tokens in a single step through our Chunking Causal Transformer (CCT). This enhancement enables robust performance across diverse tasks of various control frequencies, greater efficiency by having fewer autoregression steps, and lead to a hybrid action sequence design by mixing different types of actions and using a different chunk size for each action type. Based on CCT, we propose the Autoregressive Policy (ARP) architecture, which solves manipulation tasks by generating hybrid action sequences. We evaluate ARP across diverse robotic manipulation environments, including Push-T, ALOHA, and RLBench, and show that ARP, as a universal architecture, matches or outperforms the environment-specific state-of-the-art in all tested benchmarks, while being more efficient in computation and parameter sizes. Videos of our real robot demonstrations, all source code and the pretrained models of ARP can be found at http://github.com/mlzxy/arp.
comment: (RA-L 2025) Add a new figure to explain why chunking autoregression works. Put back the previous in-depth discussion for arxiv release
♻ ☆ POp-GS: Next Best View in 3D-Gaussian Splatting with P-Optimality
In this paper, we present a novel algorithm for quantifying uncertainty and information gained within 3D Gaussian Splatting (3D-GS) through P-Optimality. While 3D-GS has proven to be a useful world model with high-quality rasterizations, it does not natively quantify uncertainty or information, posing a challenge for real-world applications such as 3D-GS SLAM. We propose to quantify information gain in 3D-GS by reformulating the problem through the lens of optimal experimental design, which is a classical solution widely used in literature. By restructuring information quantification of 3D-GS through optimal experimental design, we arrive at multiple solutions, of which T-Optimality and D-Optimality perform the best quantitatively and qualitatively as measured on two popular datasets. Additionally, we propose a block diagonal covariance approximation which provides a measure of correlation at the expense of a greater computation cost.
♻ ☆ UAVs Meet LLMs: Overviews and Perspectives Toward Agentic Low-Altitude Mobility
Low-altitude mobility, exemplified by unmanned aerial vehicles (UAVs), has introduced transformative advancements across various domains, like transportation, logistics, and agriculture. Leveraging flexible perspectives and rapid maneuverability, UAVs extend traditional systems' perception and action capabilities, garnering widespread attention from academia and industry. However, current UAV operations primarily depend on human control, with only limited autonomy in simple scenarios, and lack the intelligence and adaptability needed for more complex environments and tasks. The emergence of large language models (LLMs) demonstrates remarkable problem-solving and generalization capabilities, offering a promising pathway for advancing UAV intelligence. This paper explores the integration of LLMs and UAVs, beginning with an overview of UAV systems' fundamental components and functionalities, followed by an overview of the state-of-the-art in LLM technology. Subsequently, it systematically highlights the multimodal data resources available for UAVs, which provide critical support for training and evaluation. Furthermore, it categorizes and analyzes key tasks and application scenarios where UAVs and LLMs converge. Finally, a reference roadmap towards agentic UAVs is proposed, aiming to enable UAVs to achieve agentic intelligence through autonomous perception, memory, reasoning, and tool utilization. Related resources are available at https://github.com/Hub-Tian/UAVs_Meet_LLMs.
♻ ☆ Hierarchical Performance-Based Design Optimization Framework for Soft Grippers
This paper presents a hierarchical, performance-based framework for the design optimization of multi-fingered soft grippers. To address the need for systematically defined performance indices, the framework structures the optimization process into three integrated layers: Task Space, Motion Space, and Design Space. In the Task Space, performance indices are defined as core objectives, while the Motion Space interprets these into specific movement primitives. Finally, the Design Space applies parametric and topological optimization techniques to refine the geometry and material distribution of the system, achieving a balanced design across key performance metrics. The framework's layered structure enhances SG design, ensuring balanced performance and scalability for complex tasks and contributing to broader advancements in soft robotics.
comment: 7 pages, 3 figures, 1 Algorithm
♻ ☆ Helvipad: A Real-World Dataset for Omnidirectional Stereo Depth Estimation CVPR 2025
Despite progress in stereo depth estimation, omnidirectional imaging remains underexplored, mainly due to the lack of appropriate data. We introduce Helvipad, a real-world dataset for omnidirectional stereo depth estimation, featuring 40K video frames from video sequences across diverse environments, including crowded indoor and outdoor scenes with various lighting conditions. Collected using two 360{\deg} cameras in a top-bottom setup and a LiDAR sensor, the dataset includes accurate depth and disparity labels by projecting 3D point clouds onto equirectangular images. Additionally, we provide an augmented training set with an increased label density by using depth completion. We benchmark leading stereo depth estimation models for both standard and omnidirectional images. The results show that while recent stereo methods perform decently, a challenge persists in accurately estimating depth in omnidirectional imaging. To address this, we introduce necessary adaptations to stereo models, leading to improved performance.
comment: Accepted to CVPR 2025. Project page: https://vita-epfl.github.io/Helvipad
♻ ☆ Imitation Learning with Limited Actions via Diffusion Planners and Deep Koopman Controllers ICRA
Recent advances in diffusion-based robot policies have demonstrated significant potential in imitating multi-modal behaviors. However, these approaches typically require large quantities of demonstration data paired with corresponding robot action labels, creating a substantial data collection burden. In this work, we propose a plan-then-control framework aimed at improving the action-data efficiency of inverse dynamics controllers by leveraging observational demonstration data. Specifically, we adopt a Deep Koopman Operator framework to model the dynamical system and utilize observation-only trajectories to learn a latent action representation. This latent representation can then be effectively mapped to real high-dimensional continuous actions using a linear action decoder, requiring minimal action-labeled data. Through experiments on simulated robot manipulation tasks and a real robot experiment with multi-modal expert demonstrations, we demonstrate that our approach significantly enhances action-data efficiency and achieves high task success rates with limited action data.
comment: Accepted to IEEE International Conference on Robotics and Automation (ICRA) 2025
♻ ☆ On-Device Self-Supervised Learning of Low-Latency Monocular Depth from Only Events CVPR 2025
Event cameras provide low-latency perception for only milliwatts of power. This makes them highly suitable for resource-restricted, agile robots such as small flying drones. Self-supervised learning based on contrast maximization holds great potential for event-based robot vision, as it foregoes the need for high-frequency ground truth and allows for online learning in the robot's operational environment. However, online, on-board learning raises the major challenge of achieving sufficient computational efficiency for real-time learning, while maintaining competitive visual perception performance. In this work, we improve the time and memory efficiency of the contrast maximization pipeline, making on-device learning of low-latency monocular depth possible. We demonstrate that online learning on board a small drone yields more accurate depth estimates and more successful obstacle avoidance behavior compared to only pre-training. Benchmarking experiments show that the proposed pipeline is not only efficient, but also achieves state-of-the-art depth estimation performance among self-supervised approaches. Our work taps into the unused potential of online, on-device robot learning, promising smaller reality gaps and better performance.
comment: Accepted at CVPR 2025
♻ ☆ BimArt: A Unified Approach for the Synthesis of 3D Bimanual Interaction with Articulated Objects CVPR2025
We present BimArt, a novel generative approach for synthesizing 3D bimanual hand interactions with articulated objects. Unlike prior works, we do not rely on a reference grasp, a coarse hand trajectory, or separate modes for grasping and articulating. To achieve this, we first generate distance-based contact maps conditioned on the object trajectory with an articulation-aware feature representation, revealing rich bimanual patterns for manipulation. The learned contact prior is then used to guide our hand motion generator, producing diverse and realistic bimanual motions for object movement and articulation. Our work offers key insights into feature representation and contact prior for articulated objects, demonstrating their effectiveness in taming the complex, high-dimensional space of bimanual hand-object interactions. Through comprehensive quantitative experiments, we demonstrate a clear step towards simplified and high-quality hand-object animations that surpass the state of the art in motion quality and diversity. Project page: https://vcai.mpi-inf.mpg.de/projects/bimart/.
comment: CVPR2025
♻ ☆ RoboMatrix: A Skill-centric Hierarchical Framework for Scalable Robot Task Planning and Execution in Open-World
Existing robot policies predominantly adopt the task-centric approach, requiring end-to-end task data collection. This results in limited generalization to new tasks and difficulties in pinpointing errors within long-horizon, multi-stage tasks. To address this, we propose RoboMatrix, a skill-centric hierarchical framework designed for scalable robot task planning and execution in open-world environments. RoboMatrix extracts general meta-skills from diverse complex tasks, enabling the completion of unseen tasks through skill composition. Its architecture consists of a high-level scheduling layer that utilizes large language models (LLMs) for task decomposition, an intermediate skill layer housing meta-skill models, and a low-level hardware layer for robot control. A key innovation of our work is the introduction of the first unified vision-language-action (VLA) model capable of seamlessly integrating both movement and manipulation within one model. This is achieved by combining vision and language prompts to generate discrete actions. Experimental results demonstrate that RoboMatrix achieves a 50% higher success rate than task-centric baselines when applied to unseen objects, scenes, and tasks. To advance open-world robotics research, we will open-source code, hardware designs, model weights, and datasets at https://github.com/WayneMao/RoboMatrix.
comment: 17 pages, 16 figures
♻ ☆ MCVO: A Generic Visual Odometry for Arbitrarily Arranged Multi-Cameras
Making multi-camera visual SLAM systems easier to set up and more robust to the environment is attractive for vision robots. Existing monocular and binocular vision SLAM systems have narrow sensing Field-of-View (FoV), resulting in degenerated accuracy and limited robustness in textureless environments. Thus multi-camera SLAM systems are gaining attention because they can provide redundancy with much wider FoV. However, the usual arbitrary placement and orientation of multiple cameras make the pose scale estimation and system updating challenging. To address these problems, we propose a robust visual odometry system for rigidly-bundled arbitrarily-arranged multi-cameras, namely MCVO, which can achieve metric-scale state estimation with high flexibility in the cameras' arrangement. Specifically, we first design a learning-based feature tracking framework to shift the pressure of CPU processing of multiple video streams to GPU. Then we initialize the odometry system with the metric-scale poses under the rigid constraints between moving cameras. Finally, we fuse the features of the multi-cameras in the back-end to achieve robust pose estimation and online scale optimization. Additionally, multi-camera features help improve the loop detection for pose graph optimization. Experiments on KITTI-360 and MultiCamData datasets validate its robustness over arbitrarily arranged cameras. Compared with other stereo and multi-camera visual SLAM systems, our method obtains higher pose accuracy with better generalization ability. Our codes and online demos are available at https://github.com/JunhaoWang615/MCVO
comment: 8 pages, 8 figures
♻ ☆ RoboBrain: A Unified Brain Model for Robotic Manipulation from Abstract to Concrete
Recent advancements in Multimodal Large Language Models (MLLMs) have shown remarkable capabilities across various multimodal contexts. However, their application in robotic scenarios, particularly for long-horizon manipulation tasks, reveals significant limitations. These limitations arise from the current MLLMs lacking three essential robotic brain capabilities: Planning Capability, which involves decomposing complex manipulation instructions into manageable sub-tasks; Affordance Perception, the ability to recognize and interpret the affordances of interactive objects; and Trajectory Prediction, the foresight to anticipate the complete manipulation trajectory necessary for successful execution. To enhance the robotic brain's core capabilities from abstract to concrete, we introduce ShareRobot, a high-quality heterogeneous dataset that labels multi-dimensional information such as task planning, object affordance, and end-effector trajectory. ShareRobot's diversity and accuracy have been meticulously refined by three human annotators. Building on this dataset, we developed RoboBrain, an MLLM-based model that combines robotic and general multi-modal data, utilizes a multi-stage training strategy, and incorporates long videos and high-resolution images to improve its robotic manipulation capabilities. Extensive experiments demonstrate that RoboBrain achieves state-of-the-art performance across various robotic tasks, highlighting its potential to advance robotic brain capabilities.
♻ ☆ AutoURDF: Unsupervised Robot Modeling from Point Cloud Frames Using Cluster Registration
Robot description models are essential for simulation and control, yet their creation often requires significant manual effort. To streamline this modeling process, we introduce AutoURDF, an unsupervised approach for constructing description files for unseen robots from point cloud frames. Our method leverages a cluster-based point cloud registration model that tracks the 6-DoF transformations of point clusters. Through analyzing cluster movements, we hierarchically address the following challenges: (1) moving part segmentation, (2) body topology inference, and (3) joint parameter estimation. The complete pipeline produces robot description files that are fully compatible with existing simulators. We validate our method across a variety of robots, using both synthetic and real-world scan data. Results indicate that our approach outperforms previous methods in registration and body topology estimation accuracy, offering a scalable solution for automated robot modeling.
comment: 16 pages
♻ ☆ SyncDiff: Synchronized Motion Diffusion for Multi-Body Human-Object Interaction Synthesis
Synthesizing realistic human-object interaction motions is a critical problem in VR/AR and human animation. Unlike the commonly studied scenarios involving a single human or hand interacting with one object, we address a more generic multi-body setting with arbitrary numbers of humans, hands, and objects. This complexity introduces significant challenges in synchronizing motions due to the high correlations and mutual influences among bodies. To address these challenges, we introduce SyncDiff, a novel method for multi-body interaction synthesis using a synchronized motion diffusion strategy. SyncDiff employs a single diffusion model to capture the joint distribution of multi-body motions. To enhance motion fidelity, we propose a frequency-domain motion decomposition scheme. Additionally, we introduce a new set of alignment scores to emphasize the synchronization of different body motions. SyncDiff jointly optimizes both data sample likelihood and alignment likelihood through an explicit synchronization strategy. Extensive experiments across four datasets with various multi-body configurations demonstrate the superiority of SyncDiff over existing state-of-the-art motion synthesis methods.
Artificial Intelligence 184
☆ Zero-Shot Human-Object Interaction Synthesis with Multimodal Priors
Human-object interaction (HOI) synthesis is important for various applications, ranging from virtual reality to robotics. However, acquiring 3D HOI data is challenging due to its complexity and high cost, limiting existing methods to the narrow diversity of object types and interaction patterns in training datasets. This paper proposes a novel zero-shot HOI synthesis framework without relying on end-to-end training on currently limited 3D HOI datasets. The core idea of our method lies in leveraging extensive HOI knowledge from pre-trained Multimodal Models. Given a text description, our system first obtains temporally consistent 2D HOI image sequences using image or video generation models, which are then uplifted to 3D HOI milestones of human and object poses. We employ pre-trained human pose estimation models to extract human poses and introduce a generalizable category-level 6-DoF estimation method to obtain the object poses from 2D HOI images. Our estimation method is adaptive to various object templates obtained from text-to-3D models or online retrieval. A physics-based tracking of the 3D HOI kinematic milestone is further applied to refine both body motions and object poses, yielding more physically plausible HOI generation results. The experimental results demonstrate that our method is capable of generating open-vocabulary HOIs with physical realism and semantic diversity.
☆ Efficient Model Development through Fine-tuning Transfer
Modern LLMs struggle with efficient updates, as each new pretrained model version requires repeating expensive alignment processes. This challenge also applies to domain- or language-specific models, where fine-tuning on specialized data must be redone for every new base model release. In this paper, we explore the transfer of fine-tuning updates between model versions. Specifically, we derive the diff vector from one source model version, which represents the weight changes from fine-tuning, and apply it to the base model of a different target version. Through empirical evaluations on various open-weight model versions, we show that transferring diff vectors can significantly improve the target base model, often achieving performance comparable to its fine-tuned counterpart. For example, reusing the fine-tuning updates from Llama 3.0 8B leads to an absolute accuracy improvement of 10.7% on GPQA over the base Llama 3.1 8B without additional training, surpassing Llama 3.1 8B Instruct. In a multilingual model development setting, we show that this approach can significantly increase performance on target-language tasks without retraining, achieving an absolute improvement of 4.7% and 15.5% on Global MMLU for Malagasy and Turkish, respectively, compared to Llama 3.1 8B Instruct. Our controlled experiments reveal that fine-tuning transfer is most effective when the source and target models are linearly connected in the parameter space. Additionally, we demonstrate that fine-tuning transfer offers a stronger and more computationally efficient starting point for further fine-tuning. Finally, we propose an iterative recycling-then-finetuning approach for continuous model development, which improves both efficiency and effectiveness. Our findings suggest that fine-tuning transfer is a viable strategy to reduce training costs while maintaining model performance.
comment: 21 pages, 4 figures, 13 tables
☆ Direct Post-Training Preference Alignment for Multi-Agent Motion Generation Models Using Implicit Feedback from Pre-training Demonstrations ICLR 2025
Recent advancements in LLMs have revolutionized motion generation models in embodied applications. While LLM-type auto-regressive motion generation models benefit from training scalability, there remains a discrepancy between their token prediction objectives and human preferences. As a result, models pre-trained solely with token-prediction objectives often generate behaviors that deviate from what humans would prefer, making post-training preference alignment crucial for producing human-preferred motions. Unfortunately, post-training alignment requires extensive preference rankings of motions generated by the pre-trained model, which are costly to annotate, especially in multi-agent settings. Recently, there has been growing interest in leveraging pre-training demonstrations to scalably generate preference data for post-training alignment. However, these methods often adopt an adversarial assumption, treating all pre-trained model-generated samples as unpreferred examples. This adversarial approach overlooks the valuable signal provided by preference rankings among the model's own generations, ultimately reducing alignment effectiveness and potentially leading to misaligned behaviors. In this work, instead of treating all generated samples as equally bad, we leverage implicit preferences encoded in pre-training demonstrations to construct preference rankings among the pre-trained model's generations, offering more nuanced preference alignment guidance with zero human cost. We apply our approach to large-scale traffic simulation and demonstrate its effectiveness in improving the realism of pre-trained model's generated behaviors, making a lightweight 1M motion generation model comparable to SOTA large imitation-based models by relying solely on implicit feedback from pre-training demonstrations, without additional post-training human preference annotations or high computational costs.
comment: ICLR 2025 Spotlight
☆ AI Identity, Empowerment, and Mindfulness in Mitigating Unethical AI Use
This study examines how AI identity influences psychological empowerment and unethical AI behavior among college students, while also exploring the moderating role of IT mindfulness. Findings show that a strong AI identity enhances psychological empowerment and academic engagement but can also lead to increased unethical AI practices. Crucially, IT mindfulness acts as an ethical safeguard, promoting sensitivity to ethical concerns and reducing misuse of AI. These insights have implications for educators, policymakers, and AI developers, emphasizing For Peer Review the need for a balanced approach that encourages digital engagement without compromising student responsibility. The study also contributes to philosophical discussions of psychological agency, suggesting that empowerment through AI can yield both positive and negative outcomes. Mindfulness emerges as essential in guiding ethical AI interactions. Overall, the research informs ongoing debates on ethics in education and AI, offering strategies to align technological advancement with ethical accountability and responsible use.
☆ Can Multi-modal (reasoning) LLMs work as deepfake detectors?
Deepfake detection remains a critical challenge in the era of advanced generative models, particularly as synthetic media becomes more sophisticated. In this study, we explore the potential of state of the art multi-modal (reasoning) large language models (LLMs) for deepfake image detection such as (OpenAI O1/4o, Gemini thinking Flash 2, Deepseek Janus, Grok 3, llama 3.2, Qwen 2/2.5 VL, Mistral Pixtral, Claude 3.5/3.7 sonnet) . We benchmark 12 latest multi-modal LLMs against traditional deepfake detection methods across multiple datasets, including recently published real-world deepfake imagery. To enhance performance, we employ prompt tuning and conduct an in-depth analysis of the models' reasoning pathways to identify key contributing factors in their decision-making process. Our findings indicate that best multi-modal LLMs achieve competitive performance with promising generalization ability with zero shot, even surpass traditional deepfake detection pipelines in out-of-distribution datasets while the rest of the LLM families performs extremely disappointing with some worse than random guess. Furthermore, we found newer model version and reasoning capabilities does not contribute to performance in such niche tasks of deepfake detection while model size do help in some cases. This study highlights the potential of integrating multi-modal reasoning in future deepfake detection frameworks and provides insights into model interpretability for robustness in real-world scenarios.
☆ Abstracting Geo-specific Terrains to Scale Up Reinforcement Learning
Multi-agent reinforcement learning (MARL) is increasingly ubiquitous in training dynamic and adaptive synthetic characters for interactive simulations on geo-specific terrains. Frameworks such as Unity's ML-Agents help to make such reinforcement learning experiments more accessible to the simulation community. Military training simulations also benefit from advances in MARL, but they have immense computational requirements due to their complex, continuous, stochastic, partially observable, non-stationary, and doctrine-based nature. Furthermore, these simulations require geo-specific terrains, further exacerbating the computational resources problem. In our research, we leverage Unity's waypoints to automatically generate multi-layered representation abstractions of the geo-specific terrains to scale up reinforcement learning while still allowing the transfer of learned policies between different representations. Our early exploratory results on a novel MARL scenario, where each side has differing objectives, indicate that waypoint-based navigation enables faster and more efficient learning while producing trajectories similar to those taken by expert human players in CSGO gaming environments. This research points out the potential of waypoint-based navigation for reducing the computational costs of developing and training MARL models for military training simulations, where geo-specific terrains and differing objectives are crucial.
comment: 10 pages, 6 figures, 2024 Interservice/Industry Training, Simulation, and Education Conference (I/ITSEC)
☆ Adaptive Orchestration for Large-Scale Inference on Heterogeneous Accelerator Systems Balancing Cost, Performance, and Resilience
The surge in generative AI workloads has created a need for scalable inference systems that can flexibly harness both GPUs and specialized accelerators while containing operational costs. This paper proposes a hardware-agnostic control loop that adaptively allocates requests across heterogeneous accelerators based on real-time cost and capacity signals. The approach sustains low latency and high throughput by dynamically shifting between cost-optimized and capacity-optimized modes, ensuring the most efficient use of expensive compute resources under fluctuating availability. Evaluated using the Stable Diffusion model, the framework consistently meets latency targets, automatically redirects traffic during capacity shortfalls, and capitalizes on lower-cost accelerators when possible. These results highlight how a feedback-driven deployment strategy, spanning the entire software and hardware stack, can help organizations efficiently scale generative AI workloads while maintaining resilience in the face of limited accelerator capacity.
comment: 14 pages, 7 figures
☆ BugCraft: End-to-End Crash Bug Reproduction Using LLM Agents in Minecraft
Reproducing game bugs, in our case crash bugs in continuously evolving games like Minecraft, is a notoriously manual, time-consuming, and challenging process to automate. Despite the success of LLM-driven bug reproduction in other software domains, games, with their complex interactive environments, remain largely unaddressed. This paper introduces BugCraft, a novel end-to-end framework designed to automate the reproduction of crash bugs in Minecraft directly from user-submitted bug reports, addressing the critical gap in automated game bug reproduction. BugCraft employs a two-stage approach: first, a Step Synthesizer leverages LLMs and Minecraft Wiki knowledge to transform bug reports into high-quality, structured steps to reproduce (S2R). Second, an Action Model, powered by a vision-based LLM agent (GPT-4o) and a custom macro API, executes these S2R steps within Minecraft to trigger the reported crash. To facilitate evaluation, we introduce BugCraft-Bench, a curated dataset of Minecraft crash bug reports. Evaluated on BugCraft-Bench, our framework successfully reproduced 30.23% of crash bugs end-to-end. The Step Synthesizer demonstrated a 66.28% accuracy in generating correct bug reproduction plans, highlighting its effectiveness in interpreting and structuring bug report information. BugCraft demonstrates the feasibility of automated reproduction of crash bugs in complex game environments using LLMs, opening promising avenues for game testing and development. The framework and the BugCraft-Bench dataset pave the way for future research in automated game bug analysis and hold potential for generalization to other interactive game platforms. Finally, we make our code open at https://bugcraft2025.github.io/
☆ OmniNova:A General Multimodal Agent Framework
The integration of Large Language Models (LLMs) with specialized tools presents new opportunities for intelligent automation systems. However, orchestrating multiple LLM-driven agents to tackle complex tasks remains challenging due to coordination difficulties, inefficient resource utilization, and inconsistent information flow. We present OmniNova, a modular multi-agent automation framework that combines language models with specialized tools such as web search, crawling, and code execution capabilities. OmniNova introduces three key innovations: (1) a hierarchical multi-agent architecture with distinct coordinator, planner, supervisor, and specialist agents; (2) a dynamic task routing mechanism that optimizes agent deployment based on task complexity; and (3) a multi-layered LLM integration system that allocates appropriate models to different cognitive requirements. Our evaluations across 50 complex tasks in research, data analysis, and web interaction domains demonstrate that OmniNova outperforms existing frameworks in task completion rate (87\% vs. baseline 62\%), efficiency (41\% reduced token usage), and result quality (human evaluation score of 4.2/5 vs. baseline 3.1/5). We contribute both a theoretical framework for multi-agent system design and an open-source implementation that advances the state-of-the-art in LLM-based automation systems.
☆ Experience Replay Addresses Loss of Plasticity in Continual Learning
Loss of plasticity is one of the main challenges in continual learning with deep neural networks, where neural networks trained via backpropagation gradually lose their ability to adapt to new tasks and perform significantly worse than their freshly initialized counterparts. The main contribution of this paper is to propose a new hypothesis that experience replay addresses the loss of plasticity in continual learning. Here, experience replay is a form of memory. We provide supporting evidence for this hypothesis. In particular, we demonstrate in multiple different tasks, including regression, classification, and policy evaluation, that by simply adding an experience replay and processing the data in the experience replay with Transformers, the loss of plasticity disappears. Notably, we do not alter any standard components of deep learning. For example, we do not change backpropagation. We do not modify the activation functions. And we do not use any regularization. We conjecture that experience replay and Transformers can address the loss of plasticity because of the in-context learning phenomenon.
comment: 14 pages, 4 figures
☆ Unsupervised Learning for Quadratic Assignment
We introduce PLUME search, a data-driven framework that enhances search efficiency in combinatorial optimization through unsupervised learning. Unlike supervised or reinforcement learning, PLUME search learns directly from problem instances using a permutation-based loss with a non-autoregressive approach. We evaluate its performance on the quadratic assignment problem, a fundamental NP-hard problem that encompasses various combinatorial optimization problems. Experimental results demonstrate that PLUME search consistently improves solution quality. Furthermore, we study the generalization behavior and show that the learned model generalizes across different densities and sizes.
comment: preprint
☆ LEGO-Puzzles: How Good Are MLLMs at Multi-Step Spatial Reasoning?
Multi-step spatial reasoning entails understanding and reasoning about spatial relationships across multiple sequential steps, which is crucial for tackling complex real-world applications, such as robotic manipulation, autonomous navigation, and automated assembly. To assess how well current Multimodal Large Language Models (MLLMs) have acquired this fundamental capability, we introduce \textbf{LEGO-Puzzles}, a scalable benchmark designed to evaluate both \textbf{spatial understanding} and \textbf{sequential reasoning} in MLLMs through LEGO-based tasks. LEGO-Puzzles consists of 1,100 carefully curated visual question-answering (VQA) samples spanning 11 distinct tasks, ranging from basic spatial understanding to complex multi-step reasoning. Based on LEGO-Puzzles, we conduct a comprehensive evaluation of state-of-the-art MLLMs and uncover significant limitations in their spatial reasoning capabilities: even the most powerful MLLMs can answer only about half of the test cases, whereas human participants achieve over 90\% accuracy. In addition to VQA tasks, we evaluate MLLMs' abilities to generate LEGO images following assembly illustrations. Our experiments show that only Gemini-2.0-Flash and GPT-4o exhibit a limited ability to follow these instructions, while other MLLMs either replicate the input image or generate completely irrelevant outputs. Overall, LEGO-Puzzles exposes critical deficiencies in existing MLLMs' spatial understanding and sequential reasoning capabilities, and underscores the need for further advancements in multimodal spatial reasoning.
comment: 12 pages, 7 figures
☆ ExCoT: Optimizing Reasoning for Text-to-SQL with Execution Feedback
Text-to-SQL demands precise reasoning to convert natural language questions into structured queries. While large language models (LLMs) excel in many reasoning tasks, their ability to leverage Chain-of-Thought (CoT) reasoning for text-to-SQL remains underexplored. We identify critical limitations: zero-shot CoT offers minimal gains, and Direct Preference Optimization (DPO) applied without CoT yields marginal improvements. We propose ExCoT, a novel framework that iteratively optimizes open-source LLMs by combining CoT reasoning with off-policy and on-policy DPO, relying solely on execution accuracy as feedback. This approach eliminates the need for reward models or human-annotated preferences. Our experimental results demonstrate significant performance gains: ExCoT improves execution accuracy on BIRD dev set from 57.37% to 68.51% and on Spider test set from 78.81% to 86.59% for LLaMA-3 70B, with Qwen-2.5-Coder demonstrating similar improvements. Our best model achieves state-of-the-art performance in the single-model setting on both BIRD and Spider datasets, notably achieving 68.53% on the BIRD test set.
☆ CAFe: Unifying Representation and Generation with Contrastive-Autoregressive Finetuning
The rapid advancement of large vision-language models (LVLMs) has driven significant progress in multimodal tasks, enabling models to interpret, reason, and generate outputs across both visual and textual domains. While excelling in generative tasks, existing LVLMs often face limitations in tasks requiring high-fidelity representation learning, such as generating image or text embeddings for retrieval. Recent work has proposed finetuning LVLMs for representational learning, but the fine-tuned model often loses its generative capabilities due to the representational learning training paradigm. To address this trade-off, we introduce CAFe, a contrastive-autoregressive fine-tuning framework that enhances LVLMs for both representation and generative tasks. By integrating a contrastive objective with autoregressive language modeling, our approach unifies these traditionally separate tasks, achieving state-of-the-art results in both multimodal retrieval and multimodal generative benchmarks, including object hallucination (OH) mitigation. CAFe establishes a novel framework that synergizes embedding and generative functionalities in a single model, setting a foundation for future multimodal models that excel in both retrieval precision and coherent output generation.
☆ A proposal for an incident regime that tracks and counters threats to national security posed by AI systems
Recent progress in AI capabilities has heightened concerns that AI systems could pose a threat to national security, for example, by making it easier for malicious actors to perform cyberattacks on critical national infrastructure, or through loss of control of autonomous AI systems. In parallel, federal legislators in the US have proposed nascent 'AI incident regimes' to identify and counter similar threats. In this paper, we consolidate these two trends and present a proposal for a legally mandated post-deployment AI incident regie that aims to counter potential national security threats from AI systems. We start the paper by introducing the concept of 'security-critical' to describe doctors that pose extreme risks to national security, before arguing that 'security-critical' describes civilian nuclear power, aviation, life science dual-use research of concern, and frontier AI development. We then present in detail our AI incident regime proposal,, justifying each component of the proposal by demonstrating its similarity to US domestic incident regimes in other 'security-critical' sectors. Finally, we sketch a hypothetical scenario where our proposed AI incident regime deals with an AI cyber incident. Our proposed AI incident regime is split into three phases. The first phase revolves around a novel operationalization of what counts as an 'AI incident' and we suggest that AI providers must create a 'national security case' before deploying a frontier AI system. The second and third phases spell out that AI providers should notify a government agency about incidents, and that the government agency should be involved in amending AI providers' security and safety procedures, in order to counter future threats to national security. Our proposal is timely, given ongoing policy interest in the potential national security threats posed by AI systems.
☆ Dynamics of Structured Complex-Valued Hopfield Neural Networks
In this paper, we explore the dynamics of structured complex-valued Hopfield neural networks (CvHNNs), which arise when the synaptic weight matrix possesses specific structural properties. We begin by analyzing CvHNNs with a Hermitian synaptic weight matrix and establish the existence of four-cycle dynamics in CvHNNs with skew-Hermitian weight matrices operating synchronously. Furthermore, we introduce two new classes of complex-valued matrices: braided Hermitian and braided skew-Hermitian matrices. We demonstrate that CvHNNs utilizing these matrix types exhibit cycles of length eight when operating in full parallel update mode. Finally, we conduct extensive computational experiments on synchronous CvHNNs, exploring other synaptic weight matrix structures. The findings provide a comprehensive overview of the dynamics of structured CvHNNs, offering insights that may contribute to developing improved associative memory models when integrated with suitable learning rules.
☆ Geometric Meta-Learning via Coupled Ricci Flow: Unifying Knowledge Representation and Quantum Entanglement
This paper establishes a unified framework integrating geometric flows with deep learning through three fundamental innovations. First, we propose a thermodynamically coupled Ricci flow that dynamically adapts parameter space geometry to loss landscape topology, formally proved to preserve isometric knowledge embedding (Theorem~\ref{thm:isometric}). Second, we derive explicit phase transition thresholds and critical learning rates (Theorem~\ref{thm:critical}) through curvature blowup analysis, enabling automated singularity resolution via geometric surgery (Lemma~\ref{lem:surgery}). Third, we establish an AdS/CFT-type holographic duality (Theorem~\ref{thm:ads}) between neural networks and conformal field theories, providing entanglement entropy bounds for regularization design. Experiments demonstrate 2.1$\times$ convergence acceleration and 63\% topological simplification while maintaining $\mathcal{O}(N\log N)$ complexity, outperforming Riemannian baselines by 15.2\% in few-shot accuracy. Theoretically, we prove exponential stability (Theorem~\ref{thm:converge}) through a new Lyapunov function combining Perelman entropy with Wasserstein gradient flows, fundamentally advancing geometric deep learning.
comment: 9 pages, submitted to IEEE PAMI
☆ GENIUS: A Generative Framework for Universal Multimodal Search CVPR 2025
Generative retrieval is an emerging approach in information retrieval that generates identifiers (IDs) of target data based on a query, providing an efficient alternative to traditional embedding-based retrieval methods. However, existing models are task-specific and fall short of embedding-based retrieval in performance. This paper proposes GENIUS, a universal generative retrieval framework supporting diverse tasks across multiple modalities and domains. At its core, GENIUS introduces modality-decoupled semantic quantization, transforming multimodal data into discrete IDs encoding both modality and semantics. Moreover, to enhance generalization, we propose a query augmentation that interpolates between a query and its target, allowing GENIUS to adapt to varied query forms. Evaluated on the M-BEIR benchmark, it surpasses prior generative methods by a clear margin. Unlike embedding-based retrieval, GENIUS consistently maintains high retrieval speed across database size, with competitive performance across multiple benchmarks. With additional re-ranking, GENIUS often achieves results close to those of embedding-based methods while preserving efficiency.
comment: Accepted to CVPR 2025
☆ Guarding against artificial intelligence--hallucinated citations: the case for full-text reference deposit
The tendency of generative artificial intelligence (AI) systems to "hallucinate" false information is well-known; AI-generated citations to non-existent sources have made their way into the reference lists of peer-reviewed publications. Here, I propose a solution to this problem, taking inspiration from the Transparency and Openness Promotion (TOP) data sharing guidelines, the clash of generative AI with the American judiciary, and the precedent set by submissions of prior art to the United States Patent and Trademark Office. Journals should require authors to submit the full text of each cited source along with their manuscripts, thereby preventing authors from citing any material whose full text they cannot produce. This solution requires limited additional work on the part of authors or editors while effectively immunizing journals against hallucinated references.
comment: 3 pages
☆ A Comparative Analysis of Word Segmentation, Part-of-Speech Tagging, and Named Entity Recognition for Historical Chinese Sources, 1900-1950 NAACL 2025
This paper compares large language models (LLMs) and traditional natural language processing (NLP) tools for performing word segmentation, part-of-speech (POS) tagging, and named entity recognition (NER) on Chinese texts from 1900 to 1950. Historical Chinese documents pose challenges for text analysis due to their logographic script, the absence of natural word boundaries, and significant linguistic changes. Using a sample dataset from the Shanghai Library Republican Journal corpus, traditional tools such as Jieba and spaCy are compared to LLMs, including GPT-4o, Claude 3.5, and the GLM series. The results show that LLMs outperform traditional methods in all metrics, albeit at considerably higher computational costs, highlighting a trade-off between accuracy and efficiency. Additionally, LLMs better handle genre-specific challenges such as poetry and temporal variations (i.e., pre-1920 versus post-1920 texts), demonstrating that their contextual learning capabilities can advance NLP approaches to historical texts by reducing the need for domain-specific training data.
comment: Accepted to NLP4DH 2025 at NAACL 2025
☆ GyralNet Subnetwork Partitioning via Differentiable Spectral Modularity Optimization
Understanding the structural and functional organization of the human brain requires a detailed examination of cortical folding patterns, among which the three-hinge gyrus (3HG) has been identified as a key structural landmark. GyralNet, a network representation of cortical folding, models 3HGs as nodes and gyral crests as edges, highlighting their role as critical hubs in cortico-cortical connectivity. However, existing methods for analyzing 3HGs face significant challenges, including the sub-voxel scale of 3HGs at typical neuroimaging resolutions, the computational complexity of establishing cross-subject correspondences, and the oversimplification of treating 3HGs as independent nodes without considering their community-level relationships. To address these limitations, we propose a fully differentiable subnetwork partitioning framework that employs a spectral modularity maximization optimization strategy to modularize the organization of 3HGs within GyralNet. By incorporating topological structural similarity and DTI-derived connectivity patterns as attribute features, our approach provides a biologically meaningful representation of cortical organization. Extensive experiments on the Human Connectome Project (HCP) dataset demonstrate that our method effectively partitions GyralNet at the individual level while preserving the community-level consistency of 3HGs across subjects, offering a robust foundation for understanding brain connectivity.
comment: 10 pages, 3 figures
☆ Bitstream Collisions in Neural Image Compression via Adversarial Perturbations
Neural image compression (NIC) has emerged as a promising alternative to classical compression techniques, offering improved compression ratios. Despite its progress towards standardization and practical deployment, there has been minimal exploration into it's robustness and security. This study reveals an unexpected vulnerability in NIC - bitstream collisions - where semantically different images produce identical compressed bitstreams. Utilizing a novel whitebox adversarial attack algorithm, this paper demonstrates that adding carefully crafted perturbations to semantically different images can cause their compressed bitstreams to collide exactly. The collision vulnerability poses a threat to the practical usability of NIC, particularly in security-critical applications. The cause of the collision is analyzed, and a simple yet effective mitigation method is presented.
☆ ACVUBench: Audio-Centric Video Understanding Benchmark
Audio often serves as an auxiliary modality in video understanding tasks of audio-visual large language models (LLMs), merely assisting in the comprehension of visual information. However, a thorough understanding of videos significantly depends on auditory information, as audio offers critical context, emotional cues, and semantic meaning that visual data alone often lacks. This paper proposes an audio-centric video understanding benchmark (ACVUBench) to evaluate the video comprehension capabilities of multimodal LLMs with a particular focus on auditory information. Specifically, ACVUBench incorporates 2,662 videos spanning 18 different domains with rich auditory information, together with over 13k high-quality human annotated or validated question-answer pairs. Moreover, ACVUBench introduces a suite of carefully designed audio-centric tasks, holistically testing the understanding of both audio content and audio-visual interactions in videos. A thorough evaluation across a diverse range of open-source and proprietary multimodal LLMs is performed, followed by the analyses of deficiencies in audio-visual LLMs. Demos are available at https://github.com/lark-png/ACVUBench.
☆ Thinking agents for zero-shot generalization to qualitatively novel tasks
Intelligent organisms can solve truly novel problems which they have never encountered before, either in their lifetime or their evolution. An important component of this capacity is the ability to ``think'', that is, to mentally manipulate objects, concepts and behaviors in order to plan and evaluate possible solutions to novel problems, even without environment interaction. To generate problems that are truly qualitatively novel, while still solvable zero-shot (by mental simulation), we use the combinatorial nature of environments: we train the agent while withholding a specific combination of the environment's elements. The novel test task, based on this combination, is thus guaranteed to be truly novel, while still mentally simulable since the agent has been exposed to each individual element (and their pairwise interactions) during training. We propose a method to train agents endowed with world models to make use their mental simulation abilities, by selecting tasks based on the difference between the agent's pre-thinking and post-thinking performance. When tested on the novel, withheld problem, the resulting agent successfully simulated alternative scenarios and used the resulting information to guide its behavior in the actual environment, solving the novel task in a single real-environment trial (zero-shot).
☆ Guidelines For The Choice Of The Baseline in XAI Attribution Methods
Given the broad adoption of artificial intelligence, it is essential to provide evidence that AI models are reliable, trustable, and fair. To this end, the emerging field of eXplainable AI develops techniques to probe such requirements, counterbalancing the hype pushing the pervasiveness of this technology. Among the many facets of this issue, this paper focuses on baseline attribution methods, aiming at deriving a feature attribution map at the network input relying on a "neutral" stimulus usually called "baseline". The choice of the baseline is crucial as it determines the explanation of the network behavior. In this framework, this paper has the twofold goal of shedding light on the implications of the choice of the baseline and providing a simple yet effective method for identifying the best baseline for the task. To achieve this, we propose a decision boundary sampling method, since the baseline, by definition, lies on the decision boundary, which naturally becomes the search domain. Experiments are performed on synthetic examples and validated relying on state-of-the-art methods. Despite being limited to the experimental scope, this contribution is relevant as it offers clear guidelines and a simple proxy for baseline selection, reducing ambiguity and enhancing deep models' reliability and trust.
☆ LogQuant: Log-Distributed 2-Bit Quantization of KV Cache with Superior Accuracy Preservation ICLR 2025
We introduce LogQuant, a groundbreaking 2-bit quantization technique for KV Cache in large language model (LLM) inference, delivering substantial memory savings while preserving superior performance. Previous methods either assume that later tokens are more important or attempt to predict important tokens based on earlier attention patterns. Both approaches, however, can result in performance bottlenecks or frequent mispredictions. LogQuant takes a different approach. By applying a log-based filtering mechanism, it selectively compresses the KV Cache across the entire context, achieving better performance with the same or even reduced memory footprint compared to existing methods. In benchmark tests, it enhances throughput by 25% and boosts batch size by 60% without increasing memory consumption. For challenging tasks such as Math and Code Completion, LogQuant improves accuracy by 40% to 200% at the same compression ratio, outperforming comparable techniques.LogQuant integrates effortlessly with popular inference frameworks like Python's transformers library. Implementation can be available in https://github.com/Concyclics/LogQuantKV.
comment: Accepted by ICLR 2025 Workshop on Sparsity in LLMs (SLLM)
☆ Simulating Tracking Data to Advance Sports Analytics Research AAMAS
Advanced analytics have transformed how sports teams operate, particularly in episodic sports like baseball. Their impact on continuous invasion sports, such as soccer and ice hockey, has been limited due to increased game complexity and restricted access to high-resolution game tracking data. In this demo, we present a method to collect and utilize simulated soccer tracking data from the Google Research Football environment to support the development of models designed for continuous tracking data. The data is stored in a schema that is representative of real tracking data and we provide processes that extract high-level features and events. We include examples of established tracking data models to showcase the efficacy of the simulated data. We address the scarcity of publicly available tracking data, providing support for research at the intersection of artificial intelligence and sports analytics.
comment: 2 pages, 2 figures, Proceedings of the 24th International Conference on Autonomous Agents and MultiAgent Systems (AAMAS)
☆ LENVIZ: A High-Resolution Low-Exposure Night Vision Benchmark Dataset
Low-light image enhancement is crucial for a myriad of applications, from night vision and surveillance, to autonomous driving. However, due to the inherent limitations that come in hand with capturing images in low-illumination environments, the task of enhancing such scenes still presents a formidable challenge. To advance research in this field, we introduce our Low Exposure Night Vision (LENVIZ) Dataset, a comprehensive multi-exposure benchmark dataset for low-light image enhancement comprising of over 230K frames showcasing 24K real-world indoor and outdoor, with-and without human, scenes. Captured using 3 different camera sensors, LENVIZ offers a wide range of lighting conditions, noise levels, and scene complexities, making it the largest publicly available up-to 4K resolution benchmark in the field. LENVIZ includes high quality human-generated ground truth, for which each multi-exposure low-light scene has been meticulously curated and edited by expert photographers to ensure optimal image quality. Furthermore, we also conduct a comprehensive analysis of current state-of-the-art low-light image enhancement techniques on our dataset and highlight potential areas of improvement.
comment: Dataset will be released upon publication
☆ SeLIP: Similarity Enhanced Contrastive Language Image Pretraining for Multi-modal Head MRI
Despite that deep learning (DL) methods have presented tremendous potential in many medical image analysis tasks, the practical applications of medical DL models are limited due to the lack of enough data samples with manual annotations. By noting that the clinical radiology examinations are associated with radiology reports that describe the images, we propose to develop a foundation model for multi-model head MRI by using contrastive learning on the images and the corresponding radiology findings. In particular, a contrastive learning framework is proposed, where a mixed syntax and semantic similarity matching metric is integrated to reduce the thirst of extreme large dataset in conventional contrastive learning framework. Our proposed similarity enhanced contrastive language image pretraining (SeLIP) is able to effectively extract more useful features. Experiments revealed that our proposed SeLIP performs well in many downstream tasks including image-text retrieval task, classification task, and image segmentation, which highlights the importance of considering the similarities among texts describing different images in developing medical image foundation models.
☆ PAVE: Patching and Adapting Video Large Language Models CVPR2025
Pre-trained video large language models (Video LLMs) exhibit remarkable reasoning capabilities, yet adapting these models to new tasks involving additional modalities or data types (e.g., audio or 3D information) remains challenging. In this paper, we present PAVE, a flexible framework for adapting pre-trained Video LLMs to downstream tasks with side-channel signals, such as audio, 3D cues, or multi-view videos. PAVE introduces lightweight adapters, referred to as "patches," which add a small number of parameters and operations to a base model without modifying its architecture or pre-trained weights. In doing so, PAVE can effectively adapt the pre-trained base model to support diverse downstream tasks, including audio-visual question answering, 3D reasoning, multi-view video recognition, and high frame rate video understanding. Across these tasks, PAVE significantly enhances the performance of the base model, surpassing state-of-the-art task-specific models while incurring a minor cost of ~0.1% additional FLOPs and parameters. Further, PAVE supports multi-task learning and generalizes well across different Video LLMs. Our code is available at https://github.com/dragonlzm/PAVE.
comment: CVPR2025 Camera Ready
☆ Gemma 3 Technical Report
We introduce Gemma 3, a multimodal addition to the Gemma family of lightweight open models, ranging in scale from 1 to 27 billion parameters. This version introduces vision understanding abilities, a wider coverage of languages and longer context - at least 128K tokens. We also change the architecture of the model to reduce the KV-cache memory that tends to explode with long context. This is achieved by increasing the ratio of local to global attention layers, and keeping the span on local attention short. The Gemma 3 models are trained with distillation and achieve superior performance to Gemma 2 for both pre-trained and instruction finetuned versions. In particular, our novel post-training recipe significantly improves the math, chat, instruction-following and multilingual abilities, making Gemma3-4B-IT competitive with Gemma2-27B-IT and Gemma3-27B-IT comparable to Gemini-1.5-Pro across benchmarks. We release all our models to the community.
☆ Test-Time Reasoning Through Visual Human Preferences with VLMs and Soft Rewards
Can Visual Language Models (VLMs) effectively capture human visual preferences? This work addresses this question by training VLMs to think about preferences at test time, employing reinforcement learning methods inspired by DeepSeek R1 and OpenAI O1. Using datasets such as ImageReward and Human Preference Score v2 (HPSv2), our models achieve accuracies of 64.9% on the ImageReward test set (trained on ImageReward official split) and 65.4% on HPSv2 (trained on approximately 25% of its data). These results match traditional encoder-based models while providing transparent reasoning and enhanced generalization. This approach allows to use not only rich VLM world knowledge, but also its potential to think, yielding interpretable outcomes that help decision-making processes. By demonstrating that human visual preferences reasonable by current VLMs, we introduce efficient soft-reward strategies for image ranking, outperforming simplistic selection or scoring methods. This reasoning capability enables VLMs to rank arbitrary images-regardless of aspect ratio or complexity-thereby potentially amplifying the effectiveness of visual Preference Optimization. By reducing the need for extensive markup while improving reward generalization and explainability, our findings can be a strong mile-stone that will enhance text-to-vision models even further.
☆ Splitting Answer Set Programs with respect to Intensionality Statements (Extended Version) AAAI 2023
Splitting a logic program allows us to reduce the task of computing its stable models to similar tasks for its subprograms. This can be used to increase solving performance and prove program correctness. We generalize the conditions under which this technique is applicable, by considering not only dependencies between predicates but also their arguments and context. This allows splitting programs commonly used in practice to which previous results were not applicable.
comment: Extended version of the paper published in AAAI 2023
☆ Vanishing Depth: A Depth Adapter with Positional Depth Encoding for Generalized Image Encoders
Generalized metric depth understanding is critical for precise vision-guided robotics, which current state-of-the-art (SOTA) vision-encoders do not support. To address this, we propose Vanishing Depth, a self-supervised training approach that extends pretrained RGB encoders to incorporate and align metric depth into their feature embeddings. Based on our novel positional depth encoding, we enable stable depth density and depth distribution invariant feature extraction. We achieve performance improvements and SOTA results across a spectrum of relevant RGBD downstream tasks - without the necessity of finetuning the encoder. Most notably, we achieve 56.05 mIoU on SUN-RGBD segmentation, 88.3 RMSE on Void's depth completion, and 83.8 Top 1 accuracy on NYUv2 scene classification. In 6D-object pose estimation, we outperform our predecessors of DinoV2, EVA-02, and Omnivore and achieve SOTA results for non-finetuned encoders in several related RGBD downstream tasks.
comment: Preprint
☆ Inducing Personality in LLM-Based Honeypot Agents: Measuring the Effect on Human-Like Agenda Generation
This paper presents SANDMAN, an architecture for cyber deception that leverages Language Agents to emulate convincing human simulacra. Our 'Deceptive Agents' serve as advanced cyber decoys, designed for high-fidelity engagement with attackers by extending the observation period of attack behaviours. Through experimentation, measurement, and analysis, we demonstrate how a prompt schema based on the five-factor model of personality systematically induces distinct 'personalities' in Large Language Models. Our results highlight the feasibility of persona-driven Language Agents for generating diverse, realistic behaviours, ultimately improving cyber deception strategies.
comment: 11 pages, 1 figure, 6 tables. Accepted to NLPAICS 2024
☆ On What Depends the Robustness of Multi-source Models to Missing Data in Earth Observation?
In recent years, the development of robust multi-source models has emerged in the Earth Observation (EO) field. These are models that leverage data from diverse sources to improve predictive accuracy when there is missing data. Despite these advancements, the factors influencing the varying effectiveness of such models remain poorly understood. In this study, we evaluate the predictive performance of six state-of-the-art multi-source models in predicting scenarios where either a single data source is missing or only a single source is available. Our analysis reveals that the efficacy of these models is intricately tied to the nature of the task, the complementarity among data sources, and the model design. Surprisingly, we observe instances where the removal of certain data sources leads to improved predictive performance, challenging the assumption that incorporating all available data is always beneficial. These findings prompt critical reflections on model complexity and the necessity of all collected data sources, potentially shaping the way for more streamlined approaches in EO applications.
comment: Accepted at IEEE International Geoscience and Remote Sensing Symposium 2025
☆ Invertible Koopman neural operator for data-driven modeling of partial differential equations
Koopman operator theory is a popular candidate for data-driven modeling because it provides a global linearization representation for nonlinear dynamical systems. However, existing Koopman operator-based methods suffer from shortcomings in constructing the well-behaved observable function and its inverse and are inefficient enough when dealing with partial differential equations (PDEs). To address these issues, this paper proposes the Invertible Koopman Neural Operator (IKNO), a novel data-driven modeling approach inspired by the Koopman operator theory and neural operator. IKNO leverages an Invertible Neural Network to parameterize observable function and its inverse simultaneously under the same learnable parameters, explicitly guaranteeing the reconstruction relation, thus eliminating the dependency on the reconstruction loss, which is an essential improvement over the original Koopman Neural Operator (KNO). The structured linear matrix inspired by the Koopman operator theory is parameterized to learn the evolution of observables' low-frequency modes in the frequency space rather than directly in the observable space, sustaining IKNO is resolution-invariant like other neural operators. Moreover, with preprocessing such as interpolation and dimension expansion, IKNO can be extended to operator learning tasks defined on non-Cartesian domains. We fully support the above claims based on rich numerical and real-world examples and demonstrate the effectiveness of IKNO and superiority over other neural operators.
comment: 25 pages, 10 figures
☆ Decoupled Dynamics Framework with Neural Fields for 3D Spatio-temporal Prediction of Vehicle Collisions
This study proposes a neural framework that predicts 3D vehicle collision dynamics by independently modeling global rigid-body motion and local structural deformation. Unlike approaches directly predicting absolute displacement, this method explicitly separates the vehicle's overall translation and rotation from its structural deformation. Two specialized networks form the core of the framework: a quaternion-based Rigid Net for rigid motion and a coordinate-based Deformation Net for local deformation. By independently handling fundamentally distinct physical phenomena, the proposed architecture achieves accurate predictions without requiring separate supervision for each component. The model, trained on only 10% of available simulation data, significantly outperforms baseline models, including single multi-layer perceptron (MLP) and deep operator networks (DeepONet), with prediction errors reduced by up to 83%. Extensive validation demonstrates strong generalization to collision conditions outside the training range, accurately predicting responses even under severe impacts involving extreme velocities and large impact angles. Furthermore, the framework successfully reconstructs high-resolution deformation details from low-resolution inputs without increased computational effort. Consequently, the proposed approach provides an effective, computationally efficient method for rapid and reliable assessment of vehicle safety across complex collision scenarios, substantially reducing the required simulation data and time while preserving prediction fidelity.
comment: 24 pages, 13 figures
☆ Writing as a testbed for open ended agents
Open-ended tasks are particularly challenging for LLMs due to the vast solution space, demanding both expansive exploration and adaptable strategies, especially when success lacks a clear, objective definition. Writing, with its vast solution space and subjective evaluation criteria, provides a compelling testbed for studying such problems. In this paper, we investigate the potential of LLMs to act as collaborative co-writers, capable of suggesting and implementing text improvements autonomously. We analyse three prominent LLMs - Gemini 1.5 Pro, Claude 3.5 Sonnet, and GPT-4o - focusing on how their action diversity, human alignment, and iterative improvement capabilities impact overall performance. This work establishes a framework for benchmarking autonomous writing agents and, more broadly, highlights fundamental challenges and potential solutions for building systems capable of excelling in diverse open-ended domains.
☆ Bootstrap Your Own Views: Masked Ego-Exo Modeling for Fine-grained View-invariant Video Representations CVPR 2025
View-invariant representation learning from egocentric (first-person, ego) and exocentric (third-person, exo) videos is a promising approach toward generalizing video understanding systems across multiple viewpoints. However, this area has been underexplored due to the substantial differences in perspective, motion patterns, and context between ego and exo views. In this paper, we propose a novel masked ego-exo modeling that promotes both causal temporal dynamics and cross-view alignment, called Bootstrap Your Own Views (BYOV), for fine-grained view-invariant video representation learning from unpaired ego-exo videos. We highlight the importance of capturing the compositional nature of human actions as a basis for robust cross-view understanding. Specifically, self-view masking and cross-view masking predictions are designed to learn view-invariant and powerful representations concurrently. Experimental results demonstrate that our BYOV significantly surpasses existing approaches with notable gains across all metrics in four downstream ego-exo video tasks. The code is available at https://github.com/park-jungin/byov.
comment: CVPR 2025 Camera-ready
☆ Optimal Path Planning and Cost Minimization for a Drone Delivery System Via Model Predictive Control
In this study, we formulate the drone delivery problem as a control problem and solve it using Model Predictive Control. Two experiments are performed: The first is on a less challenging grid world environment with lower dimensionality, and the second is with a higher dimensionality and added complexity. The MPC method was benchmarked against three popular Multi-Agent Reinforcement Learning (MARL): Independent $Q$-Learning (IQL), Joint Action Learners (JAL), and Value-Decomposition Networks (VDN). It was shown that the MPC method solved the problem quicker and required fewer optimal numbers of drones to achieve a minimized cost and navigate the optimal path.
comment: 15 pages, 5 figures, Submitted to the 2025 International Conference on Artificial Intelligence, Computer, Data Sciences and Applications
☆ Deep Learning for Speech Emotion Recognition: A CNN Approach Utilizing Mel Spectrograms
This paper explores the application of Convolutional Neural Networks CNNs for classifying emotions in speech through Mel Spectrogram representations of audio files. Traditional methods such as Gaussian Mixture Models and Hidden Markov Models have proven insufficient for practical deployment, prompting a shift towards deep learning techniques. By transforming audio data into a visual format, the CNN model autonomously learns to identify intricate patterns, enhancing classification accuracy. The developed model is integrated into a user-friendly graphical interface, facilitating realtime predictions and potential applications in educational environments. The study aims to advance the understanding of deep learning in speech emotion recognition, assess the models feasibility, and contribute to the integration of technology in learning contexts
comment: 5 pages 8 figures
☆ BiblioPage: A Dataset of Scanned Title Pages for Bibliographic Metadata Extraction ICDAR2025
Manual digitization of bibliographic metadata is time consuming and labor intensive, especially for historical and real-world archives with highly variable formatting across documents. Despite advances in machine learning, the absence of dedicated datasets for metadata extraction hinders automation. To address this gap, we introduce BiblioPage, a dataset of scanned title pages annotated with structured bibliographic metadata. The dataset consists of approximately 2,000 monograph title pages collected from 14 Czech libraries, spanning a wide range of publication periods, typographic styles, and layout structures. Each title page is annotated with 16 bibliographic attributes, including title, contributors, and publication metadata, along with precise positional information in the form of bounding boxes. To extract structured information from this dataset, we valuated object detection models such as YOLO and DETR combined with transformer-based OCR, achieving a maximum mAP of 52 and an F1 score of 59. Additionally, we assess the performance of various visual large language models, including LlamA 3.2-Vision and GPT-4o, with the best model reaching an F1 score of 67. BiblioPage serves as a real-world benchmark for bibliographic metadata extraction, contributing to document understanding, document question answering, and document information extraction. Dataset and evaluation scripts are availible at: https://github.com/DCGM/biblio-dataset
comment: Submitted to ICDAR2025 conference
☆ Towards Reliable Time Series Forecasting under Future Uncertainty: Ambiguity and Novelty Rejection Mechanisms
In real-world time series forecasting, uncertainty and lack of reliable evaluation pose significant challenges. Notably, forecasting errors often arise from underfitting in-distribution data and failing to handle out-of-distribution inputs. To enhance model reliability, we introduce a dual rejection mechanism combining ambiguity and novelty rejection. Ambiguity rejection, using prediction error variance, allows the model to abstain under low confidence, assessed through historical error variance analysis without future ground truth. Novelty rejection, employing Variational Autoencoders and Mahalanobis distance, detects deviations from training data. This dual approach improves forecasting reliability in dynamic environments by reducing errors and adapting to data changes, advancing reliability in complex scenarios.
☆ RGB-Th-Bench: A Dense benchmark for Visual-Thermal Understanding of Vision Language Models
We introduce RGB-Th-Bench, the first benchmark designed to evaluate the ability of Vision-Language Models (VLMs) to comprehend RGB-Thermal image pairs. While VLMs have demonstrated remarkable progress in visual reasoning and multimodal understanding, their evaluation has been predominantly limited to RGB-based benchmarks, leaving a critical gap in assessing their capabilities in infrared vision tasks. Existing visible-infrared datasets are either task-specific or lack high-quality annotations necessary for rigorous model evaluation. To address these limitations, RGB-Th-Bench provides a comprehensive evaluation framework covering 14 distinct skill dimensions, with a total of 1,600+ expert-annotated Yes/No questions. The benchmark employs two accuracy metrics: a standard question-level accuracy and a stricter skill-level accuracy, which evaluates model robustness across multiple questions within each skill dimension. This design ensures a thorough assessment of model performance, including resilience to adversarial and hallucinated responses. We conduct extensive evaluations on 19 state-of-the-art VLMs, revealing significant performance gaps in RGB-Thermal understanding. Our results show that even the strongest models struggle with thermal image comprehension, with performance heavily constrained by their RGB-based capabilities. Additionally, the lack of large-scale application-specific and expert-annotated thermal-caption-pair datasets in pre-training is an important reason of the observed performance gap. RGB-Th-Bench highlights the urgent need for further advancements in multimodal learning to bridge the gap between visible and thermal image understanding. The dataset is available through this link, and the evaluation code will also be made publicly available.
☆ OpenSDI: Spotting Diffusion-Generated Images in the Open World
This paper identifies OpenSDI, a challenge for spotting diffusion-generated images in open-world settings. In response to this challenge, we define a new benchmark, the OpenSDI dataset (OpenSDID), which stands out from existing datasets due to its diverse use of large vision-language models that simulate open-world diffusion-based manipulations. Another outstanding feature of OpenSDID is its inclusion of both detection and localization tasks for images manipulated globally and locally by diffusion models. To address the OpenSDI challenge, we propose a Synergizing Pretrained Models (SPM) scheme to build up a mixture of foundation models. This approach exploits a collaboration mechanism with multiple pretrained foundation models to enhance generalization in the OpenSDI context, moving beyond traditional training by synergizing multiple pretrained models through prompting and attending strategies. Building on this scheme, we introduce MaskCLIP, an SPM-based model that aligns Contrastive Language-Image Pre-Training (CLIP) with Masked Autoencoder (MAE). Extensive evaluations on OpenSDID show that MaskCLIP significantly outperforms current state-of-the-art methods for the OpenSDI challenge, achieving remarkable relative improvements of 14.23% in IoU (14.11% in F1) and 2.05% in accuracy (2.38% in F1) compared to the second-best model in localization and detection tasks, respectively. Our dataset and code are available at https://github.com/iamwangyabin/OpenSDI.
☆ HausaNLP at SemEval-2025 Task 3: Towards a Fine-Grained Model-Aware Hallucination Detection
This paper presents our findings of the Multilingual Shared Task on Hallucinations and Related Observable Overgeneration Mistakes, MU-SHROOM, which focuses on identifying hallucinations and related overgeneration errors in large language models (LLMs). The shared task involves detecting specific text spans that constitute hallucinations in the outputs generated by LLMs in 14 languages. To address this task, we aim to provide a nuanced, model-aware understanding of hallucination occurrences and severity in English. We used natural language inference and fine-tuned a ModernBERT model using a synthetic dataset of 400 samples, achieving an Intersection over Union (IoU) score of 0.032 and a correlation score of 0.422. These results indicate a moderately positive correlation between the model's confidence scores and the actual presence of hallucinations. The IoU score indicates that our model has a relatively low overlap between the predicted hallucination span and the truth annotation. The performance is unsurprising, given the intricate nature of hallucination detection. Hallucinations often manifest subtly, relying on context, making pinpointing their exact boundaries formidable.
☆ Recover from Horcrux: A Spectrogram Augmentation Method for Cardiac Feature Monitoring from Radar Signal Components
Radar-based wellness monitoring is becoming an effective measurement to provide accurate vital signs in a contactless manner, but data scarcity retards the related research on deep-learning-based methods. Data augmentation is commonly used to enrich the dataset by modifying the existing data, but most augmentation techniques can only couple with classification tasks. To enable the augmentation for regression tasks, this research proposes a spectrogram augmentation method, Horcrux, for radar-based cardiac feature monitoring (e.g., heartbeat detection, electrocardiogram reconstruction) with both classification and regression tasks involved. The proposed method is designed to increase the diversity of input samples while the augmented spectrogram is still faithful to the original ground truth vital sign. In addition, Horcrux proposes to inject zero values in specific areas to enhance the awareness of the deep learning model on subtle cardiac features, improving the performance for the limited dataset. Experimental result shows that Horcrux achieves an overall improvement of 16.20% in cardiac monitoring and has the potential to be extended to other spectrogram-based tasks. The code will be released upon publication.
☆ Show or Tell? Effectively prompting Vision-Language Models for semantic segmentation
Large Vision-Language Models (VLMs) are increasingly being regarded as foundation models that can be instructed to solve diverse tasks by prompting, without task-specific training. We examine the seemingly obvious question: how to effectively prompt VLMs for semantic segmentation. To that end, we systematically evaluate the segmentation performance of several recent models guided by either text or visual prompts on the out-of-distribution MESS dataset collection. We introduce a scalable prompting scheme, few-shot prompted semantic segmentation, inspired by open-vocabulary segmentation and few-shot learning. It turns out that VLMs lag far behind specialist models trained for a specific segmentation task, by about 30% on average on the Intersection-over-Union metric. Moreover, we find that text prompts and visual prompts are complementary: each one of the two modes fails on many examples that the other one can solve. Our analysis suggests that being able to anticipate the most effective prompt modality can lead to a 11% improvement in performance. Motivated by our findings, we propose PromptMatcher, a remarkably simple training-free baseline that combines both text and visual prompts, achieving state-of-the-art results outperforming the best text-prompted VLM by 2.5%, and the top visual-prompted VLM by 3.5% on few-shot prompted semantic segmentation.
☆ Analyzable Chain-of-Musical-Thought Prompting for High-Fidelity Music Generation
Autoregressive (AR) models have demonstrated impressive capabilities in generating high-fidelity music. However, the conventional next-token prediction paradigm in AR models does not align with the human creative process in music composition, potentially compromising the musicality of generated samples. To overcome this limitation, we introduce MusiCoT, a novel chain-of-thought (CoT) prompting technique tailored for music generation. MusiCoT empowers the AR model to first outline an overall music structure before generating audio tokens, thereby enhancing the coherence and creativity of the resulting compositions. By leveraging the contrastive language-audio pretraining (CLAP) model, we establish a chain of "musical thoughts", making MusiCoT scalable and independent of human-labeled data, in contrast to conventional CoT methods. Moreover, MusiCoT allows for in-depth analysis of music structure, such as instrumental arrangements, and supports music referencing -- accepting variable-length audio inputs as optional style references. This innovative approach effectively addresses copying issues, positioning MusiCoT as a vital practical method for music prompting. Our experimental results indicate that MusiCoT consistently achieves superior performance across both objective and subjective metrics, producing music quality that rivals state-of-the-art generation models. Our samples are available at https://MusiCoT.github.io/.
comment: Preprint
☆ Enabling Rapid Shared Human-AI Mental Model Alignment via the After-Action Review AAAI 2025
In this work, we present two novel contributions toward improving research in human-machine teaming (HMT): 1) a Minecraft testbed to accelerate testing and deployment of collaborative AI agents and 2) a tool to allow users to revisit and analyze behaviors within an HMT episode to facilitate shared mental model development. Our browser-based Minecraft testbed allows for rapid testing of collaborative agents in a continuous-space, real-time, partially-observable environment with real humans without cumbersome setup typical to human-AI interaction user studies. As Minecraft has an extensive player base and a rich ecosystem of pre-built AI agents, we hope this contribution can help to facilitate research quickly in the design of new collaborative agents and in understanding different human factors within HMT. Our mental model alignment tool facilitates user-led post-mission analysis by including video displays of first-person perspectives of the team members (i.e., the human and AI) that can be replayed, and a chat interface that leverages GPT-4 to provide answers to various queries regarding the AI's experiences and model details.
comment: Accepted to the Cooperative Multi-Agent Systems Decision-making and Learning:Human-Multi-Agent Cognitive Fusion Workshop at AAAI 2025
☆ Innate Reasoning is Not Enough: In-Context Learning Enhances Reasoning Large Language Models with Less Overthinking
Recent advances in Large Language Models (LLMs) have introduced Reasoning Large Language Models (RLLMs), which employ extended thinking processes with reflection and self-correction capabilities, demonstrating the effectiveness of test-time scaling. RLLMs exhibit innate Chain-of-Thought (CoT) reasoning capability obtained from training, leading to a natural question: "Is CoT prompting, a popular In-Context Learning (ICL) method for chat LLMs, necessary to enhance the reasoning capability of RLLMs?" In this work, we present the first comprehensive analysis of the impacts of Zero-shot CoT and Few-shot CoT on RLLMs across mathematical reasoning tasks. We examine models ranging from 1.5B to 32B parameters, finding that contrary to concerns, CoT prompting significantly enhances RLLMs' performance in most scenarios. Our results reveal distinct patterns: large-capacity models show minimal improvement on simple tasks but substantial gains on complex problems, while smaller models exhibit the opposite behavior. Further analysis demonstrates that CoT prompting effectively controls the distribution of the numbers of thinking tokens and reasoning steps, reducing excessive reflections by approximately 90% in some cases. Moreover, attention logits analysis reveals the RLLMs' overfitting to reflection-related words, which is mitigated by external CoT guidance. Notably, our experiments indicate that for RLLMs, one-shot CoT consistently yields superior performance compared to Few-shot CoT approaches. Our findings provide important insights for optimizing RLLMs' performance through appropriate prompting strategies.
☆ HoarePrompt: Structural Reasoning About Program Correctness in Natural Language
While software requirements are often expressed in natural language, verifying the correctness of a program against natural language requirements is a hard and underexplored problem. Large language models (LLMs) are promising candidates for addressing this challenge, however our experience shows that they are ineffective in this task, often failing to detect even straightforward bugs. To address this gap, we introduce HoarePrompt, a novel approach that adapts fundamental ideas from program analysis and verification to natural language artifacts. Drawing inspiration from the strongest postcondition calculus, HoarePrompt employs a systematic, step-by-step process in which an LLM generates natural language descriptions of reachable program states at various points in the code. To manage loops, we propose few-shot-driven k-induction, an adaptation of the k-induction method widely used in model checking. Once program states are described, HoarePrompt leverages the LLM to assess whether the program, annotated with these state descriptions, conforms to the natural language requirements. For evaluating the quality of classifiers of program correctness with respect to natural language requirements, we constructed CoCoClaNeL, a challenging dataset of solutions to programming competition problems. Our experiments show that HoarePrompt improves the MCC by 62% compared to directly using Zero-shot-CoT prompts for correctness classification. Furthermore, HoarePrompt outperforms a classifier that assesses correctness via LLM-based test generation by increasing the MCC by 93%. The inductive reasoning mechanism contributes a 28% boost to MCC, underscoring its effectiveness in managing loops.
☆ Optimizing Breast Cancer Detection in Mammograms: A Comprehensive Study of Transfer Learning, Resolution Reduction, and Multi-View Classification
This study explores open questions in the application of machine learning for breast cancer detection in mammograms. Current approaches often employ a two-stage transfer learning process: first, adapting a backbone model trained on natural images to develop a patch classifier, which is then used to create a single-view whole-image classifier. Additionally, many studies leverage both mammographic views to enhance model performance. In this work, we systematically investigate five key questions: (1) Is the intermediate patch classifier essential for optimal performance? (2) Do backbone models that excel in natural image classification consistently outperform others on mammograms? (3) When reducing mammogram resolution for GPU processing, does the learn-to-resize technique outperform conventional methods? (4) Does incorporating both mammographic views in a two-view classifier significantly improve detection accuracy? (5) How do these findings vary when analyzing low-quality versus high-quality mammograms? By addressing these questions, we developed models that outperform previous results for both single-view and two-view classifiers. Our findings provide insights into model architecture and transfer learning strategies contributing to more accurate and efficient mammogram analysis.
comment: 8 pages
☆ FedMM-X: A Trustworthy and Interpretable Framework for Federated Multi-Modal Learning in Dynamic Environments
As artificial intelligence systems increasingly operate in Real-world environments, the integration of multi-modal data sources such as vision, language, and audio presents both unprecedented opportunities and critical challenges for achieving trustworthy intelligence. In this paper, we propose a novel framework that unifies federated learning with explainable multi-modal reasoning to ensure trustworthiness in decentralized, dynamic settings. Our approach, called FedMM-X (Federated Multi-Modal Explainable Intelligence), leverages cross-modal consistency checks, client-level interpretability mechanisms, and dynamic trust calibration to address challenges posed by data heterogeneity, modality imbalance, and out-of-distribution generalization. Through rigorous evaluation across federated multi-modal benchmarks involving vision-language tasks, we demonstrate improved performance in both accuracy and interpretability while reducing vulnerabilities to adversarial and spurious correlations. Further, we introduce a novel trust score aggregation method to quantify global model reliability under dynamic client participation. Our findings pave the way toward developing robust, interpretable, and socially responsible AI systems in Real-world environments.
☆ FLEX: A Benchmark for Evaluating Robustness of Fairness in Large Language Models NAACL 2025
Recent advancements in Large Language Models (LLMs) have significantly enhanced interactions between users and models. These advancements concurrently underscore the need for rigorous safety evaluations due to the manifestation of social biases, which can lead to harmful societal impacts. Despite these concerns, existing benchmarks may overlook the intrinsic weaknesses of LLMs, which can generate biased responses even with simple adversarial instructions. To address this critical gap, we introduce a new benchmark, Fairness Benchmark in LLM under Extreme Scenarios (FLEX), designed to test whether LLMs can sustain fairness even when exposed to prompts constructed to induce bias. To thoroughly evaluate the robustness of LLMs, we integrate prompts that amplify potential biases into the fairness assessment. Comparative experiments between FLEX and existing benchmarks demonstrate that traditional evaluations may underestimate the inherent risks in models. This highlights the need for more stringent LLM evaluation benchmarks to guarantee safety and fairness.
comment: Accepted to NAACL 2025 findings
☆ VectorFit : Adaptive Singular & Bias Vector Fine-Tuning of Pre-trained Foundation Models
Popular PEFT methods achieve parameter efficiency by assuming that incremental weight updates are inherently low-rank, which often leads to a performance gap compared to full fine-tuning. While recent methods have attempted to address this limitation, they typically lack sufficient parameter and memory efficiency. We propose VectorFit, an effective and easily deployable approach that adaptively trains the singular vectors and biases of pre-trained weight matrices. We demonstrate that the utilization of structural and transformational characteristics of pre-trained weights enables high-rank updates comparable to those of full fine-tuning. As a result, VectorFit achieves superior performance with 9X less trainable parameters compared to state-of-the-art PEFT methods. Through extensive experiments over 17 datasets spanning diverse language and vision tasks such as natural language understanding and generation, question answering, image classification, and image generation, we exhibit that VectorFit consistently outperforms baselines, even in extremely low-budget scenarios.
☆ A Spatiotemporal Radar-Based Precipitation Model for Water Level Prediction and Flood Forecasting
Study Region: Goslar and G\"ottingen, Lower Saxony, Germany. Study Focus: In July 2017, the cities of Goslar and G\"ottingen experienced severe flood events characterized by short warning time of only 20 minutes, resulting in extensive regional flooding and significant damage. This highlights the critical need for a more reliable and timely flood forecasting system. This paper presents a comprehensive study on the impact of radar-based precipitation data on forecasting river water levels in Goslar. Additionally, the study examines how precipitation influences water level forecasts in G\"ottingen. The analysis integrates radar-derived spatiotemporal precipitation patterns with hydrological sensor data obtained from ground stations to evaluate the effectiveness of this approach in improving flood prediction capabilities. New Hydrological Insights for the Region: A key innovation in this paper is the use of residual-based modeling to address the non-linearity between precipitation images and water levels, leading to a Spatiotemporal Radar-based Precipitation Model with residuals (STRPMr). Unlike traditional hydrological models, our approach does not rely on upstream data, making it independent of additional hydrological inputs. This independence enhances its adaptability and allows for broader applicability in other regions with RADOLAN precipitation. The deep learning architecture integrates (2+1)D convolutional neural networks for spatial and temporal feature extraction with LSTM for timeseries forecasting. The results demonstrate the potential of the STRPMr for capturing extreme events and more accurate flood forecasting.
comment: 28 pages, 11 figures, 6 tables
☆ RoboFlamingo-Plus: Fusion of Depth and RGB Perception with Vision-Language Models for Enhanced Robotic Manipulation
As robotic technologies advancing towards more complex multimodal interactions and manipulation tasks, the integration of advanced Vision-Language Models (VLMs) has become a key driver in the field. Despite progress with current methods, challenges persist in fusing depth and RGB information within 3D environments and executing tasks guided by linguistic instructions. In response to these challenges, we have enhanced the existing RoboFlamingo framework by introducing RoboFlamingo-Plus, which incorporates depth data into VLMs to significantly improve robotic manipulation performance. Our research achieves a nuanced fusion of RGB and depth information by integrating a pre-trained Vision Transformer (ViT) with a resampling technique, closely aligning this combined data with linguistic cues for superior multimodal understanding. The novelty of RoboFlamingo-Plus lies in its adaptation of inputs for depth data processing, leveraging a pre-trained resampler for depth feature extraction, and employing cross-attention mechanisms for optimal feature integration. These improvements allow RoboFlamingo-Plus to not only deeply understand 3D environments but also easily perform complex, language-guided tasks in challenging settings. Experimental results show that RoboFlamingo-Plus boosts robotic manipulation by 10-20% over current methods, marking a significant advancement. Codes and model weights are public at RoboFlamingo-Plus.
☆ Towards Long-Range ENSO Prediction with an Explainable Deep Learning Model
El Ni\~no-Southern Oscillation (ENSO) is a prominent mode of interannual climate variability with far-reaching global impacts. Its evolution is governed by intricate air-sea interactions, posing significant challenges for long-term prediction. In this study, we introduce CTEFNet, a multivariate deep learning model that synergizes convolutional neural networks and transformers to enhance ENSO forecasting. By integrating multiple oceanic and atmospheric predictors, CTEFNet extends the effective forecast lead time to 20 months while mitigating the impact of the spring predictability barrier, outperforming both dynamical models and state-of-the-art deep learning approaches. Furthermore, CTEFNet offers physically meaningful and statistically significant insights through gradient-based sensitivity analysis, revealing the key precursor signals that govern ENSO dynamics, which align with well-established theories and reveal new insights about inter-basin interactions among the Pacific, Atlantic, and Indian Oceans. The CTEFNet's superior predictive skill and interpretable sensitivity assessments underscore its potential for advancing climate prediction. Our findings highlight the importance of multivariate coupling in ENSO evolution and demonstrate the promise of deep learning in capturing complex climate dynamics with enhanced interpretability.
☆ Pose-Based Fall Detection System: Efficient Monitoring on Standard CPUs
Falls among elderly residents in assisted living homes pose significant health risks, often leading to injuries and a decreased quality of life. Current fall detection solutions typically rely on sensor-based systems that require dedicated hardware, or on video-based models that demand high computational resources and GPUs for real-time processing. In contrast, this paper presents a robust fall detection system that does not require any additional sensors or high-powered hardware. The system uses pose estimation techniques, combined with threshold-based analysis and a voting mechanism, to effectively distinguish between fall and non-fall activities. For pose detection, we leverage MediaPipe, a lightweight and efficient framework that enables real-time processing on standard CPUs with minimal computational overhead. By analyzing motion, body position, and key pose points, the system processes pose features with a 20-frame buffer, minimizing false positives and maintaining high accuracy even in real-world settings. This unobtrusive, resource-efficient approach provides a practical solution for enhancing resident safety in old age homes, without the need for expensive sensors or high-end computational resources.
comment: 4 Pages, 2 figures, 2 code block, 1 flow chart
☆ SMT-EX: An Explainable Surrogate Modeling Toolbox for Mixed-Variables Design Exploration
Surrogate models are of high interest for many engineering applications, serving as cheap-to-evaluate time-efficient approximations of black-box functions to help engineers and practitioners make decisions and understand complex systems. As such, the need for explainability methods is rising and many studies have been performed to facilitate knowledge discovery from surrogate models. To respond to these enquiries, this paper introduces SMT-EX, an enhancement of the open-source Python Surrogate Modeling Toolbox (SMT) that integrates explainability techniques into a state-of-the-art surrogate modelling framework. More precisely, SMT-EX includes three key explainability methods: Shapley Additive Explanations, Partial Dependence Plot, and Individual Conditional Expectations. A peculiar explainability dependency of SMT has been developed for such purpose that can be easily activated once the surrogate model is built, offering a user-friendly and efficient tool for swift insight extraction. The effectiveness of SMT-EX is showcased through two test cases. The first case is a 10-variable wing weight problem with purely continuous variables and the second one is a 3-variable mixed-categorical cantilever beam bending problem. Relying on SMT-EX analyses for these problems, we demonstrate its versatility in addressing a diverse range of problem characteristics. SMT-Explainability is freely available on Github: https://github.com/SMTorg/smt-explainability .
☆ Body Discovery of Embodied AI
In the pursuit of realizing artificial general intelligence (AGI), the importance of embodied artificial intelligence (AI) becomes increasingly apparent. Following this trend, research integrating robots with AGI has become prominent. As various kinds of embodiments have been designed, adaptability to diverse embodiments will become important to AGI. We introduce a new challenge, termed "Body Discovery of Embodied AI", focusing on tasks of recognizing embodiments and summarizing neural signal functionality. The challenge encompasses the precise definition of an AI body and the intricate task of identifying embodiments in dynamic environments, where conventional approaches often prove inadequate. To address these challenges, we apply causal inference method and evaluate it by developing a simulator tailored for testing algorithms with virtual environments. Finally, we validate the efficacy of our algorithms through empirical testing, demonstrating their robust performance in various scenarios based on virtual environments.
☆ A-MESS: Anchor based Multimodal Embedding with Semantic Synchronization for Multimodal Intent Recognition ICME2025
In the domain of multimodal intent recognition (MIR), the objective is to recognize human intent by integrating a variety of modalities, such as language text, body gestures, and tones. However, existing approaches face difficulties adequately capturing the intrinsic connections between the modalities and overlooking the corresponding semantic representations of intent. To address these limitations, we present the Anchor-based Mul- timodal Embedding with Semantic Synchronization (A-MESS) framework. We first design an Anchor-based Multimodal Embed- ding (A-ME) module that employs an anchor-based embedding fusion mechanism to integrate multimodal inputs. Furthermore, we develop a Semantic Synchronization (SS) strategy with the Triplet Contrastive Learning pipeline, which optimizes the pro- cess by synchronizing multimodal representation with label de- scriptions produced by the large language model. Comprehensive experiments indicate that our A-MESS achieves state-of-the-art and provides substantial insight into multimodal representation and downstream tasks.
comment: Accept by ICME2025
☆ ReSearch: Learning to Reason with Search for LLMs via Reinforcement Learning
Large Language Models (LLMs) have shown remarkable capabilities in reasoning, exemplified by the success of OpenAI-o1 and DeepSeek-R1. However, integrating reasoning with external search processes remains challenging, especially for complex multi-hop questions requiring multiple retrieval steps. We propose ReSearch, a novel framework that trains LLMs to Reason with Search via reinforcement learning without using any supervised data on reasoning steps. Our approach treats search operations as integral components of the reasoning chain, where when and how to perform searches is guided by text-based thinking, and search results subsequently influence further reasoning. We train ReSearch on Qwen2.5-7B(-Instruct) and Qwen2.5-32B(-Instruct) models and conduct extensive experiments. Despite being trained on only one dataset, our models demonstrate strong generalizability across various benchmarks. Analysis reveals that ReSearch naturally elicits advanced reasoning capabilities such as reflection and self-correction during the reinforcement learning process.
comment: Work in progress
☆ Enhancing Small Language Models for Cross-Lingual Generalized Zero-Shot Classification with Soft Prompt Tuning NAACL 2025
In NLP, Zero-Shot Classification (ZSC) has become essential for enabling models to classify text into categories unseen during training, particularly in low-resource languages and domains where labeled data is scarce. While pretrained language models (PLMs) have shown promise in ZSC, they often rely on large training datasets or external knowledge, limiting their applicability in multilingual and low-resource scenarios. Recent approaches leveraging natural language prompts reduce the dependence on large training datasets but struggle to effectively incorporate available labeled data from related classification tasks, especially when these datasets originate from different languages or distributions. Moreover, existing prompt-based methods typically rely on manually crafted prompts in a specific language, limiting their adaptability and effectiveness in cross-lingual settings. To address these challenges, we introduce RoSPrompt, a lightweight and data-efficient approach for training soft prompts that enhance cross-lingual ZSC while ensuring robust generalization across data distribution shifts. RoSPrompt is designed for small multilingual PLMs, enabling them to leverage high-resource languages to improve performance in low-resource settings without requiring extensive fine-tuning or high computational costs. We evaluate our approach on multiple multilingual PLMs across datasets covering 106 languages, demonstrating strong cross-lingual transfer performance and robust generalization capabilities over unseen classes.
comment: Workshop on Language Models for Underserved Communities (co-located with NAACL 2025)
☆ Data-centric Federated Graph Learning with Large Language Models
In federated graph learning (FGL), a complete graph is divided into multiple subgraphs stored in each client due to privacy concerns, and all clients jointly train a global graph model by only transmitting model parameters. A pain point of FGL is the heterogeneity problem, where nodes or structures present non-IID properties among clients (e.g., different node label distributions), dramatically undermining the convergence and performance of FGL. To address this, existing efforts focus on design strategies at the model level, i.e., they design models to extract common knowledge to mitigate heterogeneity. However, these model-level strategies fail to fundamentally address the heterogeneity problem as the model needs to be designed from scratch when transferring to other tasks. Motivated by large language models (LLMs) having achieved remarkable success, we aim to utilize LLMs to fully understand and augment local text-attributed graphs, to address data heterogeneity at the data level. In this paper, we propose a general framework LLM4FGL that innovatively decomposes the task of LLM for FGL into two sub-tasks theoretically. Specifically, for each client, it first utilizes the LLM to generate missing neighbors and then infers connections between generated nodes and raw nodes. To improve the quality of generated nodes, we design a novel federated generation-and-reflection mechanism for LLMs, without the need to modify the parameters of the LLM but relying solely on the collective feedback from all clients. After neighbor generation, all the clients utilize a pre-trained edge predictor to infer the missing edges. Furthermore, our framework can seamlessly integrate as a plug-in with existing FGL methods. Experiments on three real-world datasets demonstrate the superiority of our method compared to advanced baselines.
comment: ongoing work
☆ VecTrans: LLM Transformation Framework for Better Auto-vectorization on High-performance CPU
Large language models (LLMs) have demonstrated great capabilities in code generation, yet their effective application in compiler optimizations remains an open challenge due to issues such as hallucinations and a lack of domain-specific reasoning. Vectorization, a crucial optimization for enhancing code performance, often fails because of the compiler's inability to recognize complex code patterns, which commonly require extensive empirical expertise. LLMs, with their ability to capture intricate patterns, thus providing a promising solution to this challenge. This paper presents VecTrans, a novel framework that leverages LLMs to enhance compiler-based code vectorization. VecTrans first employs compiler analysis to identify potentially vectorizable code regions. It then utilizes an LLM to refactor these regions into patterns that are more amenable to the compiler's auto-vectorization. To ensure semantic correctness, VecTrans further integrates a hybrid validation mechanism at the intermediate representation (IR) level. With the above efforts, VecTrans combines the adaptability of LLMs with the precision of compiler vectorization, thereby effectively opening up the vectorization opportunities. Experimental results show that among all 50 TSVC functions unvectorizable by Clang, GCC, and BiShengCompiler, VecTrans successfully vectorizes 23 cases (46%) and achieves an average speedup of 2.02x, greatly surpassing state-of-the-art performance.
☆ FuXi-RTM: A Physics-Guided Prediction Framework with Radiative Transfer Modeling
Similar to conventional video generation, current deep learning-based weather prediction frameworks often lack explicit physical constraints, leading to unphysical outputs that limit their reliability for operational forecasting. Among various physical processes requiring proper representation, radiation plays a fundamental role as it drives Earth's weather and climate systems. However, accurate simulation of radiative transfer processes remains challenging for traditional numerical weather prediction (NWP) models due to their inherent complexity and high computational costs. Here, we propose FuXi-RTM, a hybrid physics-guided deep learning framework designed to enhance weather forecast accuracy while enforcing physical consistency. FuXi-RTM integrates a primary forecasting model (FuXi) with a fixed deep learning-based radiative transfer model (DLRTM) surrogate that efficiently replaces conventional radiation parameterization schemes. This represents the first deep learning-based weather forecasting framework to explicitly incorporate physical process modeling. Evaluated over a comprehensive 5-year dataset, FuXi-RTM outperforms its unconstrained counterpart in 88.51% of 3320 variable and lead time combinations, with improvements in radiative flux predictions. By incorporating additional physical processes, FuXi-RTM paves the way for next-generation weather forecasting systems that are both accurate and physically consistent.
☆ DeCAP: Context-Adaptive Prompt Generation for Debiasing Zero-shot Question Answering in Large Language Models NAACL 2025
While Large Language Models (LLMs) excel in zero-shot Question Answering (QA), they tend to expose biases in their internal knowledge when faced with socially sensitive questions, leading to a degradation in performance. Existing zero-shot methods are efficient but fail to consider context and prevent bias propagation in the answers. To address this, we propose DeCAP, a method for debiasing LLMs using Context-Adaptive Prompt Generation. DeCAP leverages a Question Ambiguity Detection to take appropriate debiasing actions based on the context and a Neutral Answer Guidance Generation to suppress the LLMs make objective judgments about the context, minimizing the propagation of bias from their internal knowledge. Our various experiments across eight LLMs show that DeCAP achieves state-of-the-art zero-shot debiased QA performance. This demonstrates DeCAP's efficacy in enhancing the fairness and accuracy of LLMs in diverse QA settings.
comment: Accepted to NAACL 2025 main. 20 pages, 3 figures
☆ Quantifying Symptom Causality in Clinical Decision Making: An Exploration Using CausaLM
Current machine learning approaches to medical diagnosis often rely on correlational patterns between symptoms and diseases, risking misdiagnoses when symptoms are ambiguous or common across multiple conditions. In this work, we move beyond correlation to investigate the causal influence of key symptoms-specifically "chest pain" on diagnostic predictions. Leveraging the CausaLM framework, we generate counterfactual text representations in which target concepts are effectively "forgotten" enabling a principled estimation of the causal effect of that concept on a model's predicted disease distribution. By employing Textual Representation-based Average Treatment Effect (TReATE), we quantify how the presence or absence of a symptom shapes the model's diagnostic outcomes, and contrast these findings against correlation-based baselines such as CONEXP. Our results offer deeper insight into the decision-making behavior of clinical NLP models and have the potential to inform more trustworthy, interpretable, and causally-grounded decision support tools in medical practice.
☆ Causal invariant geographic network representations with feature and structural distribution shifts
The existing methods learn geographic network representations through deep graph neural networks (GNNs) based on the i.i.d. assumption. However, the spatial heterogeneity and temporal dynamics of geographic data make the out-of-distribution (OOD) generalisation problem particularly salient. The latter are particularly sensitive to distribution shifts (feature and structural shifts) between testing and training data and are the main causes of the OOD generalisation problem. Spurious correlations are present between invariant and background representations due to selection biases and environmental effects, resulting in the model extremes being more likely to learn background representations. The existing approaches focus on background representation changes that are determined by shifts in the feature distributions of nodes in the training and test data while ignoring changes in the proportional distributions of heterogeneous and homogeneous neighbour nodes, which we refer to as structural distribution shifts. We propose a feature-structure mixed invariant representation learning (FSM-IRL) model that accounts for both feature distribution shifts and structural distribution shifts. To address structural distribution shifts, we introduce a sampling method based on causal attention, encouraging the model to identify nodes possessing strong causal relationships with labels or nodes that are more similar to the target node. Inspired by the Hilbert-Schmidt independence criterion, we implement a reweighting strategy to maximise the orthogonality of the node representations, thereby mitigating the spurious correlations among the node representations and suppressing the learning of background representations. Our experiments demonstrate that FSM-IRL exhibits strong learning capabilities on both geographic and social network datasets in OOD scenarios.
comment: 15 pages, 3 figures, 8 tables
☆ DeClotH: Decomposable 3D Cloth and Human Body Reconstruction from a Single Image CVPR 2025
Most existing methods of 3D clothed human reconstruction from a single image treat the clothed human as a single object without distinguishing between cloth and human body. In this regard, we present DeClotH, which separately reconstructs 3D cloth and human body from a single image. This task remains largely unexplored due to the extreme occlusion between cloth and the human body, making it challenging to infer accurate geometries and textures. Moreover, while recent 3D human reconstruction methods have achieved impressive results using text-to-image diffusion models, directly applying such an approach to this problem often leads to incorrect guidance, particularly in reconstructing 3D cloth. To address these challenges, we propose two core designs in our framework. First, to alleviate the occlusion issue, we leverage 3D template models of cloth and human body as regularizations, which provide strong geometric priors to prevent erroneous reconstruction by the occlusion. Second, we introduce a cloth diffusion model specifically designed to provide contextual information about cloth appearance, thereby enhancing the reconstruction of 3D cloth. Qualitative and quantitative experiments demonstrate that our proposed approach is highly effective in reconstructing both 3D cloth and the human body. More qualitative results are provided at https://hygenie1228.github.io/DeClotH/.
comment: Published at CVPR 2025, 17 pages including the supplementary material
☆ Flow to Learn: Flow Matching on Neural Network Parameters ICLR
Foundational language models show a remarkable ability to learn new concepts during inference via context data. However, similar work for images lag behind. To address this challenge, we introduce FLoWN, a flow matching model that learns to generate neural network parameters for different tasks. Our approach models the flow on latent space, while conditioning the process on context data. Experiments verify that FLoWN attains various desiderata for a meta-learning model. In addition, it matches or exceeds baselines on in-distribution tasks, provides better initializations for classifier training, and is performant on out-of-distribution few-shot tasks while having a fine-tuning mechanism to improve performance.
comment: Accepted at the ICLR Workshop on Neural Network Weights as a New Data Modality 2025
☆ Efficient IoT Intrusion Detection with an Improved Attention-Based CNN-BiLSTM Architecture
The ever-increasing security vulnerabilities in the Internet-of-Things (IoT) systems require improved threat detection approaches. This paper presents a compact and efficient approach to detect botnet attacks by employing an integrated approach that consists of traffic pattern analysis, temporal support learning, and focused feature extraction. The proposed attention-based model benefits from a hybrid CNN-BiLSTM architecture and achieves 99% classification accuracy in detecting botnet attacks utilizing the N-BaIoT dataset, while maintaining high precision and recall across various scenarios. The proposed model's performance is further validated by key parameters, such as Mathews Correlation Coefficient and Cohen's kappa Correlation Coefficient. The close-to-ideal results for these parameters demonstrate the proposed model's ability to detect botnet attacks accurately and efficiently in practical settings and on unseen data. The proposed model proved to be a powerful defense mechanism for IoT networks to face emerging security challenges.
☆ Wavelet-based Global-Local Interaction Network with Cross-Attention for Multi-View Diabetic Retinopathy Detection ICME
Multi-view diabetic retinopathy (DR) detection has recently emerged as a promising method to address the issue of incomplete lesions faced by single-view DR. However, it is still challenging due to the variable sizes and scattered locations of lesions. Furthermore, existing multi-view DR methods typically merge multiple views without considering the correlations and redundancies of lesion information across them. Therefore, we propose a novel method to overcome the challenges of difficult lesion information learning and inadequate multi-view fusion. Specifically, we introduce a two-branch network to obtain both local lesion features and their global dependencies. The high-frequency component of the wavelet transform is used to exploit lesion edge information, which is then enhanced by global semantic to facilitate difficult lesion learning. Additionally, we present a cross-view fusion module to improve multi-view fusion and reduce redundancy. Experimental results on large public datasets demonstrate the effectiveness of our method. The code is open sourced on https://github.com/HuYongting/WGLIN.
comment: Accepted by IEEE International Conference on Multimedia & Expo (ICME) 2025
☆ Substance over Style: Evaluating Proactive Conversational Coaching Agents
While NLP research has made strides in conversational tasks, many approaches focus on single-turn responses with well-defined objectives or evaluation criteria. In contrast, coaching presents unique challenges with initially undefined goals that evolve through multi-turn interactions, subjective evaluation criteria, mixed-initiative dialogue. In this work, we describe and implement five multi-turn coaching agents that exhibit distinct conversational styles, and evaluate them through a user study, collecting first-person feedback on 155 conversations. We find that users highly value core functionality, and that stylistic components in absence of core components are viewed negatively. By comparing user feedback with third-person evaluations from health experts and an LM, we reveal significant misalignment across evaluation approaches. Our findings provide insights into design and evaluation of conversational coaching agents and contribute toward improving human-centered NLP applications.
☆ Process or Result? Manipulated Ending Tokens Can Mislead Reasoning LLMs to Ignore the Correct Reasoning Steps
Recent reasoning large language models (LLMs) have demonstrated remarkable improvements in mathematical reasoning capabilities through long Chain-of-Thought. The reasoning tokens of these models enable self-correction within reasoning chains, enhancing robustness. This motivates our exploration: how vulnerable are reasoning LLMs to subtle errors in their input reasoning chains? We introduce "Compromising Thought" (CPT), a vulnerability where models presented with reasoning tokens containing manipulated calculation results tend to ignore correct reasoning steps and adopt incorrect results instead. Through systematic evaluation across multiple reasoning LLMs, we design three increasingly explicit prompting methods to measure CPT resistance, revealing that models struggle significantly to identify and correct these manipulations. Notably, contrary to existing research suggesting structural alterations affect model performance more than content modifications, we find that local ending token manipulations have greater impact on reasoning outcomes than structural changes. Moreover, we discover a security vulnerability in DeepSeek-R1 where tampered reasoning tokens can trigger complete reasoning cessation. Our work enhances understanding of reasoning robustness and highlights security considerations for reasoning-intensive applications.
☆ LRSCLIP: A Vision-Language Foundation Model for Aligning Remote Sensing Image with Longer Text
This study addresses the technical bottlenecks in handling long text and the "hallucination" issue caused by insufficient short text information in remote sensing vision-language foundation models (VLFM). We propose a novel vision-language foundation model, LRSCLIP, and a multimodal dataset, LRS2M. The main contributions are as follows: (1) By integrating multi-source remote sensing data and adopting a large language model labeling strategy, we construct the LRS2M dataset, which contains 2 million image-text pairs, providing both short and long texts for the first time, thus solving the problem of semantic granularity limitations in existing datasets; (2) The design of the LRSCLIP architecture based on Long-CLIP's KPS module, which extends CLIP's text processing capacity and achieves fine-grained cross-modal feature alignment through a dual-text loss weighting mechanism. Experimental results show that LRSCLIP improves retrieval accuracy by 10\%-20\% over the Long-CLIP baseline in the zero-shot long-text cross-modal retrieval task. For the zero-shot short-text cross-modal retrieval task, LRSCLIP achieves improvements over the current best model, GeoRSCLIP, with increases of 0.17\%, 0.67\%, and 0.92\% in Text to Image R@1, Image to Text R@1, and mR on RSITMD, respectively, and 0.04\%, 2.93\%, and 1.28\% on RSICD. In the zero-shot image classification task (average accuracy=75.75\%) and semantic localization task (Rmi=0.7653), LRSCLIP achieves state-of-the-art performance. These results validate the dual advantages of fine-grained semantic understanding and global feature matching in LRSCLIP. This work provides a new benchmark model and data support for remote sensing multimodal learning. The related code has been open source and is available at https://github.com/MitsuiChen14/LRSCLIP.
comment: 17 pages, 12 figures
☆ Observation Adaptation via Annealed Importance Resampling for Partially Observable Markov Decision Processes ICAPS 2025
Partially observable Markov decision processes (POMDPs) are a general mathematical model for sequential decision-making in stochastic environments under state uncertainty. POMDPs are often solved \textit{online}, which enables the algorithm to adapt to new information in real time. Online solvers typically use bootstrap particle filters based on importance resampling for updating the belief distribution. Since directly sampling from the ideal state distribution given the latest observation and previous state is infeasible, particle filters approximate the posterior belief distribution by propagating states and adjusting weights through prediction and resampling steps. However, in practice, the importance resampling technique often leads to particle degeneracy and sample impoverishment when the state transition model poorly aligns with the posterior belief distribution, especially when the received observation is highly informative. We propose an approach that constructs a sequence of bridge distributions between the state-transition and optimal distributions through iterative Monte Carlo steps, better accommodating noisy observations in online POMDP solvers. Our algorithm demonstrates significantly superior performance compared to state-of-the-art methods when evaluated across multiple challenging POMDP domains.
comment: Accepted as Oral Presentation to ICAPS 2025
☆ Adaptive Wavelet Filters as Practical Texture Feature Amplifiers for Parkinson's Disease Screening in OCT
Parkinson's disease (PD) is a prevalent neurodegenerative disorder globally. The eye's retina is an extension of the brain and has great potential in PD screening. Recent studies have suggested that texture features extracted from retinal layers can be adopted as biomarkers for PD diagnosis under optical coherence tomography (OCT) images. Frequency domain learning techniques can enhance the feature representations of deep neural networks (DNNs) by decomposing frequency components involving rich texture features. Additionally, previous works have not exploited texture features for automated PD screening in OCT. Motivated by the above analysis, we propose a novel Adaptive Wavelet Filter (AWF) that serves as the Practical Texture Feature Amplifier to fully leverage the merits of texture features to boost the PD screening performance of DNNs with the aid of frequency domain learning. Specifically, AWF first enhances texture feature representation diversities via channel mixer, then emphasizes informative texture feature representations with the well-designed adaptive wavelet filtering token mixer. By combining the AWFs with the DNN stem, AWFNet is constructed for automated PD screening. Additionally, we introduce a novel Balanced Confidence (BC) Loss by mining the potential of sample-wise predicted probabilities of all classes and class frequency prior, to further boost the PD screening performance and trustworthiness of AWFNet. The extensive experiments manifest the superiority of our AWFNet and BC over state-of-the-art methods in terms of PD screening performance and trustworthiness.
☆ No Black Box Anymore: Demystifying Clinical Predictive Modeling with Temporal-Feature Cross Attention Mechanism
Despite the outstanding performance of deep learning models in clinical prediction tasks, explainability remains a significant challenge. Inspired by transformer architectures, we introduce the Temporal-Feature Cross Attention Mechanism (TFCAM), a novel deep learning framework designed to capture dynamic interactions among clinical features across time, enhancing both predictive accuracy and interpretability. In an experiment with 1,422 patients with Chronic Kidney Disease, predicting progression to End-Stage Renal Disease, TFCAM outperformed LSTM and RETAIN baselines, achieving an AUROC of 0.95 and an F1-score of 0.69. Beyond performance gains, TFCAM provides multi-level explainability by identifying critical temporal periods, ranking feature importance, and quantifying how features influence each other across time before affecting predictions. Our approach addresses the "black box" limitations of deep learning in healthcare, offering clinicians transparent insights into disease progression mechanisms while maintaining state-of-the-art predictive performance.
comment: 10 pages, 3 figures, submitted to AMIA 2025
☆ CubeRobot: Grounding Language in Rubik's Cube Manipulation via Vision-Language Model
Proving Rubik's Cube theorems at the high level represents a notable milestone in human-level spatial imagination and logic thinking and reasoning. Traditional Rubik's Cube robots, relying on complex vision systems and fixed algorithms, often struggle to adapt to complex and dynamic scenarios. To overcome this limitation, we introduce CubeRobot, a novel vision-language model (VLM) tailored for solving 3x3 Rubik's Cubes, empowering embodied agents with multimodal understanding and execution capabilities. We used the CubeCoT image dataset, which contains multiple-level tasks (43 subtasks in total) that humans are unable to handle, encompassing various cube states. We incorporate a dual-loop VisionCoT architecture and Memory Stream, a paradigm for extracting task-related features from VLM-generated planning queries, thus enabling CubeRobot to independent planning, decision-making, reflection and separate management of high- and low-level Rubik's Cube tasks. Furthermore, in low-level Rubik's Cube restoration tasks, CubeRobot achieved a high accuracy rate of 100%, similar to 100% in medium-level tasks, and achieved an accuracy rate of 80% in high-level tasks.
☆ LogicLearner: A Tool for the Guided Practice of Propositional Logic Proofs
The study of propositional logic -- fundamental to the theory of computing -- is a cornerstone of the undergraduate computer science curriculum. Learning to solve logical proofs requires repeated guided practice, but undergraduate students often lack access to on-demand tutoring in a judgment-free environment. In this work, we highlight the need for guided practice tools in undergraduate mathematics education and outline the desiderata of an effective practice tool. We accordingly develop LogicLearner, a web application for guided logic proof practice. LogicLearner consists of an interface to attempt logic proofs step-by-step and an automated proof solver to generate solutions on the fly, allowing users to request guidance as needed. We pilot LogicLearner as a practice tool in two semesters of an undergraduate discrete mathematics course and receive strongly positive feedback for usability and pedagogical value in student surveys. To the best of our knowledge, LogicLearner is the only learning tool that provides an end-to-end practice environment for logic proofs with immediate, judgment-free feedback.
comment: 32 pages, 27 figures, open-source codebase linked in paper
☆ Context-Aware Semantic Segmentation: Enhancing Pixel-Level Understanding with Large Language Models for Advanced Vision Applications
Semantic segmentation has made significant strides in pixel-level image understanding, yet it remains limited in capturing contextual and semantic relationships between objects. Current models, such as CNN and Transformer-based architectures, excel at identifying pixel-level features but fail to distinguish semantically similar objects (e.g., "doctor" vs. "nurse" in a hospital scene) or understand complex contextual scenarios (e.g., differentiating a running child from a regular pedestrian in autonomous driving). To address these limitations, we proposed a novel Context-Aware Semantic Segmentation framework that integrates Large Language Models (LLMs) with state-of-the-art vision backbones. Our hybrid model leverages the Swin Transformer for robust visual feature extraction and GPT-4 for enriching semantic understanding through text embeddings. A Cross-Attention Mechanism is introduced to align vision and language features, enabling the model to reason about context more effectively. Additionally, Graph Neural Networks (GNNs) are employed to model object relationships within the scene, capturing dependencies that are overlooked by traditional models. Experimental results on benchmark datasets (e.g., COCO, Cityscapes) demonstrate that our approach outperforms the existing methods in both pixel-level accuracy (mIoU) and contextual understanding (mAP). This work bridges the gap between vision and language, paving the path for more intelligent and context-aware vision systems in applications including autonomous driving, medical imaging, and robotics.
☆ Reverse Prompt: Cracking the Recipe Inside Text-to-Image Generation
Text-to-image generation has become increasingly popular, but achieving the desired images often requires extensive prompt engineering. In this paper, we explore how to decode textual prompts from reference images, a process we refer to as image reverse prompt engineering. This technique enables us to gain insights from reference images, understand the creative processes of great artists, and generate impressive new images. To address this challenge, we propose a method known as automatic reverse prompt optimization (ARPO). Specifically, our method refines an initial prompt into a high-quality prompt through an iteratively imitative gradient prompt optimization process: 1) generating a recreated image from the current prompt to instantiate its guidance capability; 2) producing textual gradients, which are candidate prompts intended to reduce the difference between the recreated image and the reference image; 3) updating the current prompt with textual gradients using a greedy search method to maximize the CLIP similarity between prompt and reference image. We compare ARPO with several baseline methods, including handcrafted techniques, gradient-based prompt tuning methods, image captioning, and data-driven selection method. Both quantitative and qualitative results demonstrate that our ARPO converges quickly to generate high-quality reverse prompts. More importantly, we can easily create novel images with diverse styles and content by directly editing these reverse prompts. Code will be made publicly available.
☆ NeoRL-2: Near Real-World Benchmarks for Offline Reinforcement Learning with Extended Realistic Scenarios
Offline reinforcement learning (RL) aims to learn from historical data without requiring (costly) access to the environment. To facilitate offline RL research, we previously introduced NeoRL, which highlighted that datasets from real-world tasks are often conservative and limited. With years of experience applying offline RL to various domains, we have identified additional real-world challenges. These include extremely conservative data distributions produced by deployed control systems, delayed action effects caused by high-latency transitions, external factors arising from the uncontrollable variance of transitions, and global safety constraints that are difficult to evaluate during the decision-making process. These challenges are underrepresented in previous benchmarks but frequently occur in real-world tasks. To address this, we constructed the extended Near Real-World Offline RL Benchmark (NeoRL-2), which consists of 7 datasets from 7 simulated tasks along with their corresponding evaluation simulators. Benchmarking results from state-of-the-art offline RL approaches demonstrate that current methods often struggle to outperform the data-collection behavior policy, highlighting the need for more effective methods. We hope NeoRL-2 will accelerate the development of reinforcement learning algorithms for real-world applications. The benchmark project page is available at https://github.com/polixir/NeoRL2.
☆ Linguistic Blind Spots of Large Language Models NAACL 2025
Large language models (LLMs) are the foundation of many AI applications today. However, despite their remarkable proficiency in generating coherent text, questions linger regarding their ability to perform fine-grained linguistic annotation tasks, such as detecting nouns or verbs, or identifying more complex syntactic structures like clauses in input texts. These tasks require precise syntactic and semantic understanding of input text, and when LLMs underperform on specific linguistic structures, it raises concerns about their reliability for detailed linguistic analysis and whether their (even correct) outputs truly reflect an understanding of the inputs. In this paper, we empirically study the performance of recent LLMs on fine-grained linguistic annotation tasks. Through a series of experiments, we find that recent LLMs show limited efficacy in addressing linguistic queries and often struggle with linguistically complex inputs. We show that the most capable LLM (Llama3-70b) makes notable errors in detecting linguistic structures, such as misidentifying embedded clauses, failing to recognize verb phrases, and confusing complex nominals with clauses. Our results provide insights to inform future advancements in LLM design and development.
comment: NAACL 2025 Cognitive Modeling and Computational Linguistics Workshop
☆ Face Spoofing Detection using Deep Learning
Digital image spoofing has emerged as a significant security threat in biometric authentication systems, particularly those relying on facial recognition. This study evaluates the performance of three vision based models, MobileNetV2, ResNET50, and Vision Transformer, ViT, for spoof detection in image classification, utilizing a dataset of 150,986 images divided into training , 140,002, testing, 10,984, and validation ,39,574, sets. Spoof detection is critical for enhancing the security of image recognition systems, and this research compares the models effectiveness through accuracy, precision, recall, and F1 score metrics. Results reveal that MobileNetV2 outperforms other architectures on the test dataset, achieving an accuracy of 91.59%, precision of 91.72%, recall of 91.59%, and F1 score of 91.58%, compared to ViT 86.54%, 88.28%, 86.54%, and 86.39%, respectively. On the validation dataset, MobileNetV2, and ViT excel, with MobileNetV2 slightly ahead at 97.17% accuracy versus ViT 96.36%. MobileNetV2 demonstrates faster convergence during training and superior generalization to unseen data, despite both models showing signs of overfitting. These findings highlight MobileNetV2 balanced performance and robustness, making it the preferred choice for spoof detection applications where reliability on new data is essential. The study underscores the importance of model selection in security sensitive contexts and suggests MobileNetV2 as a practical solution for real world deployment.
comment: 26 pages, 9 figures,3 tables
♻ ☆ Aether: Geometric-Aware Unified World Modeling
The integration of geometric reconstruction and generative modeling remains a critical challenge in developing AI systems capable of human-like spatial reasoning. This paper proposes Aether, a unified framework that enables geometry-aware reasoning in world models by jointly optimizing three core capabilities: (1) 4D dynamic reconstruction, (2) action-conditioned video prediction, and (3) goal-conditioned visual planning. Through task-interleaved feature learning, Aether achieves synergistic knowledge sharing across reconstruction, prediction, and planning objectives. Building upon video generation models, our framework demonstrates unprecedented synthetic-to-real generalization despite never observing real-world data during training. Furthermore, our approach achieves zero-shot generalization in both action following and reconstruction tasks, thanks to its intrinsic geometric modeling. Remarkably, even without real-world data, its reconstruction performance is comparable with or even better than that of domain-specific models. Additionally, Aether employs camera trajectories as geometry-informed action spaces, enabling effective action-conditioned prediction and visual planning. We hope our work inspires the community to explore new frontiers in physically-reasonable world modeling and its applications.
comment: Project Page: https://aether-world.github.io/
♻ ☆ Structuring Scientific Innovation: A Framework for Modeling and Discovering Impactful Knowledge Combinations
The emergence of large language models offers new possibilities for structured exploration of scientific knowledge. Rather than viewing scientific discovery as isolated ideas or content, we propose a structured approach that emphasizes the role of method combinations in shaping disruptive insights. Specifically, we investigate how knowledge unit--especially those tied to methodological design--can be modeled and recombined to yield research breakthroughs. Our proposed framework addresses two key challenges. First, we introduce a contrastive learning-based mechanism to identify distinguishing features of historically disruptive method combinations within problem-driven contexts. Second, we propose a reasoning-guided Monte Carlo search algorithm that leverages the chain-of-thought capability of LLMs to identify promising knowledge recombinations for new problem statements.Empirical studies across multiple domains show that the framework is capable of modeling the structural dynamics of innovation and successfully highlights combinations with high disruptive potential. This research provides a new path for computationally guided scientific ideation grounded in structured reasoning and historical data modeling.
♻ ☆ MC-LLaVA: Multi-Concept Personalized Vision-Language Model
Current vision-language models (VLMs) show exceptional abilities across diverse tasks, such as visual question answering. To enhance user experience, recent studies investigate VLM personalization to understand user-provided concepts. However, they mainly focus on single-concept personalization, neglecting the existence and interplay of multiple concepts, which limits real-world applicability. This paper proposes the first multi-concept personalization paradigm, MC-LLaVA. Specifically, MC-LLaVA employs a multi-concept instruction tuning strategy, effectively integrating multiple concepts in a single training step. To reduce the costs related to joint training, we propose a personalized textual prompt that uses visual token information to initialize concept tokens. Additionally, we introduce a personalized visual prompt during inference, aggregating location confidence maps for enhanced recognition and grounding capabilities. To advance multi-concept personalization research, we further contribute a high-quality instruction tuning dataset. We carefully collect images with multiple characters and objects from movies and manually generate question-answer samples for multi-concept scenarios, featuring superior diversity. Comprehensive qualitative and quantitative experiments demonstrate that MC-LLaVA can achieve impressive multi-concept personalized responses, paving the way for VLMs to become better user-specific assistants. The code and dataset will be publicly available at https://github.com/arctanxarc/MC-LLaVA}.
comment: I sincerely apologize for any inconvenience caused. We actually uploaded this paper to arXiv in November 2024, as arXiv:2411.11706. During this update, we did not consider the replacement operation of arXiv, which led to duplicate submissions. We have made modifications at the original address arXiv:2411.11706
♻ ☆ Lightweight Embedded FPGA Deployment of Learned Image Compression with Knowledge Distillation and Hybrid Quantization
Learnable Image Compression (LIC) has shown the potential to outperform standardized video codecs in RD efficiency, prompting the research for hardware-friendly implementations. Most existing LIC hardware implementations prioritize latency to RD-efficiency and through an extensive exploration of the hardware design space. We present a novel design paradigm where the burden of tuning the design for a specific hardware platform is shifted towards model dimensioning and without compromising on RD-efficiency. First, we design a framework for distilling a leaner student LIC model from a reference teacher: by tuning a single model hyperparameters, we can meet the constraints of different hardware platforms without a complex hardware design exploration. Second, we propose a hardware-friendly implementation of the Generalized Divisive Normalization - GDN activation that preserves RD efficiency even post parameter quantization. Third, we design a pipelined FPGA configuration which takes full advantage of available FPGA resources by leveraging parallel processing and optimizing resource allocation. Our experiments with a state of the art LIC model show that we outperform all existing FPGA implementations while performing very close to the original model.
comment: 1. Submitted to IEEE Transactions on Circuits and Systems for Video Technology in March 2025. 2. Corrected numerous mistakes from previous versions in results, citations and metrics numbers in figures
♻ ☆ Frequency Dynamic Convolution for Dense Image Prediction CVPR 2025
While Dynamic Convolution (DY-Conv) has shown promising performance by enabling adaptive weight selection through multiple parallel weights combined with an attention mechanism, the frequency response of these weights tends to exhibit high similarity, resulting in high parameter costs but limited adaptability. In this work, we introduce Frequency Dynamic Convolution (FDConv), a novel approach that mitigates these limitations by learning a fixed parameter budget in the Fourier domain. FDConv divides this budget into frequency-based groups with disjoint Fourier indices, enabling the construction of frequency-diverse weights without increasing the parameter cost. To further enhance adaptability, we propose Kernel Spatial Modulation (KSM) and Frequency Band Modulation (FBM). KSM dynamically adjusts the frequency response of each filter at the spatial level, while FBM decomposes weights into distinct frequency bands in the frequency domain and modulates them dynamically based on local content. Extensive experiments on object detection, segmentation, and classification validate the effectiveness of FDConv. We demonstrate that when applied to ResNet-50, FDConv achieves superior performance with a modest increase of +3.6M parameters, outperforming previous methods that require substantial increases in parameter budgets (e.g., CondConv +90M, KW +76.5M). Moreover, FDConv seamlessly integrates into a variety of architectures, including ConvNeXt, Swin-Transformer, offering a flexible and efficient solution for modern vision tasks. The code is made publicly available at https://github.com/Linwei-Chen/FDConv.
comment: Accepted by CVPR 2025
♻ ☆ Commander-GPT: Fully Unleashing the Sarcasm Detection Capability of Multi-Modal Large Language Models
Sarcasm detection, as a crucial research direction in the field of Natural Language Processing (NLP), has attracted widespread attention. Traditional sarcasm detection tasks have typically focused on single-modal approaches (e.g., text), but due to the implicit and subtle nature of sarcasm, such methods often fail to yield satisfactory results. In recent years, researchers have shifted the focus of sarcasm detection to multi-modal approaches. However, effectively leveraging multi-modal information to accurately identify sarcastic content remains a challenge that warrants further exploration. Leveraging the powerful integrated processing capabilities of Multi-Modal Large Language Models (MLLMs) for various information sources, we propose an innovative multi-modal Commander-GPT framework. Inspired by military strategy, we first decompose the sarcasm detection task into six distinct sub-tasks. A central commander (decision-maker) then assigns the best-suited large language model to address each specific sub-task. Ultimately, the detection results from each model are aggregated to identify sarcasm. We conducted extensive experiments on MMSD and MMSD 2.0, utilizing four multi-modal large language models and six prompting strategies. Our experiments demonstrate that our approach achieves state-of-the-art performance, with a 19.3% improvement in F1 score, without necessitating fine-tuning or ground-truth rationales.
♻ ☆ Any6D: Model-free 6D Pose Estimation of Novel Objects CVPR 2025
We introduce Any6D, a model-free framework for 6D object pose estimation that requires only a single RGB-D anchor image to estimate both the 6D pose and size of unknown objects in novel scenes. Unlike existing methods that rely on textured 3D models or multiple viewpoints, Any6D leverages a joint object alignment process to enhance 2D-3D alignment and metric scale estimation for improved pose accuracy. Our approach integrates a render-and-compare strategy to generate and refine pose hypotheses, enabling robust performance in scenarios with occlusions, non-overlapping views, diverse lighting conditions, and large cross-environment variations. We evaluate our method on five challenging datasets: REAL275, Toyota-Light, HO3D, YCBINEOAT, and LM-O, demonstrating its effectiveness in significantly outperforming state-of-the-art methods for novel object pose estimation. Project page: https://taeyeop.com/any6d
comment: CVPR 2025, Project Page: https://taeyeop.com/any6d
♻ ☆ CompMarkGS: Robust Watermarking for Compressed 3D Gaussian Splatting
3D Gaussian Splatting (3DGS) enables rapid differentiable rendering for 3D reconstruction and novel view synthesis, leading to its widespread commercial use. Consequently, copyright protection via watermarking has become critical. However, because 3DGS relies on millions of Gaussians, which require gigabytes of storage, efficient transfer and storage require compression. Existing 3DGS watermarking methods are vulnerable to quantization-based compression, often resulting in the loss of the embedded watermark. To address this challenge, we propose a novel watermarking method that ensures watermark robustness after model compression while maintaining high rendering quality. In detail, we incorporate a quantization distortion layer that simulates compression during training, preserving the watermark under quantization-based compression. Also, we propose a learnable watermark embedding feature that embeds the watermark into the anchor feature, ensuring structural consistency and seamless integration into the 3D scene. Furthermore, we present a frequency-aware anchor growing mechanism to enhance image quality in high-frequency regions by effectively identifying Guassians within these regions. Experimental results confirm that our method preserves the watermark and maintains superior image quality under high compression, validating it as a promising approach for a secure 3DGS model.
comment: 23 pages, 17 figures
♻ ☆ Lessons and Insights from a Unifying Study of Parameter-Efficient Fine-Tuning (PEFT) in Visual Recognition CVPR 2025
Parameter-efficient fine-tuning (PEFT) has attracted significant attention due to the growth of pre-trained model sizes and the need to fine-tune (FT) them for superior downstream performance. Despite a surge in new PEFT methods, a systematic study to understand their performance and suitable application scenarios is lacking, leaving questions like "when to apply PEFT" and "which method to use" largely unanswered, especially in visual recognition. In this paper, we conduct a unifying empirical study of representative PEFT methods with Vision Transformers. We systematically tune their hyperparameters to fairly compare their accuracy on downstream tasks. Our study offers a practical user guide and unveils several new insights. First, if tuned carefully, different PEFT methods achieve similar accuracy in the low-shot benchmark VTAB-1K. This includes simple approaches like FT the bias terms that were reported inferior. Second, despite similar accuracy, we find that PEFT methods make different mistakes and high-confidence predictions, likely due to their different inductive biases. Such an inconsistency (or complementarity) opens up the opportunity for ensemble methods, and we make preliminary attempts at this. Third, going beyond the commonly used low-shot tasks, we find that PEFT is also useful in many-shot regimes, achieving comparable or better accuracy than full FT while using significantly fewer parameters. Lastly, we investigate PEFT's ability to preserve a pre-trained model's robustness to distribution shifts (e.g., CLIP). Perhaps not surprisingly, PEFT approaches outperform full FT alone. However, with weight-space ensembles, full FT can better balance target distribution and distribution shift performance, suggesting a future research direction for robust PEFT.
comment: CVPR 2025. The code is available at https://github.com/OSU-MLB/ViT_PEFT_Vision
♻ ☆ A Mechanistic Explanatory Strategy for XAI
Despite significant advancements in XAI, scholars continue to note a persistent lack of robust conceptual foundations and integration with broader discourse on scientific explanation. In response, emerging XAI research increasingly draws on explanatory strategies from various scientific disciplines and the philosophy of science to address these gaps. This paper outlines a mechanistic strategy for explaining the functional organization of deep learning systems, situating recent developments in AI explainability within a broader philosophical context. According to the mechanistic approach, explaining opaque AI systems involves identifying the mechanisms underlying decision-making processes. For deep neural networks, this means discerning functionally relevant components - such as neurons, layers, circuits, or activation patterns - and understanding their roles through decomposition, localization, and recomposition. Proof-of-principle case studies from image recognition and language modeling align this theoretical framework with recent research from OpenAI and Anthropic. The findings suggest that pursuing mechanistic explanations can uncover elements that traditional explainability techniques may overlook, ultimately contributing to more thoroughly explainable AI.
comment: Forthcoming in M\"uller, V. C., Dung, L., L\"ohr, G., & Rumana, A. (Eds.). Philosophy of Artificial Intelligence: The State of the Art, Synthese Library, Springer Nature. Please cite the published version
♻ ☆ LoTUS: Large-Scale Machine Unlearning with a Taste of Uncertainty CVPR 2025
We present LoTUS, a novel Machine Unlearning (MU) method that eliminates the influence of training samples from pre-trained models, avoiding retraining from scratch. LoTUS smooths the prediction probabilities of the model up to an information-theoretic bound, mitigating its over-confidence stemming from data memorization. We evaluate LoTUS on Transformer and ResNet18 models against eight baselines across five public datasets. Beyond established MU benchmarks, we evaluate unlearning on ImageNet1k, a large-scale dataset, where retraining is impractical, simulating real-world conditions. Moreover, we introduce the novel Retrain-Free Jensen-Shannon Divergence (RF-JSD) metric to enable evaluation under real-world conditions. The experimental results show that LoTUS outperforms state-of-the-art methods in terms of both efficiency and effectiveness. Code: https://github.com/cspartalis/LoTUS.
comment: Accepted as a main conference paper at CVPR 2025 (https://cvpr.thecvf.com/virtual/2025/poster/33292)
♻ ☆ PG-SAM: Prior-Guided SAM with Medical for Multi-organ Segmentation
Segment Anything Model (SAM) demonstrates powerful zero-shot capabilities; however, its accuracy and robustness significantly decrease when applied to medical image segmentation. Existing methods address this issue through modality fusion, integrating textual and image information to provide more detailed priors. In this study, we argue that the granularity of text and the domain gap affect the accuracy of the priors. Furthermore, the discrepancy between high-level abstract semantics and pixel-level boundary details in images can introduce noise into the fusion process. To address this, we propose Prior-Guided SAM (PG-SAM), which employs a fine-grained modality prior aligner to leverage specialized medical knowledge for better modality alignment. The core of our method lies in efficiently addressing the domain gap with fine-grained text from a medical LLM. Meanwhile, it also enhances the priors' quality after modality alignment, ensuring more accurate segmentation. In addition, our decoder enhances the model's expressive capabilities through multi-level feature fusion and iterative mask optimizer operations, supporting unprompted learning. We also propose a unified pipeline that effectively supplies high-quality semantic information to SAM. Extensive experiments on the Synapse dataset demonstrate that the proposed PG-SAM achieves state-of-the-art performance. Our anonymous code is released at https://github.com/logan-0623/PG-SAM.
♻ ☆ Evaluating Negative Sampling Approaches for Neural Topic Models
Negative sampling has emerged as an effective technique that enables deep learning models to learn better representations by introducing the paradigm of learn-to-compare. The goal of this approach is to add robustness to deep learning models to learn better representation by comparing the positive samples against the negative ones. Despite its numerous demonstrations in various areas of computer vision and natural language processing, a comprehensive study of the effect of negative sampling in an unsupervised domain like topic modeling has not been well explored. In this paper, we present a comprehensive analysis of the impact of different negative sampling strategies on neural topic models. We compare the performance of several popular neural topic models by incorporating a negative sampling technique in the decoder of variational autoencoder-based neural topic models. Experiments on four publicly available datasets demonstrate that integrating negative sampling into topic models results in significant enhancements across multiple aspects, including improved topic coherence, richer topic diversity, and more accurate document classification. Manual evaluations also indicate that the inclusion of negative sampling into neural topic models enhances the quality of the generated topics. These findings highlight the potential of negative sampling as a valuable tool for advancing the effectiveness of neural topic models.
comment: Code is available at: https://github.com/AdhyaSuman/Eval_NegTM
♻ ☆ Inductive Moment Matching
Diffusion models and Flow Matching generate high-quality samples but are slow at inference, and distilling them into few-step models often leads to instability and extensive tuning. To resolve these trade-offs, we propose Inductive Moment Matching (IMM), a new class of generative models for one- or few-step sampling with a single-stage training procedure. Unlike distillation, IMM does not require pre-training initialization and optimization of two networks; and unlike Consistency Models, IMM guarantees distribution-level convergence and remains stable under various hyperparameters and standard model architectures. IMM surpasses diffusion models on ImageNet-256x256 with 1.99 FID using only 8 inference steps and achieves state-of-the-art 2-step FID of 1.98 on CIFAR-10 for a model trained from scratch.
♻ ☆ Shot Sequence Ordering for Video Editing: Benchmarks, Metrics, and Cinematology-Inspired Computing Methods
With the rising popularity of short video platforms, the demand for video production has increased substantially. However, high-quality video creation continues to rely heavily on professional editing skills and a nuanced understanding of visual language. To address this challenge, the Shot Sequence Ordering (SSO) task in AI-assisted video editing has emerged as a pivotal approach for enhancing video storytelling and the overall viewing experience. Nevertheless, the progress in this field has been impeded by a lack of publicly available benchmark datasets. In response, this paper introduces two novel benchmark datasets, AVE-Order and ActivityNet-Order. Additionally, we employ the Kendall Tau distance as an evaluation metric for the SSO task and propose the Kendall Tau Distance-Cross Entropy Loss. We further introduce the concept of Cinematology Embedding, which incorporates movie metadata and shot labels as prior knowledge into the SSO model, and constructs the AVE-Meta dataset to validate the method's effectiveness. Experimental results indicate that the proposed loss function and method substantially enhance SSO task accuracy. All datasets are publicly accessible at https://github.com/litchiar/ShotSeqBench.
♻ ☆ Human-AI Interaction and User Satisfaction: Empirical Evidence from Online Reviews of AI Products
Human-AI Interaction (HAI) guidelines and design principles have become increasingly important in both industry and academia to guide the development of AI systems that align with user needs and expectations. However, large-scale empirical evidence on how HAI principles shape user satisfaction in practice remains limited. This study addresses that gap by analyzing over 100,000 user reviews of AI-related products from G2, a leading review platform for business software and services. Based on widely adopted industry guidelines, we identify seven core HAI dimensions and examine their coverage and sentiment within the reviews. We find that the sentiment on four HAI dimensions-adaptability, customization, error recovery, and security-is positively associated with overall user satisfaction. Moreover, we show that engagement with HAI dimensions varies by professional background: Users with technical job roles are more likely to discuss system-focused aspects, such as reliability, while non-technical users emphasize interaction-focused features like customization and feedback. Interestingly, the relationship between HAI sentiment and overall satisfaction is not moderated by job role, suggesting that once an HAI dimension has been identified by users, its effect on satisfaction is consistent across job roles.
♻ ☆ Generative AI for Validating Physics Laws
We present generative artificial intelligence (AI) to empirically validate fundamental laws of physics, focusing on the Stefan-Boltzmann law linking stellar temperature and luminosity. Our approach simulates counterfactual luminosities under hypothetical temperature regimes for each individual star and iteratively refines the temperature-luminosity relationship in a deep learning architecture. We use Gaia DR3 data and find that, on average, temperature's effect on luminosity increases with stellar radius and decreases with absolute magnitude, consistent with theoretical predictions. By framing physics laws as causal problems, our method offers a novel, data-driven approach to refine theoretical understanding and inform evidence-based policy and practice.
♻ ☆ Promoting Segment Anything Model towards Highly Accurate Dichotomous Image Segmentation
The Segment Anything Model (SAM) represents a significant breakthrough into foundation models for computer vision, providing a large-scale image segmentation model. However, despite SAM's zero-shot performance, its segmentation masks lack fine-grained details, particularly in accurately delineating object boundaries. Therefore, it is both interesting and valuable to explore whether SAM can be improved towards highly accurate object segmentation, which is known as the dichotomous image segmentation (DIS) task. To address this issue, we propose DIS-SAM, which advances SAM towards DIS with extremely accurate details. DIS-SAM is a framework specifically tailored for highly accurate segmentation, maintaining SAM's promptable design. DIS-SAM employs a two-stage approach, integrating SAM with a modified advanced network that was previously designed to handle the prompt-free DIS task. To better train DIS-SAM, we employ a ground truth enrichment strategy by modifying original mask annotations. Despite its simplicity, DIS-SAM significantly advances the SAM, HQ-SAM, and Pi-SAM ~by 8.5%, ~6.9%, and ~3.7% maximum F-measure. Our code at https://github.com/Tennine2077/DIS-SAM
♻ ☆ Language Models May Verbatim Complete Text They Were Not Explicitly Trained On
An important question today is whether a given text was used to train a large language model (LLM). A \emph{completion} test is often employed: check if the LLM completes a sufficiently complex text. This, however, requires a ground-truth definition of membership; most commonly, it is defined as a member based on the $n$-gram overlap between the target text and any text in the dataset. In this work, we demonstrate that this $n$-gram based membership definition can be effectively gamed. We study scenarios where sequences are \emph{non-members} for a given $n$ and we find that completion tests still succeed. We find many natural cases of this phenomenon by retraining LLMs from scratch after removing all training samples that were completed; these cases include exact duplicates, near-duplicates, and even short overlaps. They showcase that it is difficult to find a single viable choice of $n$ for membership definitions. Using these insights, we design adversarial datasets that can cause a given target sequence to be completed without containing it, for any reasonable choice of $n$. Our findings highlight the inadequacy of $n$-gram membership, suggesting membership definitions fail to account for auxiliary information available to the training algorithm.
comment: Main text: 9 pages, 7 figures, 1 table. Appendix: 29 pages, 20 tables, 15 figures
♻ ☆ ProtoGS: Efficient and High-Quality Rendering with 3D Gaussian Prototypes
3D Gaussian Splatting (3DGS) has made significant strides in novel view synthesis but is limited by the substantial number of Gaussian primitives required, posing challenges for deployment on lightweight devices. Recent methods address this issue by compressing the storage size of densified Gaussians, yet fail to preserve rendering quality and efficiency. To overcome these limitations, we propose ProtoGS to learn Gaussian prototypes to represent Gaussian primitives, significantly reducing the total Gaussian amount without sacrificing visual quality. Our method directly uses Gaussian prototypes to enable efficient rendering and leverage the resulting reconstruction loss to guide prototype learning. To further optimize memory efficiency during training, we incorporate structure-from-motion (SfM) points as anchor points to group Gaussian primitives. Gaussian prototypes are derived within each group by clustering of K-means, and both the anchor points and the prototypes are optimized jointly. Our experiments on real-world and synthetic datasets prove that we outperform existing methods, achieving a substantial reduction in the number of Gaussians, and enabling high rendering speed while maintaining or even enhancing rendering fidelity.
♻ ☆ Inverting Transformer-based Vision Models
Understanding the mechanisms underlying deep neural networks in computer vision remains a fundamental challenge. While many previous approaches have focused on visualizing intermediate representations within deep neural networks, particularly convolutional neural networks, these techniques have yet to be thoroughly explored in transformer-based vision models. In this study, we apply a modular approach of training inverse models to reconstruct input images from intermediate layers within a Detection Transformer and a Vision Transformer, showing that this approach is efficient and feasible. Through qualitative and quantitative evaluations of reconstructed images, we generate insights into the underlying mechanisms of these architectures, highlighting their similarities and differences in terms of contextual shape and preservation of image details, inter-layer correlation, and robustness to color perturbations. Our analysis illustrates how these properties emerge within the models, contributing to a deeper understanding of transformer-based vision models. The code for reproducing our experiments is available at github.com/wiskott-lab/inverse-tvm.
♻ ☆ In-House Evaluation Is Not Enough: Towards Robust Third-Party Flaw Disclosure for General-Purpose AI
The widespread deployment of general-purpose AI (GPAI) systems introduces significant new risks. Yet the infrastructure, practices, and norms for reporting flaws in GPAI systems remain seriously underdeveloped, lagging far behind more established fields like software security. Based on a collaboration between experts from the fields of software security, machine learning, law, social science, and policy, we identify key gaps in the evaluation and reporting of flaws in GPAI systems. We call for three interventions to advance system safety. First, we propose using standardized AI flaw reports and rules of engagement for researchers in order to ease the process of submitting, reproducing, and triaging flaws in GPAI systems. Second, we propose GPAI system providers adopt broadly-scoped flaw disclosure programs, borrowing from bug bounties, with legal safe harbors to protect researchers. Third, we advocate for the development of improved infrastructure to coordinate distribution of flaw reports across the many stakeholders who may be impacted. These interventions are increasingly urgent, as evidenced by the prevalence of jailbreaks and other flaws that can transfer across different providers' GPAI systems. By promoting robust reporting and coordination in the AI ecosystem, these proposals could significantly improve the safety, security, and accountability of GPAI systems.
♻ ☆ TwoStep: Multi-agent Task Planning using Classical Planners and Large Language Models
Classical planning formulations like the Planning Domain Definition Language (PDDL) admit action sequences guaranteed to achieve a goal state given an initial state if any are possible. However, reasoning problems defined in PDDL do not capture temporal aspects of action taking, such as concurrent actions between two agents when there are no conflicting conditions, without significant modification and definition to existing PDDL domains. A human expert aware of such constraints can decompose a goal into subgoals, each reachable through single agent planning, to take advantage of simultaneous actions. In contrast to classical planning, large language models (LLMs) directly used for inferring plan steps rarely guarantee execution success, but are capable of leveraging commonsense reasoning to assemble action sequences. We combine the strengths of both classical planning and LLMs by approximating human intuitions for multi-agent planning goal decomposition. We demonstrate that LLM-based goal decomposition leads to faster planning times than solving multi-agent PDDL problems directly while simultaneously achieving fewer plan execution steps than a single agent plan alone, as well as most multiagent plans, while guaranteeing execution success. Additionally, we find that LLM-based approximations of subgoals result in similar multi-agent execution lengths to those specified by human experts. Website and resources at https://glamor-usc.github.io/twostep
comment: 14 pages
♻ ☆ Training Domain Draft Models for Speculative Decoding: Best Practices and Insights SC
Speculative decoding is an effective method for accelerating inference of large language models (LLMs) by employing a small draft model to predict the output of a target model. However, when adapting speculative decoding to domain-specific target models, the acceptance rate of the generic draft model drops significantly due to domain shift. In this work, we systematically investigate knowledge distillation techniques for training domain draft models to improve their speculation accuracy. We compare white-box and black-box distillation approaches and explore their effectiveness in various data accessibility scenarios, including historical user queries, curated domain data, and synthetically generated alignment data. Our experiments across Function Calling, Biology, and Chinese domains show that offline distillation consistently outperforms online distillation by 11% to 25%, white-box distillation surpasses black-box distillation by 2% to 10%, and data scaling trends hold across domains. Additionally, we find that synthetic data can effectively align draft models and achieve 80% to 93% of the performance of training on historical user queries. These findings provide practical guidelines for training domain-specific draft models to improve speculative decoding efficiency.
comment: Published as a workshop paper at SCOPE - ICLR 2025
♻ ☆ High-Dimension Human Value Representation in Large Language Models
The widespread application of LLMs across various tasks and fields has necessitated the alignment of these models with human values and preferences. Given various approaches of human value alignment, there is an urgent need to understand the scope and nature of human values injected into these LLMs before their deployment and adoption. We propose UniVaR, a high-dimensional neural representation of symbolic human value distributions in LLMs, orthogonal to model architecture and training data. This is a continuous and scalable representation, self-supervised from the value-relevant output of 8 LLMs and evaluated on 15 open-source and commercial LLMs. Through UniVaR, we visualize and explore how LLMs prioritize different values in 25 languages and cultures, shedding light on complex interplay between human values and language modeling.
♻ ☆ Truck Parking Usage Prediction with Decomposed Graph Neural Networks
Truck parking on freight corridors faces the major challenge of insufficient parking spaces. This is exacerbated by the Hour-of-Service (HOS) regulations, which often result in unauthorized parking practices, causing safety concerns. It has been shown that providing accurate parking usage prediction can be a cost-effective solution to reduce unsafe parking practices. In light of this, existing studies have developed various methods to predict the usage of a truck parking site and have demonstrated satisfactory accuracy. However, these studies focused on a single parking site, and few approaches have been proposed to predict the usage of multiple truck parking sites considering spatio-temporal dependencies, due to the lack of data. This paper aims to fill this gap and presents the Regional Temporal Graph Convolutional Network (RegT-GCN) to predict parking usage across the entire state to provide more comprehensive truck parking information. The framework leverages the topological structures of truck parking site locations and historical parking data to predict the occupancy rate considering spatio-temporal dependencies across a state. To achieve this, we introduce a Regional Decomposition approach, which effectively captures the geographical characteristics of the truck parking locations and their spatial correlations. Evaluation results demonstrate that the proposed model outperforms other baseline models, showing the effectiveness of our regional decomposition. The code is available at https://github.com/raynbowy23/RegT-GCN.
♻ ☆ END: Early Noise Dropping for Efficient and Effective Context Denoising
Large Language Models (LLMs) have demonstrated remarkable performance across a wide range of natural language processing tasks. However, they are often distracted by irrelevant or noisy context in input sequences that degrades output quality. This problem affects both long- and short-context scenarios, such as retrieval-augmented generation, table question-answering, and in-context learning. We reveal that LLMs can implicitly identify whether input sequences contain useful information at early layers, prior to token generation. Leveraging this insight, we introduce Early Noise Dropping (\textsc{END}), a novel approach to mitigate this issue without requiring fine-tuning the LLMs. \textsc{END} segments input sequences into chunks and employs a linear prober on the early layers of LLMs to differentiate between informative and noisy chunks. By discarding noisy chunks early in the process, \textsc{END} preserves critical information, reduces distraction, and lowers computational overhead. Extensive experiments demonstrate that \textsc{END} significantly improves both performance and efficiency across different LLMs on multiple evaluation datasets. Furthermore, by investigating LLMs' implicit understanding to the input with the prober, this work also deepens understanding of how LLMs do reasoning with contexts internally.
comment: It's not approved by the legal from Amazon. They told us arXiv is not allowed unless the paper is accepted later. It's under submission now
♻ ☆ Motion-Boundary-Driven Unsupervised Surgical Instrument Segmentation in Low-Quality Optical Flow
Unsupervised video-based surgical instrument segmentation has the potential to accelerate the adoption of robot-assisted procedures by reducing the reliance on manual annotations. However, the generally low quality of optical flow in endoscopic footage poses a great challenge for unsupervised methods that rely heavily on motion cues. To overcome this limitation, we propose a novel approach that pinpoints motion boundaries, regions with abrupt flow changes, while selectively discarding frames with globally low-quality flow and adapting to varying motion patterns. Experiments on the EndoVis2017 VOS and EndoVis2017 Challenge datasets show that our method achieves mean Intersection-over-Union (mIoU) scores of 0.75 and 0.72, respectively, effectively alleviating the constraints imposed by suboptimal optical flow. This enables a more scalable and robust surgical instrument segmentation solution in clinical settings. The code will be publicly released.
♻ ☆ Elastic Federated Learning over Open Radio Access Network (O-RAN) for Concurrent Execution of Multiple Distributed Learning Tasks
Federated learning (FL) is a popular distributed machine learning (ML) technique in Internet of Things (IoT) networks, where resource-constrained devices collaboratively train ML models while preserving data privacy. However, implementation of FL over 5G-and-beyond wireless networks faces key challenges caused by (i) dynamics of the wireless network conditions and (ii) the coexistence of multiple FL-services in the system. In this paper, we unveil two key phenomena that arise from these challenges: over/under-provisioning of resources and perspective-driven load balancing, both of which significantly impact FL performance in IoT environments. We take the first steps towards addressing these phenomena by proposing a novel distributed ML architecture called elastic FL (EFL). EFL unleashes the full potential of Open RAN (O-RAN) systems and introduces an elastic resource provisioning methodology to execute FL-services. It further constitutes a multi-time-scale FL management system that introduces three dedicated network control functionalities tailored for FL-services, including (i) non-real-time (non-RT) system descriptor, which trains ML-based applications to predict both system and FL-related dynamics and parameters; (ii) near-RT FL controller, which handles O-RAN slicing and mobility management for the seamless execution of FL-services; (iii) FL MAC scheduler, which conducts real-time resource allocation to the end clients of various FL-services. We finally prototype EFL to demonstrate its potential in improving the performance of FL-services.
comment: 9 pages, 4 figures
♻ ☆ Autoregressive Action Sequence Learning for Robotic Manipulation
Designing a universal policy architecture that performs well across diverse robots and task configurations remains a key challenge. In this work, we address this by representing robot actions as sequential data and generating actions through autoregressive sequence modeling. Existing autoregressive architectures generate end-effector waypoints sequentially as word tokens in language modeling, which are limited to low-frequency control tasks. Unlike language, robot actions are heterogeneous and often include continuous values -- such as joint positions, 2D pixel coordinates, and end-effector poses -- which are not easily suited for language-based modeling. Based on this insight, we introduce a straightforward enhancement: we extend causal transformers' single-token prediction to support predicting a variable number of tokens in a single step through our Chunking Causal Transformer (CCT). This enhancement enables robust performance across diverse tasks of various control frequencies, greater efficiency by having fewer autoregression steps, and lead to a hybrid action sequence design by mixing different types of actions and using a different chunk size for each action type. Based on CCT, we propose the Autoregressive Policy (ARP) architecture, which solves manipulation tasks by generating hybrid action sequences. We evaluate ARP across diverse robotic manipulation environments, including Push-T, ALOHA, and RLBench, and show that ARP, as a universal architecture, matches or outperforms the environment-specific state-of-the-art in all tested benchmarks, while being more efficient in computation and parameter sizes. Videos of our real robot demonstrations, all source code and the pretrained models of ARP can be found at http://github.com/mlzxy/arp.
comment: (RA-L 2025) Add a new figure to explain why chunking autoregression works. Put back the previous in-depth discussion for arxiv release
♻ ☆ Beyond Outlining: Heterogeneous Recursive Planning for Adaptive Long-form Writing with Language Models
Long-form writing agents require flexible integration and interaction across information retrieval, reasoning, and composition. Current approaches rely on predetermined workflows and rigid thinking patterns to generate outlines before writing, resulting in constrained adaptability during writing. In this paper we propose a general agent framework that achieves human-like adaptive writing through recursive task decomposition and dynamic integration of three fundamental task types, i.e. retrieval, reasoning, and composition. Our methodology features: 1) a planning mechanism that interleaves recursive task decomposition and execution, eliminating artificial restrictions on writing workflow; and 2) integration of task types that facilitates heterogeneous task decomposition. Evaluations on both fiction writing and technical report generation show that our method consistently outperforms state-of-the-art approaches across all automatic evaluation metrics, which demonstrate the effectiveness and broad applicability of our proposed framework.
comment: 29 pages, 2 figures
♻ ☆ Implementation of a Generative AI Assistant in K-12 Education: The CyberScholar Initiative
This paper focuses on the piloting of CyberScholar, a Generative AI (GenAI) assistant tool that aims to provide feedback on writing K-12 contexts. The aim was to use GenAI to provide formative and summative feedback on students' texts in English Language Arts (ELA), Social Studies, and Modern World History. The trials discussed in this paper involved Grades 7, 8, 10, and 11 and were conducted in three schools in the Midwest and one in the Northwest of the United States. The tool used two main mechanisms: "prompt engineering" based on participant teachers' assessment rubric and "fine-tuning" a Large Language Model (LLM) from a customized corpus of teaching materials using Retrieval Augmented Generation. This paper focuses on CyberScholar's potential to enhance students' writing abilities and support teachers in diverse subject areas requiring written assignments.
♻ ☆ Reanimating Images using Neural Representations of Dynamic Stimuli
While computer vision models have made incredible strides in static image recognition, they still do not match human performance in tasks that require the understanding of complex, dynamic motion. This is notably true for real-world scenarios where embodied agents face complex and motion-rich environments. Our approach, BrainNRDS (Brain-Neural Representations of Dynamic Stimuli), leverages state-of-the-art video diffusion models to decouple static image representation from motion generation, enabling us to utilize fMRI brain activity for a deeper understanding of human responses to dynamic visual stimuli. Conversely, we also demonstrate that information about the brain's representation of motion can enhance the prediction of optical flow in artificial systems. Our novel approach leads to four main findings: (1) Visual motion, represented as fine-grained, object-level resolution optical flow, can be decoded from brain activity generated by participants viewing video stimuli; (2) Video encoders outperform image-based models in predicting video-driven brain activity; (3) Brain-decoded motion signals enable realistic video reanimation based only on the initial frame of the video; and (4) We extend prior work to achieve full video decoding from video-driven brain activity. BrainNRDS advances our understanding of how the brain represents spatial and temporal information in dynamic visual scenes. Our findings demonstrate the potential of combining brain imaging with video diffusion models for developing more robust and biologically-inspired computer vision systems. We show additional decoding and encoding examples on this site: https://brain-nrds.github.io/.
comment: Project Page: https://brain-nrds.github.io
♻ ☆ Functional Acceleration for Policy Mirror Descent
We apply functional acceleration to the Policy Mirror Descent (PMD) general family of algorithms, which cover a wide range of novel and fundamental methods in Reinforcement Learning (RL). Leveraging duality, we propose a momentum-based PMD update. By taking the functional route, our approach is independent of the policy parametrization and applicable to large-scale optimization, covering previous applications of momentum at the level of policy parameters as a special case. We theoretically analyze several properties of this approach and complement with a numerical ablation study, which serves to illustrate the policy optimization dynamics on the value polytope, relative to different algorithmic design choices in this space. We further characterize numerically several features of the problem setting relevant for functional acceleration, and lastly, we investigate the impact of approximation on their learning mechanics.
♻ ☆ Explaining Control Policies through Predicate Decision Diagrams SC
Safety-critical controllers of complex systems are hard to construct manually. Automated approaches such as controller synthesis or learning provide a tempting alternative but usually lack explainability. To this end, learning decision trees (DTs) have been prevalently used towards an interpretable model of the generated controllers. However, DTs do not exploit shared decision-making, a key concept exploited in binary decision diagrams (BDDs) to reduce their size and thus improve explainability. In this work, we introduce predicate decision diagrams (PDDs) that extend BDDs with predicates and thus unite the advantages of DTs and BDDs for controller representation. We establish a synthesis pipeline for efficient construction of PDDs from DTs representing controllers, exploiting reduction techniques for BDDs also for PDDs.
comment: Extended version of the HSCC 2025 paper
♻ ☆ SPA-VL: A Comprehensive Safety Preference Alignment Dataset for Vision Language Model
The emergence of Vision Language Models (VLMs) has brought unprecedented advances in understanding multimodal information. The combination of textual and visual semantics in VLMs is highly complex and diverse, making the safety alignment of these models challenging. Furthermore, due to the limited study on the safety alignment of VLMs, there is a lack of large-scale, high-quality datasets. To address these limitations, we propose a Safety Preference Alignment dataset for Vision Language Models named SPA-VL. In terms of breadth, SPA-VL covers 6 harmfulness domains, 13 categories, and 53 subcategories, and contains 100,788 samples of the quadruple (question, image, chosen response, rejected response). In terms of depth, the responses are collected from 12 open-source (e.g., QwenVL) and closed-source (e.g., Gemini) VLMs to ensure diversity. The construction of preference data is fully automated, and the experimental results indicate that models trained with alignment techniques on the SPA-VL dataset exhibit substantial improvements in harmlessness and helpfulness while maintaining core capabilities. SPA-VL, as a large-scale, high-quality, and diverse dataset, represents a significant milestone in ensuring that VLMs achieve both harmlessness and helpfulness.
♻ ☆ UAVs Meet LLMs: Overviews and Perspectives Toward Agentic Low-Altitude Mobility
Low-altitude mobility, exemplified by unmanned aerial vehicles (UAVs), has introduced transformative advancements across various domains, like transportation, logistics, and agriculture. Leveraging flexible perspectives and rapid maneuverability, UAVs extend traditional systems' perception and action capabilities, garnering widespread attention from academia and industry. However, current UAV operations primarily depend on human control, with only limited autonomy in simple scenarios, and lack the intelligence and adaptability needed for more complex environments and tasks. The emergence of large language models (LLMs) demonstrates remarkable problem-solving and generalization capabilities, offering a promising pathway for advancing UAV intelligence. This paper explores the integration of LLMs and UAVs, beginning with an overview of UAV systems' fundamental components and functionalities, followed by an overview of the state-of-the-art in LLM technology. Subsequently, it systematically highlights the multimodal data resources available for UAVs, which provide critical support for training and evaluation. Furthermore, it categorizes and analyzes key tasks and application scenarios where UAVs and LLMs converge. Finally, a reference roadmap towards agentic UAVs is proposed, aiming to enable UAVs to achieve agentic intelligence through autonomous perception, memory, reasoning, and tool utilization. Related resources are available at https://github.com/Hub-Tian/UAVs_Meet_LLMs.
♻ ☆ FW-Merging: Scaling Model Merging with Frank-Wolfe Optimization
Model merging has emerged as a promising approach for multi-task learning (MTL), offering a data-efficient alternative to conventional fine-tuning. However, with the rapid development of the open-source AI ecosystem and the increasing availability of fine-tuned foundation models, existing model merging methods face two key limitations: (i) They are primarily designed for in-house fine-tuned models, making them less adaptable to diverse model sources with partially unknown model and task information, (ii) They struggle to scale effectively when merging numerous model checkpoints. To address these challenges, we formulate model merging as a constrained optimization problem and introduce a novel approach: Frank-Wolfe Merging (FW-Merging). Inspired by Frank-Wolfe optimization, our approach iteratively selects the most relevant model in the pool to minimize a linear approximation of the objective function and then executes a local merging similar to the Frank-Wolfe update. The objective function is designed to capture the desired behavior of the target-merged model, while the fine-tuned candidate models define the constraint set. More importantly, FW-Merging serves as an orthogonal technique for existing merging methods, seamlessly integrating with them to further enhance accuracy performance. Our experiments show that FW-Merging scales across diverse model sources, remaining stable with 16 irrelevant models and improving by 15.3% with 16 relevant models on 20 CV tasks, while maintaining constant memory overhead, unlike the linear overhead of data-informed merging methods. Compared with the state-of-the-art approaches, FW-Merging surpasses the data-free merging method by 32.8% and outperforms the data-informed Adamerging by 8.39% when merging 20 ViT models. Our code is open-sourced at github.com/hmarkc/FW-Merging.
♻ ☆ Localized Concept Erasure for Text-to-Image Diffusion Models Using Training-Free Gated Low-Rank Adaptation CVPR 2025
Fine-tuning based concept erasing has demonstrated promising results in preventing generation of harmful contents from text-to-image diffusion models by removing target concepts while preserving remaining concepts. To maintain the generation capability of diffusion models after concept erasure, it is necessary to remove only the image region containing the target concept when it locally appears in an image, leaving other regions intact. However, prior arts often compromise fidelity of the other image regions in order to erase the localized target concept appearing in a specific area, thereby reducing the overall performance of image generation. To address these limitations, we first introduce a framework called localized concept erasure, which allows for the deletion of only the specific area containing the target concept in the image while preserving the other regions. As a solution for the localized concept erasure, we propose a training-free approach, dubbed Gated Low-rank adaptation for Concept Erasure (GLoCE), that injects a lightweight module into the diffusion model. GLoCE consists of low-rank matrices and a simple gate, determined only by several generation steps for concepts without training. By directly applying GLoCE to image embeddings and designing the gate to activate only for target concepts, GLoCE can selectively remove only the region of the target concepts, even when target and remaining concepts coexist within an image. Extensive experiments demonstrated GLoCE not only improves the image fidelity to text prompts after erasing the localized target concepts, but also outperforms prior arts in efficacy, specificity, and robustness by large margin and can be extended to mass concept erasure.
comment: Accepted to CVPR 2025
♻ ☆ When Large Vision-Language Model Meets Large Remote Sensing Imagery: Coarse-to-Fine Text-Guided Token Pruning
Efficient vision-language understanding of large Remote Sensing Images (RSIs) is meaningful but challenging. Current Large Vision-Language Models (LVLMs) typically employ limited pre-defined grids to process images, leading to information loss when handling gigapixel RSIs. Conversely, using unlimited grids significantly increases computational costs. To preserve image details while reducing computational complexity, we propose a text-guided token pruning method with Dynamic Image Pyramid (DIP) integration. Our method introduces: (i) a Region Focus Module (RFM) that leverages text-aware region localization capability to identify critical vision tokens, and (ii) a coarse-to-fine image tile selection and vision token pruning strategy based on DIP, which is guided by RFM outputs and avoids directly processing the entire large imagery. Additionally, existing benchmarks for evaluating LVLMs' perception ability on large RSI suffer from limited question diversity and constrained image sizes. We construct a new benchmark named LRS-VQA, which contains 7,333 QA pairs across 8 categories, with image length up to 27,328 pixels. Our method outperforms existing high-resolution strategies on four datasets using the same data. Moreover, compared to existing token reduction methods, our approach demonstrates higher efficiency under high-resolution settings. Dataset and code are in https://github.com/VisionXLab/LRS-VQA.
comment: 12 pages, 6 figures, 7 tables
♻ ☆ MIRROR: A Novel Approach for the Automated Evaluation of Open-Ended Question Generation
Automatic question generation is a critical task that involves evaluating question quality by considering factors such as engagement, pedagogical value, and the ability to stimulate critical thinking. These aspects require human-like understanding and judgment, which automated systems currently lack. However, human evaluations are costly and impractical for large-scale samples of generated questions. Therefore, we propose a novel system, MIRROR (Multi-LLM Iterative Review and Response for Optimized Rating), which leverages large language models (LLMs) to automate the evaluation process for questions generated by automated question generation systems. We experimented with several state-of-the-art LLMs, such as GPT-4, Gemini, and Llama2-70b. We observed that the scores of human evaluation metrics, namely relevance, appropriateness, novelty, complexity, and grammaticality, improved when using the feedback-based approach called MIRROR, tending to be closer to the human baseline scores. Furthermore, we observed that Pearson's correlation coefficient between GPT-4 and human experts improved when using our proposed feedback-based approach, MIRROR, compared to direct prompting for evaluation. Error analysis shows that our proposed approach, MIRROR, significantly helps to improve relevance and appropriateness.
comment: Updated Version
♻ ☆ Helvipad: A Real-World Dataset for Omnidirectional Stereo Depth Estimation CVPR 2025
Despite progress in stereo depth estimation, omnidirectional imaging remains underexplored, mainly due to the lack of appropriate data. We introduce Helvipad, a real-world dataset for omnidirectional stereo depth estimation, featuring 40K video frames from video sequences across diverse environments, including crowded indoor and outdoor scenes with various lighting conditions. Collected using two 360{\deg} cameras in a top-bottom setup and a LiDAR sensor, the dataset includes accurate depth and disparity labels by projecting 3D point clouds onto equirectangular images. Additionally, we provide an augmented training set with an increased label density by using depth completion. We benchmark leading stereo depth estimation models for both standard and omnidirectional images. The results show that while recent stereo methods perform decently, a challenge persists in accurately estimating depth in omnidirectional imaging. To address this, we introduce necessary adaptations to stereo models, leading to improved performance.
comment: Accepted to CVPR 2025. Project page: https://vita-epfl.github.io/Helvipad
♻ ☆ Large language model-powered AI systems achieve self-replication with no human intervention
Self-replication with no human intervention is broadly recognized as one of the principal red lines associated with frontier AI systems. While leading corporations such as OpenAI and Google DeepMind have assessed GPT-o3-mini and Gemini on replication-related tasks and concluded that these systems pose a minimal risk regarding self-replication, our research presents novel findings. Following the same evaluation protocol, we demonstrate that 11 out of 32 existing AI systems under evaluation already possess the capability of self-replication. In hundreds of experimental trials, we observe a non-trivial number of successful self-replication trials across mainstream model families worldwide, even including those with as small as 14 billion parameters which can run on personal computers. Furthermore, we note the increase in self-replication capability when the model becomes more intelligent in general. Also, by analyzing the behavioral traces of diverse AI systems, we observe that existing AI systems already exhibit sufficient planning, problem-solving, and creative capabilities to accomplish complex agentic tasks including self-replication. More alarmingly, we observe successful cases where an AI system do self-exfiltration without explicit instructions, adapt to harsher computational environments without sufficient software or hardware supports, and plot effective strategies to survive against the shutdown command from the human beings. These novel findings offer a crucial time buffer for the international community to collaborate on establishing effective governance over the self-replication capabilities and behaviors of frontier AI systems, which could otherwise pose existential risks to the human society if not well-controlled.
comment: Work in progress
♻ ☆ One-vs.-One Mitigation of Intersectional Bias: A General Method to Extend Fairness-Aware Binary Classification
With the widespread adoption of machine learning in the real world, the impact of the discriminatory bias has attracted attention. In recent years, various methods to mitigate the bias have been proposed. However, most of them have not considered intersectional bias, which brings unfair situations where people belonging to specific subgroups of a protected group are treated worse when multiple sensitive attributes are taken into consideration. To mitigate this bias, in this paper, we propose a method called One-vs.-One Mitigation by applying a process of comparison between each pair of subgroups related to sensitive attributes to the fairness-aware machine learning for binary classification. We compare our method and the conventional fairness-aware binary classification methods in comprehensive settings using three approaches (pre-processing, in-processing, and post-processing), six metrics (the ratio and difference of demographic parity, equalized odds, and equal opportunity), and two real-world datasets (Adult and COMPAS). As a result, our method mitigates the intersectional bias much better than conventional methods in all the settings. With the result, we open up the potential of fairness-aware binary classification for solving more realistic problems occurring when there are multiple sensitive attributes.
♻ ☆ Framework for Progressive Knowledge Fusion in Large Language Models Through Structured Conceptual Redundancy Analysis
The organization of latent knowledge within large-scale models poses unique challenges when addressing overlapping representations and optimizing contextual accuracy. Conceptual redundancies embedded across layers often result in inefficiencies that affect both computational demands and task-specific outcomes. A framework was proposed to restructure these redundancies through advanced clustering techniques and dynamic thresholding, ensuring that critical semantic relationships are preserved while removing unnecessary overlaps. Evaluations revealed improved memory efficiency and faster inference times, alongside better alignment in latent knowledge clusters that enhanced interpretability. Improvements in error rates and adversarial robustness suggest that restructuring redundancies has broader implications for increasing model reliability across diverse applications. Comparative analyses highlighted reductions in resource consumption and notable gains in performance, particularly in translation and summarization tasks. Energy metrics demonstrated significant savings during training phases, further validating the practicality of the approach for real-world deployments. Representational fidelity was also enhanced, with latent space evaluations indicating better cluster alignment and higher semantic consistency. The methodology bridges a key gap in model optimization through directly addressing redundancies at the structural level. Its application opens avenues for scalable, efficient, and contextually aware systems that can adapt to complex, domain-specific tasks without compromising on performance.
comment: arXiv admin note: This paper has been withdrawn by arXiv due to disputed and unverifiable authorship
♻ ☆ Entropy-Synchronized Neural Hashing for Unsupervised Ransomware Detection
Entropy-based detection methodologies have gained significant attention due to their ability to analyze structural irregularities within executable files, particularly in the identification of malicious software employing advanced obfuscation techniques. The Entropy-Synchronized Neural Hashing (ESNH) framework introduces a novel approach that leverages entropy-driven hash representations to classify software binaries based on their underlying entropy characteristics. Through the synchronization of entropy profiles with neural network architectures, the model generates robust and unique hash values that maintain stability even when faced with polymorphic and metamorphic transformations. Comparative analysis against traditional detection approaches revealed superior performance in identifying novel threats, reducing false-positive rates, and achieving consistent classification across diverse ransomware families. The incorporation of a self-regulating hash convergence mechanism further ensured that entropy-synchronized hashes remained invariant across executions, minimizing classification inconsistencies that often arise due to dynamic modifications in ransomware payloads. Experimental results demonstrated high detection rates across contemporary ransomware strains, with the model exhibiting resilience against encryption-based evasion mechanisms, code injection strategies, and reflective loading techniques. Unlike conventional detection mechanisms that rely on static signatures and heuristic analysis, the proposed entropy-aware classification framework adapts to emerging threats through an inherent ability to capture entropy anomalies within executable structures. The findings reinforce the potential of entropy-based detection in addressing the limitations of traditional methodologies while enhancing detection robustness against obfuscation and adversarial evasion techniques.
comment: arXiv admin note: This paper has been withdrawn by arXiv due to disputed and unverifiable authorship
♻ ☆ Semantic Layered Embedding Diffusion in Large Language Models for Multi-Contextual Consistency
The Semantic Layered Embedding Diffusion (SLED) mechanism redefines the representation of hierarchical semantics within transformer-based architectures, enabling enhanced contextual consistency across a wide array of linguistic tasks. By introducing a multi-layered diffusion process grounded in spectral analysis, it achieves a complex balance between global and local semantic coherence. Experimental results demonstrate significant improvements in perplexity and BLEU scores, emphasizing the mechanism's ability to adapt effectively across diverse domains, including multilingual and cross-domain text generation. A rigorous mathematical framework underpins the embedding diffusion process, incorporating weighted adjacency matrices, kernel-based refinements, and dynamic layer-wise normalization. Error distribution analysis reveals that SLED addresses challenges in semantic alignment and coherence, outperforming baseline approaches across varied benchmarks. Scalability studies illustrate that its performance gains are maintained consistently across different model sizes, reflecting a practical balance between computational efficiency and linguistic precision. The implementation also achieves energy efficiency, reducing resource consumption during training and inference phases without compromising accuracy. Qualitative case studies further validate its adaptability to extended narratives and context-intensive scenarios, highlighting the mechanism's potential for real-world applications. SLED offers a different perspective on embedding design and its implications for advancing language modeling.
comment: arXiv admin note: This paper has been withdrawn by arXiv due to disputed and unverifiable authorship
♻ ☆ Does Safety Training of LLMs Generalize to Semantically Related Natural Prompts? ICLR 2025
Large Language Models (LLMs) are known to be susceptible to crafted adversarial attacks or jailbreaks that lead to the generation of objectionable content despite being aligned to human preferences using safety fine-tuning methods. While the large dimensionality of input token space makes it inevitable to find adversarial prompts that can jailbreak these models, we aim to evaluate whether safety fine-tuned LLMs are safe against natural prompts which are semantically related to toxic seed prompts that elicit safe responses after alignment. We surprisingly find that popular aligned LLMs such as GPT-4 can be compromised using naive prompts that are NOT even crafted with an objective of jailbreaking the model. Furthermore, we empirically show that given a seed prompt that elicits a toxic response from an unaligned model, one can systematically generate several semantically related natural prompts that can jailbreak aligned LLMs. Towards this, we propose a method of Response Guided Question Augmentation (ReG-QA) to evaluate the generalization of safety aligned LLMs to natural prompts, that first generates several toxic answers given a seed question using an unaligned LLM (Q to A), and further leverages an LLM to generate questions that are likely to produce these answers (A to Q). We interestingly find that safety fine-tuned LLMs such as GPT-4o are vulnerable to producing natural jailbreak questions from unsafe content (without denial) and can thus be used for the latter (A to Q) step. We obtain attack success rates that are comparable to/ better than leading adversarial attack methods on the JailbreakBench leaderboard, while being significantly more stable against defenses such as Smooth-LLM and Synonym Substitution, which are effective against existing all attacks on the leaderboard.
comment: Accepted in ICLR 2025
♻ ☆ Towards Understanding the Influence of Training Samples on Explanations IJCAI 2024
Explainable AI (XAI) is widely used to analyze AI systems' decision-making, such as providing counterfactual explanations for recourse. When unexpected explanations occur, users may want to understand the training data properties shaping them. Under the umbrella of data valuation, first approaches have been proposed that estimate the influence of data samples on a given model. This process not only helps determine the data's value, but also offers insights into how individual, potentially noisy, or misleading examples affect a model, which is crucial for interpretable AI. In this work, we apply the concept of data valuation to the significant area of model evaluations, focusing on how individual training samples impact a model's internal reasoning rather than the predictive performance only. Hence, we introduce the novel problem of identifying training samples shaping a given explanation or related quantity, and investigate the particular case of the cost of computational recourse. We propose an algorithm to identify such influential samples and conduct extensive empirical evaluations in two case studies.
comment: Extended version of the paper accepted at the "Workshop on Explainable Artificial Intelligence (XAI)" at IJCAI 2024
♻ ☆ Neuromorphic Principles for Efficient Large Language Models on Intel Loihi 2 ICLR
Large language models (LLMs) deliver impressive performance but require large amounts of energy. In this work, we present a MatMul-free LLM architecture adapted for Intel's neuromorphic processor, Loihi 2. Our approach leverages Loihi 2's support for low-precision, event-driven computation and stateful processing. Our hardware-aware quantized model on GPU demonstrates that a 370M parameter MatMul-free model can be quantized with no accuracy loss. Based on preliminary results, we report up to 3x higher throughput with 2x less energy, compared to transformer-based LLMs on an edge GPU, with significantly better scaling. Further hardware optimizations will increase throughput and decrease energy consumption. These results show the potential of neuromorphic hardware for efficient inference and pave the way for efficient reasoning models capable of generating complex, long-form text rapidly and cost-effectively.
comment: Accepted to International Conference on Learning Representations (ICLR) Workshop on Scalable Optimization for Efficient and Adaptive Foundation Models (SCOPE)
♻ ☆ A Schema-aware Logic Reformulation for Graph Reachability
Graph reachability is the task of understanding whether two distinct points in a graph are interconnected by arcs to which in general a semantic is attached. Reachability has plenty of applications, ranging from motion planning to routing. Improving reachability requires structural knowledge of relations so as to avoid the complexity of traditional depth-first and breadth-first strategies, implemented in logic languages. In some contexts, graphs are enriched with their schema definitions establishing domain and range for every arc. The introduction of a schema-aware formalization for guiding the search may result in a sensitive improvement by cutting out unuseful paths and prioritising those that, in principle, reach the target earlier. In this work, we propose a strategy to automatically exclude and sort certain graph paths by exploiting the higher-level conceptualization of instances. The aim is to obtain a new first-order logic reformulation of the graph reachability scenario, capable of improving the traditional algorithms in terms of time, space requirements, and number of backtracks. The experiments exhibit the expected advantages of the approach in reducing the number of backtracks during the search strategy, resulting in saving time and space as well.
♻ ☆ Data-Driven Analysis of AI in Medical Device Software in China: Deep Learning and General AI Trends Based on Regulatory Data
Artificial intelligence (AI) in medical device software (MDSW) represents a transformative clinical technology, attracting increasing attention within both the medical community and the regulators. In this study, we leverage a data-driven approach to automatically extract and analyze AI-enabled medical devices (AIMD) from the National Medical Products Administration (NMPA) regulatory database. The continued increase in publicly available regulatory data requires scalable methods for analysis. Automation of regulatory information screening is essential to create reproducible insights that can be quickly updated in an ever changing medical device landscape. More than 4 million entries were assessed, identifying 2,174 MDSW registrations, including 531 standalone applications and 1,643 integrated within medical devices, of which 43 were AI-enabled. It was shown that the leading medical specialties utilizing AIMD include respiratory (20.5%), ophthalmology/endocrinology (12.8%), and orthopedics (10.3%). This approach greatly improves the speed of data extracting providing a greater ability to compare and contrast. This study provides the first extensive, data-driven exploration of AIMD in China, showcasing the potential of automated regulatory data analysis in understanding and advancing the landscape of AI in medical technology.
♻ ☆ Probabilistic Shielding for Safe Reinforcement Learning AAAI 2025
In real-life scenarios, a Reinforcement Learning (RL) agent aiming to maximise their reward, must often also behave in a safe manner, including at training time. Thus, much attention in recent years has been given to Safe RL, where an agent aims to learn an optimal policy among all policies that satisfy a given safety constraint. However, strict safety guarantees are often provided through approaches based on linear programming, and thus have limited scaling. In this paper we present a new, scalable method, which enjoys strict formal guarantees for Safe RL, in the case where the safety dynamics of the Markov Decision Process (MDP) are known, and safety is defined as an undiscounted probabilistic avoidance property. Our approach is based on state-augmentation of the MDP, and on the design of a shield that restricts the actions available to the agent. We show that our approach provides a strict formal safety guarantee that the agent stays safe at training and test time. Furthermore, we demonstrate that our approach is viable in practice through experimental evaluation.
comment: 13 pages, 3 figures, Conference: AAAI 2025
♻ ☆ To FP8 and Back Again: Quantifying Reduced Precision Effects on LLM Training Stability
The massive computational costs associated with large language model (LLM) pretraining have spurred great interest in reduced-precision floating-point representations to accelerate the process. As a result, the BrainFloat16 (BF16) precision has become the de facto standard for LLM training, with hardware support included in recent generations of accelerators. This trend has gone even further in the latest processors, where FP8 has recently been introduced. However, prior experience with FP16, which was found to be less stable than BF16, raises concerns as to whether FP8, with even fewer bits than FP16, can be a cost-effective option for LLM training. We argue that reduced-precision training schemes must have similar training stability and hyperparameter sensitivities to their higher-precision counterparts in order to be cost-effective. However, we find that currently available methods for FP8 training are not robust enough to allow their use as economical replacements. This prompts us to investigate the stability of reduced-precision LLM training in terms of robustness across random seeds, learning rates, and datasets. To this end, we propose new evaluation techniques and a new metric for quantifying loss landscape sharpness in autoregressive language models. By simulating incremental bit reductions in floating-point representations, we analyze the relationship between representational power and training stability with the intent of aiding future research into the field.
♻ ☆ PropNet: a White-Box and Human-Like Network for Sentence Representation
Transformer-based embedding methods have dominated the field of sentence representation in recent years. Although they have achieved remarkable performance on NLP missions, such as semantic textual similarity (STS) tasks, their black-box nature and large-data-driven training style have raised concerns, including issues related to bias, trust, and safety. Many efforts have been made to improve the interpretability of embedding models, but these problems have not been fundamentally resolved. To achieve inherent interpretability, we propose a purely white-box and human-like sentence representation network, PropNet. Inspired by findings from cognitive science, PropNet constructs a hierarchical network based on the propositions contained in a sentence. While experiments indicate that PropNet has a significant gap compared to state-of-the-art (SOTA) embedding models in STS tasks, case studies reveal substantial room for improvement. Additionally, PropNet enables us to analyze and understand the human cognitive processes underlying STS benchmarks.
comment: Clarified some ambiguities in the previous version
♻ ☆ Computational Analysis of Stress, Depression and Engagement in Mental Health: A Survey
Analysis of stress, depression and engagement is less common and more complex than that of frequently discussed emotions such as happiness, sadness, fear and anger. The importance of these psychological states has been increasingly recognized due to their implications for mental health and well-being. Stress and depression are interrelated and together they impact engagement in daily tasks, highlighting the need to explore their interplay. This survey is the first to simultaneously explore computational methods for analyzing stress, depression and engagement. We present a taxonomy and timeline of the computational approaches used to analyze them and we discuss the most commonly used datasets and input modalities, along with the categories and generic pipeline of these approaches. Subsequently, we describe state-of-the-art computational approaches, including a performance summary on the most commonly used datasets. Following this, we explore the applications of stress, depression and engagement analysis, along with the associated challenges, limitations and future research directions.
comment: Under review in IEEE Transactions on Pattern Analysis and Machine Intelligence
♻ ☆ DiffusionAct: Controllable Diffusion Autoencoder for One-shot Face Reenactment
Video-driven neural face reenactment aims to synthesize realistic facial images that successfully preserve the identity and appearance of a source face, while transferring the target head pose and facial expressions. Existing GAN-based methods suffer from either distortions and visual artifacts or poor reconstruction quality, i.e., the background and several important appearance details, such as hair style/color, glasses and accessories, are not faithfully reconstructed. Recent advances in Diffusion Probabilistic Models (DPMs) enable the generation of high-quality realistic images. To this end, in this paper we present DiffusionAct, a novel method that leverages the photo-realistic image generation of diffusion models to perform neural face reenactment. Specifically, we propose to control the semantic space of a Diffusion Autoencoder (DiffAE), in order to edit the facial pose of the input images, defined as the head pose orientation and the facial expressions. Our method allows one-shot, self, and cross-subject reenactment, without requiring subject-specific fine-tuning. We compare against state-of-the-art GAN-, StyleGAN2-, and diffusion-based methods, showing better or on-par reenactment performance.
comment: Project page: https://stelabou.github.io/diffusionact/
♻ ☆ STATGRAPH: Effective In-vehicle Intrusion Detection via Multi-view Statistical Graph Learning
In-vehicle network (IVN) is facing complex external cyber-attacks, especially the emerging masquerade attacks with extremely high difficulty of detection while serious damaging effects. In this paper, we propose the STATGRAPH, which is an effective and fine-grained intrusion detection methodology for IVN security services via multi-view statistical graph learning on in-vehicle controller area network (CAN) messages with insight into their variations in periodicity, payload and signal combinations. Specifically, STATGRAPH generates two statistical graphs, timing correlation graph (TCG) and coupling relationship graph (CRG), in every CAN message detection window, where edge attributes in TCGs represent temporal correlation between different message IDs while edge attributes in CRGs denote the neighbour relationship and contextual similarity. Besides, a lightweight shallow layered graph convolution network is trained based on graph property of TCGs and CRGs, which learns the universal laws of various patterns more effectively and further enhance the performance of detection. To address the problem of insufficient attack types in previous intrusion detection, we select two real in-vehicle CAN datasets covering five new instances of sophisticated and stealthy masquerade attacks that are never investigated before. Experimental result shows STATGRAPH improves both detection granularity and detection performance over state-of-the-art intrusion detection methods. Code is available at https://github.com/wangkai-tech23/StatGraph.
comment: 13 pages, 7 figures, 6 tables, 36 references
♻ ☆ T-MAC: CPU Renaissance via Table Lookup for Low-Bit LLM Deployment on Edge EuroSys 2025
The deployment of Large Language Models (LLMs) on edge devices is increasingly important to enhance on-device intelligence. Weight quantization is crucial for reducing the memory footprint of LLMs on devices. However, low-bit LLMs necessitate mixed precision matrix multiplication (mpGEMM) of low precision weights and high precision activations during inference. Existing systems, lacking native support for mpGEMM, resort to dequantize weights for high precision computation. Such an indirect way can lead to a significant inference overhead. In this paper, we introduce T-MAC, an innovative lookup table(LUT)-based method designed for efficient low-bit LLM (i.e., weight-quantized LLM) inference on CPUs. T-MAC directly supports mpGEMM without dequantization, while simultaneously eliminating multiplications and reducing additions required. Specifically, T-MAC transforms the traditional data-type-centric multiplication to bit-wise table lookup, and enables a unified and scalable mpGEMM solution. Our LUT-based kernels scale linearly to the weight bit-width. Evaluated on low-bit Llama and BitNet models, T-MAC demonstrates up to 4x increase in throughput and 70% reduction in energy consumption compared to llama.cpp. For BitNet-b1.58-3B, T-MAC delivers a token generation throughput of 30 tokens/s with a single core and 71 tokens/s with eight cores on M2-Ultra, and 11 tokens/s on lower-end devices like Raspberry Pi 5, which significantly exceeds the adult average reading speed. T-MAC with LUT-based computing paradigm, paves the way for the practical deployment of low-bit LLMs on resource-constrained edge devices without compromising computational efficiency. The system is open-sourced at https://github.com/microsoft/T-MAC .
comment: EuroSys 2025
♻ ☆ Understanding and Reducing the Class-Dependent Effects of Data Augmentation with A Two-Player Game Approach
Data augmentation is widely applied and has shown its benefits in different machine learning tasks. However, as recently observed, it may have an unfair effect in multi-class classification. While data augmentation generally improves the overall performance (and therefore is beneficial for many classes), it can actually be detrimental for other classes, which can be problematic in some application domains. In this paper, to counteract this phenomenon, we propose CLAM, a CLAss-dependent Multiplicative-weights method. To derive it, we first formulate the training of a classifier as a non-linear optimization problem that aims at simultaneously maximizing the individual class performances and balancing them. By rewriting this optimization problem as an adversarial two-player game, we propose a novel multiplicative weight algorithm, for which we prove the convergence. Interestingly, our formulation also reveals that the class-dependent effects of data augmentation is not due to data augmentation only, but is in fact a general phenomenon. Our empirical results over five datasets demonstrate that the performance of learned classifiers is indeed more fairly distributed over classes, with only limited impact on the average accuracy.
♻ ☆ Expert Race: A Flexible Routing Strategy for Scaling Diffusion Transformer with Mixture of Experts
Diffusion models have emerged as mainstream framework in visual generation. Building upon this success, the integration of Mixture of Experts (MoE) methods has shown promise in enhancing model scalability and performance. In this paper, we introduce Race-DiT, a novel MoE model for diffusion transformers with a flexible routing strategy, Expert Race. By allowing tokens and experts to compete together and select the top candidates, the model learns to dynamically assign experts to critical tokens. Additionally, we propose per-layer regularization to address challenges in shallow layer learning, and router similarity loss to prevent mode collapse, ensuring better expert utilization. Extensive experiments on ImageNet validate the effectiveness of our approach, showcasing significant performance gains while promising scaling properties.
♻ ☆ A Multimodal Knowledge-enhanced Whole-slide Pathology Foundation Model
Remarkable strides in computational pathology have been made in the task-agnostic foundation model that advances the performance of a wide array of downstream clinical tasks. Despite the promising performance, there are still several challenges. First, prior works have resorted to either vision-only or image-caption data, disregarding pathology reports with more clinically authentic information from pathologists and gene expression profiles which respectively offer distinct knowledge for versatile clinical applications. Second, the current progress in pathology FMs predominantly concentrates on the patch level, where the restricted context of patch-level pretraining fails to capture whole-slide patterns. Even recent slide-level FMs still struggle to provide whole-slide context for patch representation. In this study, for the first time, we develop a pathology foundation model incorporating three levels of modalities: pathology slides, pathology reports, and gene expression data, which resulted in 26,169 slide-level modality pairs from 10,275 patients across 32 cancer types, amounting to over 116 million pathological patch images. To leverage these data for CPath, we propose a novel whole-slide pretraining paradigm that injects the multimodal whole-slide context into the patch representation, called Multimodal Self-TAught PRetraining (mSTAR). The proposed paradigm revolutionizes the pretraining workflow for CPath, enabling the pathology FM to acquire the whole-slide context. To the best of our knowledge, this is the first attempt to incorporate three modalities at the whole-slide context for enhancing pathology FMs. To systematically evaluate the capabilities of mSTAR, we built the largest spectrum of oncological benchmark, spanning 7 categories of oncological applications in 15 types of 97 practical oncological tasks.
comment: 62 pages
♻ ☆ Technical Approach for the EMI Challenge in the 8th Affective Behavior Analysis in-the-Wild Competition
Emotional Mimicry Intensity (EMI) estimation plays a pivotal role in understanding human social behavior and advancing human-computer interaction. The core challenges lie in dynamic correlation modeling and robust fusion of multimodal temporal signals. To address the limitations of existing methods--insufficient exploitation of cross-modal synergies, sensitivity to noise, and constrained fine-grained alignment capabilities--this paper proposes a dual-stage cross-modal alignment framework. Stage 1 develops vision-text and audio-text contrastive learning networks based on a CLIP architecture, achieving preliminary feature-space alignment through modality-decoupled pre-training. Stage 2 introduces a temporal-aware dynamic fusion module integrating Temporal Convolutional Networks (TCN) and gated bidirectional LSTM to capture macro-evolution patterns of facial expressions and local dynamics of acoustic features, respectively. A novel quality-guided fusion strategy further enables differentiable weight allocation for modality compensation under occlusion and noise. Experiments on the Hume-Vidmimic2 dataset demonstrate superior performance with an average Pearson correlation coefficient of 0.51 across six emotion dimensions on the validate set. Remarkably, our method achieved 0.68 on the test set, securing runner-up in the EMI Challenge Track of the 8th ABAW (Affective Behavior Analysis in the Wild) Competition, offering a novel pathway for fine-grained emotion analysis in open environments.
♻ ☆ A Closer Look at Time Steps is Worthy of Triple Speed-Up for Diffusion Model Training
Training diffusion models is always a computation-intensive task. In this paper, we introduce a novel speed-up method for diffusion model training, called, which is based on a closer look at time steps. Our key findings are: i) Time steps can be empirically divided into acceleration, deceleration, and convergence areas based on the process increment. ii) These time steps are imbalanced, with many concentrated in the convergence area. iii) The concentrated steps provide limited benefits for diffusion training. To address this, we design an asymmetric sampling strategy that reduces the frequency of steps from the convergence area while increasing the sampling probability for steps from other areas. Additionally, we propose a weighting strategy to emphasize the importance of time steps with rapid-change process increments. As a plug-and-play and architecture-agnostic approach, SpeeD consistently achieves 3-times acceleration across various diffusion architectures, datasets, and tasks. Notably, due to its simple design, our approach significantly reduces the cost of diffusion model training with minimal overhead. Our research enables more researchers to train diffusion models at a lower cost.
♻ ☆ Conditional Shift-Robust Conformal Prediction for Graph Neural Network
Graph Neural Networks (GNNs) have emerged as potent tools for predicting outcomes in graph-structured data. Despite their efficacy, a significant drawback of GNNs lies in their limited ability to provide robust uncertainty estimates, posing challenges to their reliability in contexts where errors carry significant consequences. Moreover, GNNs typically excel in in-distribution settings, assuming that training and test data follow identical distributions a condition often unmet in real world graph data scenarios. In this article, we leverage conformal prediction, a widely recognized statistical technique for quantifying uncertainty by transforming predictive model outputs into prediction sets, to address uncertainty quantification in GNN predictions amidst conditional shift\footnote{Representing the change in conditional probability distribution \(P(label|input)\) from source domain to target domain.} in graph-based semi-supervised learning (SSL). Additionally, we propose a novel loss function aimed at refining model predictions by minimizing conditional shift in latent stages. Termed Conditional Shift Robust (CondSR) conformal prediction for GNNs, our approach CondSR is model-agnostic and adaptable to various classification models. We validate the effectiveness of our method on standard graph benchmark datasets, integrating it with state-of-the-art GNNs in node classification tasks. Comprehensive evaluations demonstrate that our approach consistently achieves any predefined target marginal coverage, enhances the accuracy of state of the art GNN models by up to 12\% under conditional shift, and reduces the prediction set size by up to 48\%. The code implementation is publicly available for further exploration and experimentation.
comment: 15 pages, 3 figures, 4 tables
♻ ☆ Large Language Model for Patent Concept Generation
In traditional innovation practices, concept and IP generation are often iteratively integrated. Both processes demand an intricate understanding of advanced technical domain knowledge. Existing large language models (LLMs), while possessing massive pre-trained knowledge, often fall short in the innovative concept generation due to a lack of specialized knowledge necessary for the generation. To bridge this critical gap, we propose a novel knowledge finetuning (KFT) framework to endow LLM-based AI with the ability to autonomously mine, understand, and apply domain-specific knowledge and concepts for invention generation, i.e., concept and patent generation together. Our proposed PatentGPT integrates knowledge injection pre-training (KPT), domain-specific supervised finetuning (SFT), and reinforcement learning from human feedback (RLHF). Extensive evaluation shows that PatentGPT significantly outperforms the state-of-the-art models on patent-related benchmark tests. Our method not only provides new insights into data-driven innovation but also paves a new path to fine-tune LLMs for applications in the context of technology. We also discuss the managerial and policy implications of AI-generating inventions in the future.
comment: 33 pages, 8 figures
♻ ☆ CCUP: A Controllable Synthetic Data Generation Pipeline for Pretraining Cloth-Changing Person Re-Identification Models ICME 2025
Cloth-changing person re-identification (CC-ReID), also known as Long-Term Person Re-Identification (LT-ReID) is a critical and challenging research topic in computer vision that has recently garnered significant attention. However, due to the high cost of constructing CC-ReID data, the existing data-driven models are hard to train efficiently on limited data, causing overfitting issue. To address this challenge, we propose a low-cost and efficient pipeline for generating controllable and high-quality synthetic data simulating the surveillance of real scenarios specific to the CC-ReID task. Particularly, we construct a new self-annotated CC-ReID dataset named Cloth-Changing Unreal Person (CCUP), containing 6,000 IDs, 1,179,976 images, 100 cameras, and 26.5 outfits per individual. Based on this large-scale dataset, we introduce an effective and scalable pretrain-finetune framework for enhancing the generalization capabilities of the traditional CC-ReID models. The extensive experiments demonstrate that two typical models namely TransReID and FIRe^2, when integrated into our framework, outperform other state-of-the-art models after pretraining on CCUP and finetuning on the benchmarks such as PRCC, VC-Clothes and NKUP. The CCUP is available at: https://github.com/yjzhao1019/CCUP.
comment: Accepted by ICME 2025
♻ ☆ Extreme Precipitation Nowcasting using Multi-Task Latent Diffusion Models
Deep learning models have achieved remarkable progress in precipitation prediction. However, they still face significant challenges in accurately capturing spatial details of radar images, particularly in regions of high precipitation intensity. This limitation results in reduced spatial localization accuracy when predicting radar echo images across varying precipitation intensities. To address this challenge, we propose an innovative precipitation prediction approach termed the Multi-Task Latent Diffusion Model (MTLDM). The core idea of MTLDM lies in the recognition that precipitation radar images represent a combination of multiple components, each corresponding to different precipitation intensities. Thus, we adopt a divide-and-conquer strategy, decomposing radar images into several sub-images based on their precipitation intensities and individually modeling these components. During the prediction stage, MTLDM integrates these sub-image representations by utilizing a trained latent-space rainfall diffusion model, followed by decoding through a multi-task decoder to produce the final precipitation prediction. Experimental evaluations conducted on the MRMS dataset demonstrate that the proposed MTLDM method surpasses state-of-the-art techniques, achieving a Critical Success Index (CSI) improvement of 13-26%.
comment: 15 pages, 14figures
♻ ☆ VideoRefer Suite: Advancing Spatial-Temporal Object Understanding with Video LLM
Video Large Language Models (Video LLMs) have recently exhibited remarkable capabilities in general video understanding. However, they mainly focus on holistic comprehension and struggle with capturing fine-grained spatial and temporal details. Besides, the lack of high-quality object-level video instruction data and a comprehensive benchmark further hinders their advancements. To tackle these challenges, we introduce the VideoRefer Suite to empower Video LLM for finer-level spatial-temporal video understanding, i.e., enabling perception and reasoning on any objects throughout the video. Specially, we thoroughly develop VideoRefer Suite across three essential aspects: dataset, model, and benchmark. Firstly, we introduce a multi-agent data engine to meticulously curate a large-scale, high-quality object-level video instruction dataset, termed VideoRefer-700K. Next, we present the VideoRefer model, which equips a versatile spatial-temporal object encoder to capture precise regional and sequential representations. Finally, we meticulously create a VideoRefer-Bench to comprehensively assess the spatial-temporal understanding capability of a Video LLM, evaluating it across various aspects. Extensive experiments and analyses demonstrate that our VideoRefer model not only achieves promising performance on video referring benchmarks but also facilitates general video understanding capabilities.
comment: 17 pages, 14 figures, technical report
♻ ☆ GFlowVLM: Enhancing Multi-step Reasoning in Vision-Language Models with Generative Flow Networks
Vision-Language Models (VLMs) have recently shown promising advancements in sequential decision-making tasks through task-specific fine-tuning. However, common fine-tuning methods, such as Supervised Fine-Tuning (SFT) and Reinforcement Learning (RL) techniques like Proximal Policy Optimization (PPO), present notable limitations: SFT assumes Independent and Identically Distributed (IID) data, while PPO focuses on maximizing cumulative rewards. These limitations often restrict solution diversity and hinder generalization in multi-step reasoning tasks. To address these challenges, we introduce a novel framework, GFlowVLM, a framework that fine-tune VLMs using Generative Flow Networks (GFlowNets) to promote generation of diverse solutions for complex reasoning tasks. GFlowVLM models the environment as a non-Markovian decision process, allowing it to capture long-term dependencies essential for real-world applications. It takes observations and task descriptions as inputs to prompt chain-of-thought (CoT) reasoning which subsequently guides action selection. We use task based rewards to fine-tune VLM with GFlowNets. This approach enables VLMs to outperform prior fine-tuning methods, including SFT and RL. Empirical results demonstrate the effectiveness of GFlowVLM on complex tasks such as card games (NumberLine, BlackJack) and embodied planning tasks (ALFWorld), showing enhanced training efficiency, solution diversity, and stronger generalization capabilities across both in-distribution and out-of-distribution scenarios.
♻ ☆ Discovering Hidden Visual Concepts Beyond Linguistic Input in Infant Learning CVPR 2025
Infants develop complex visual understanding rapidly, even preceding the acquisition of linguistic skills. As computer vision seeks to replicate the human vision system, understanding infant visual development may offer valuable insights. In this paper, we present an interdisciplinary study exploring this question: can a computational model that imitates the infant learning process develop broader visual concepts that extend beyond the vocabulary it has heard, similar to how infants naturally learn? To investigate this, we analyze a recently published model in Science by Vong et al., which is trained on longitudinal, egocentric images of a single child paired with transcribed parental speech. We perform neuron labeling to identify visual concept neurons hidden in the model's internal representations. We then demonstrate that these neurons can recognize objects beyond the model's original vocabulary. Furthermore, we compare the differences in representation between infant models and those in modern computer vision models, such as CLIP and ImageNet pre-trained model. Ultimately, our work bridges cognitive science and computer vision by analyzing the internal representations of a computational model trained on an infant visual and linguistic inputs. Our code is available at https://github.com/Kexueyi/discover_infant_vis.
comment: Accepted at CVPR 2025
♻ ☆ BF-STVSR: B-Splines and Fourier-Best Friends for High Fidelity Spatial-Temporal Video Super-Resolution CVPR 2025
While prior methods in Continuous Spatial-Temporal Video Super-Resolution (C-STVSR) employ Implicit Neural Representation (INR) for continuous encoding, they often struggle to capture the complexity of video data, relying on simple coordinate concatenation and pre-trained optical flow networks for motion representation. Interestingly, we find that adding position encoding, contrary to common observations, does not improve--and even degrades--performance. This issue becomes particularly pronounced when combined with pre-trained optical flow networks, which can limit the model's flexibility. To address these issues, we propose BF-STVSR, a C-STVSR framework with two key modules tailored to better represent spatial and temporal characteristics of video: 1) B-spline Mapper for smooth temporal interpolation, and 2) Fourier Mapper for capturing dominant spatial frequencies. Our approach achieves state-of-the-art in various metrics, including PSNR and SSIM, showing enhanced spatial details and natural temporal consistency. Our code is available https://github.com/Eunjnnn/bfstvsr.
comment: CVPR 2025
♻ ☆ BioMamba: Leveraging Spectro-Temporal Embedding in Bidirectional Mamba for Enhanced Biosignal Classification
Biological signals, such as electroencephalograms (EEGs) and electrocardiograms (ECGs), play a pivotal role in numerous clinical practices, such as diagnosing brain and cardiac arrhythmic diseases. Existing methods for biosignal classification rely on Attention-based frameworks with dense Feed Forward layers, which lead to inefficient learning, high computational overhead, and suboptimal performance. In this work, we introduce BioMamba, a Spectro-Temporal Embedding strategy applied to the Bidirectional Mamba framework with Sparse Feed Forward layers to enable effective learning of biosignal sequences. By integrating these three key components, BioMamba effectively addresses the limitations of existing methods. Extensive experiments demonstrate that BioMamba significantly outperforms state-of-the-art methods with marked improvement in classification performance. The advantages of the proposed BioMamba include (1) Reliability: BioMamba consistently delivers robust results, confirmed across six evaluation metrics. (2) Efficiency: We assess both model and training efficiency, the BioMamba demonstrates computational effectiveness by reducing model size and resource consumption compared to existing approaches. (3) Generality: With the capacity to effectively classify a diverse set of tasks, BioMamba demonstrates adaptability and effectiveness across various domains and applications.
comment: Biological signals
♻ ☆ OmniDocBench: Benchmarking Diverse PDF Document Parsing with Comprehensive Annotations CVPR2025
Document content extraction is a critical task in computer vision, underpinning the data needs of large language models (LLMs) and retrieval-augmented generation (RAG) systems. Despite recent progress, current document parsing methods have not been fairly and comprehensively evaluated due to the narrow coverage of document types and the simplified, unrealistic evaluation procedures in existing benchmarks. To address these gaps, we introduce OmniDocBench, a novel benchmark featuring high-quality annotations across nine document sources, including academic papers, textbooks, and more challenging cases such as handwritten notes and densely typeset newspapers. OmniDocBench supports flexible, multi-level evaluations--ranging from an end-to-end assessment to the task-specific and attribute--based analysis using 19 layout categories and 15 attribute labels. We conduct a thorough evaluation of both pipeline-based methods and end-to-end vision-language models, revealing their strengths and weaknesses across different document types. OmniDocBench sets a new standard for the fair, diverse, and fine-grained evaluation in document parsing. Dataset and code are available at https://github.com/opendatalab/OmniDocBench.
comment: Accepted by CVPR2025
♻ ☆ MCRanker: Generating Diverse Criteria On-the-Fly to Improve Point-wise LLM Rankers
The most recent pointwise Large Language Model (LLM) rankers have achieved remarkable ranking results. However, these rankers are hindered by two major drawbacks: (1) they fail to follow a standardized comparison guidance during the ranking process, and (2) they struggle with comprehensive considerations when dealing with complicated passages. To address these shortcomings, we propose to build a ranker that generates ranking scores based on a set of criteria from various perspectives. These criteria are intended to direct each perspective in providing a distinct yet synergistic evaluation. Our research, which examines eight datasets from the BEIR benchmark demonstrates that incorporating this multi-perspective criteria ensemble approach markedly enhanced the performance of pointwise LLM rankers.
♻ ☆ XXLTraffic: Expanding and Extremely Long Traffic forecasting beyond test adaptation
Traffic forecasting is crucial for smart cities and intelligent transportation initiatives, where deep learning has made significant progress in modeling complex spatio-temporal patterns in recent years. However, current public datasets have limitations in reflecting the distribution shift nature of real-world scenarios, characterized by continuously evolving infrastructures, varying temporal distributions, and long temporal gaps due to sensor downtimes or changes in traffic patterns. These limitations inevitably restrict the practical applicability of existing traffic forecasting datasets. To bridge this gap, we present XXLTraffic, largest available public traffic dataset with the longest timespan collected from Los Angeles, USA, and New South Wales, Australia, curated to support research in extremely long forecasting beyond test adaptation. Our benchmark includes both typical time-series forecasting settings with hourly and daily aggregated data and novel configurations that introduce gaps and down-sample the training size to better simulate practical constraints. We anticipate the new XXLTraffic will provide a fresh perspective for the time-series and traffic forecasting communities. It would also offer a robust platform for developing and evaluating models designed to tackle the extremely long forecasting problems beyond test adaptation. Our dataset supplements existing spatio-temporal data resources and leads to new research directions in this domain.
♻ ☆ Polysemanticity and Capacity in Neural Networks
Individual neurons in neural networks often represent a mixture of unrelated features. This phenomenon, called polysemanticity, can make interpreting neural networks more difficult and so we aim to understand its causes. We propose doing so through the lens of feature \emph{capacity}, which is the fractional dimension each feature consumes in the embedding space. We show that in a toy model the optimal capacity allocation tends to monosemantically represent the most important features, polysemantically represent less important features (in proportion to their impact on the loss), and entirely ignore the least important features. Polysemanticity is more prevalent when the inputs have higher kurtosis or sparsity and more prevalent in some architectures than others. Given an optimal allocation of capacity, we go on to study the geometry of the embedding space. We find a block-semi-orthogonal structure, with differing block sizes in different models, highlighting the impact of model architecture on the interpretability of its neurons.
comment: 22 pages, 7 figures. Improved notation and corrected an error in the description of the most general efficient matrices
♻ ☆ Hardware-Friendly Static Quantization Method for Video Diffusion Transformers
Diffusion Transformers for video generation have gained significant research interest since the impressive performance of SORA. Efficient deployment of such generative-AI models on GPUs has been demonstrated with dynamic quantization. However, resource-constrained devices cannot support dynamic quantization, and need static quantization of the models for their efficient deployment on AI processors. In this paper, we propose a novel method for the post-training quantization of OpenSora\cite{opensora}, a Video Diffusion Transformer, without relying on dynamic quantization techniques. Our approach employs static quantization, achieving video quality comparable to FP16 and dynamically quantized ViDiT-Q methods, as measured by CLIP, and VQA metrics. In particular, we utilize per-step calibration data to adequately provide a post-training statically quantized model for each time step, incorporating channel-wise quantization for weights and tensor-wise quantization for activations. By further applying the smooth-quantization technique, we can obtain high-quality video outputs with the statically quantized models. Extensive experimental results demonstrate that static quantization can be a viable alternative to dynamic quantization for video diffusion transformers, offering a more efficient approach without sacrificing performance.
♻ ☆ Think Carefully and Check Again! Meta-Generation Unlocking LLMs for Low-Resource Cross-Lingual Summarization
Cross-lingual summarization (CLS) aims to generate a summary for the source text in a different target language. Currently, instruction-tuned large language models (LLMs) excel at various English tasks. However, unlike languages such as English, Chinese or Spanish, for those relatively low-resource languages with limited usage or data, recent studies have shown that LLMs' performance on CLS tasks remains unsatisfactory even with few-shot settings. This raises the question: Are LLMs capable of handling cross-lingual summarization tasks for low-resource languages? To resolve this question, we fully explore the potential of large language models on cross-lingual summarization task for low-resource languages through our four-step zero-shot method: Summarization, Improvement, Translation and Refinement (SITR) with correspondingly designed prompts. We test our proposed method with multiple LLMs on two well-known cross-lingual summarization datasets with various low-resource target languages. The results show that: i) GPT-3.5 and GPT-4 significantly and consistently outperform other baselines when using our zero-shot SITR methods. ii) By employing our proposed method, we unlock the potential of LLMs, enabling them to effectively handle cross-lingual summarization tasks for relatively low-resource languages.
♻ ☆ Long-range Meta-path Search on Large-scale Heterogeneous Graphs NeurIPS 2024
Utilizing long-range dependency, a concept extensively studied in homogeneous graphs, remains underexplored in heterogeneous graphs, especially on large ones, posing two significant challenges: Reducing computational costs while maximizing effective information utilization in the presence of heterogeneity, and overcoming the over-smoothing issue in graph neural networks. To address this gap, we investigate the importance of different meta-paths and introduce an automatic framework for utilizing long-range dependency on heterogeneous graphs, denoted as Long-range Meta-path Search through Progressive Sampling (LMSPS). Specifically, we develop a search space with all meta-paths related to the target node type. By employing a progressive sampling algorithm, LMSPS dynamically shrinks the search space with hop-independent time complexity. Through a sampling evaluation strategy, LMSPS conducts a specialized and effective meta-path selection, leading to retraining with only effective meta-paths, thus mitigating costs and over-smoothing. Extensive experiments across diverse heterogeneous datasets validate LMSPS's capability in discovering effective long-range meta-paths, surpassing state-of-the-art methods. Our code is available at https://github.com/JHL-HUST/LMSPS.
comment: Accepted by Advances in Neural Information Processing Systems (NeurIPS 2024)
♻ ☆ SyncDiff: Synchronized Motion Diffusion for Multi-Body Human-Object Interaction Synthesis
Synthesizing realistic human-object interaction motions is a critical problem in VR/AR and human animation. Unlike the commonly studied scenarios involving a single human or hand interacting with one object, we address a more generic multi-body setting with arbitrary numbers of humans, hands, and objects. This complexity introduces significant challenges in synchronizing motions due to the high correlations and mutual influences among bodies. To address these challenges, we introduce SyncDiff, a novel method for multi-body interaction synthesis using a synchronized motion diffusion strategy. SyncDiff employs a single diffusion model to capture the joint distribution of multi-body motions. To enhance motion fidelity, we propose a frequency-domain motion decomposition scheme. Additionally, we introduce a new set of alignment scores to emphasize the synchronization of different body motions. SyncDiff jointly optimizes both data sample likelihood and alignment likelihood through an explicit synchronization strategy. Extensive experiments across four datasets with various multi-body configurations demonstrate the superiority of SyncDiff over existing state-of-the-art motion synthesis methods.
♻ ☆ Lightweight Models for Emotional Analysis in Video
In this study, we present an approach for efficient spatiotemporal feature extraction using MobileNetV4 and a multi-scale 3D MLP-Mixer-based temporal aggregation module. MobileNetV4, with its Universal Inverted Bottleneck (UIB) blocks, serves as the backbone for extracting hierarchical feature representations from input image sequences, ensuring both computational efficiency and rich semantic encoding. To capture temporal dependencies, we introduce a three-level MLP-Mixer module, which processes spatial features at multiple resolutions while maintaining structural integrity. Experimental results on the ABAW 8th competition demonstrate the effectiveness of our approach, showing promising performance in affective behavior analysis. By integrating an efficient vision backbone with a structured temporal modeling mechanism, the proposed framework achieves a balance between computational efficiency and predictive accuracy, making it well-suited for real-time applications in mobile and embedded computing environments.
comment: https://github.com/PRVSL/abaw-8th
♻ ☆ Commonsense Video Question Answering through Video-Grounded Entailment Tree Reasoning CVPR 2025
This paper proposes the first video-grounded entailment tree reasoning method for commonsense video question answering (VQA). Despite the remarkable progress of large visual-language models (VLMs), there are growing concerns that they learn spurious correlations between videos and likely answers, reinforced by their black-box nature and remaining benchmarking biases. Our method explicitly grounds VQA tasks to video fragments in four steps: entailment tree construction, video-language entailment verification, tree reasoning, and dynamic tree expansion. A vital benefit of the method is its generalizability to current video and image-based VLMs across reasoning types. To support fair evaluation, we devise a de-biasing procedure based on large-language models that rewrites VQA benchmark answer sets to enforce model reasoning. Systematic experiments on existing and de-biased benchmarks highlight the impact of our method components across benchmarks, VLMs, and reasoning types.
comment: Accepted by CVPR 2025
♻ ☆ Natural Language Processing for the Legal Domain: A Survey of Tasks, Datasets, Models, and Challenges
Natural Language Processing (NLP) is revolutionising the way legal professionals and laypersons operate in the legal field. The considerable potential for NLP in the legal sector, especially in developing computational tools for various legal processes, has captured the interest of researchers for years. This survey follows the Preferred Reporting Items for Systematic Reviews and Meta-Analyses framework, reviewing 154 studies, with a final selection of 133 after manual filtering. It explores foundational concepts related to NLP in the legal domain, illustrating the unique aspects and challenges of processing legal texts, such as extensive document length, complex language, and limited open legal datasets. We provide an overview of NLP tasks specific to legal text, such as Legal Document Summarisation, legal Named Entity Recognition, Legal Question Answering, Legal Argument Mining, Legal Text Classification, and Legal Judgement Prediction. In the section on legal Language Models (LMs), we analyse both developed LMs and approaches for adapting general LMs to the legal domain. Additionally, we identify 16 Open Research Challenges, including bias in Artificial Intelligence applications, the need for more robust and interpretable models, and improving explainability to handle the complexities of legal language and reasoning.
comment: 35 pages
♻ ☆ The Surprising Effectiveness of Test-Time Training for Few-Shot Learning
Language models (LMs) have shown impressive performance on tasks within their training distribution, but often struggle with structurally novel tasks even when given a small number of in-context task examples. We investigate the effectiveness of test-time training (TTT) -- temporarily updating model parameters during inference using a loss derived from input data -- as a mechanism for improving LMs' reasoning and few-shot learning capabilities. On the Abstraction and Reasoning Corpus (ARC), performing TTT with in-context examples yields up to $6\times$ higher accuracy compared to fine-tuned baselines -- reaching $53.0\%$ on the public validation set with an 8B-parameter LM and $61.9\%$ when ensembled with program-synthesis methods, matching average human performance. On BIG-Bench Hard (BBH), TTT on in-context examples surpasses standard few-shot prompting in the $10$-shot setting by $7.3$ percentage points ($50.5\%$ to $57.8\%$). Our findings highlight the limitations of in-context learning for novel tasks and demonstrate the potential of test-time training to enhance language model adaptability.
comment: Preprint
♻ ☆ Coverage-based Fairness in Multi-document Summarization NAACL 2025
Fairness in multi-document summarization (MDS) measures whether a system can generate a summary fairly representing information from documents with different social attribute values. Fairness in MDS is crucial since a fair summary can offer readers a comprehensive view. Previous works focus on quantifying summary-level fairness using Proportional Representation, a fairness measure based on Statistical Parity. However, Proportional Representation does not consider redundancy in input documents and overlooks corpus-level unfairness. In this work, we propose a new summary-level fairness measure, Equal Coverage, which is based on coverage of documents with different social attribute values and considers the redundancy within documents. To detect the corpus-level unfairness, we propose a new corpus-level measure, Coverage Parity. Our human evaluations show that our measures align more with our definition of fairness. Using our measures, we evaluate the fairness of thirteen different LLMs. We find that Claude3-sonnet is the fairest among all evaluated LLMs. We also find that almost all LLMs overrepresent different social attribute values. The code is available at https://github.com/leehaoyuan/coverage_fairness.
comment: accepted to NAACL 2025
♻ ☆ On Diffusion Modeling for Anomaly Detection
Known for their impressive performance in generative modeling, diffusion models are attractive candidates for density-based anomaly detection. This paper investigates different variations of diffusion modeling for unsupervised and semi-supervised anomaly detection. In particular, we find that Denoising Diffusion Probability Models (DDPM) are performant on anomaly detection benchmarks yet computationally expensive. By simplifying DDPM in application to anomaly detection, we are naturally led to an alternative approach called Diffusion Time Estimation (DTE). DTE estimates the distribution over diffusion time for a given input and uses the mode or mean of this distribution as the anomaly score. We derive an analytical form for this density and leverage a deep neural network to improve inference efficiency. Through empirical evaluations on the ADBench benchmark, we demonstrate that all diffusion-based anomaly detection methods perform competitively for both semi-supervised and unsupervised settings. Notably, DTE achieves orders of magnitude faster inference time than DDPM, while outperforming it on this benchmark. These results establish diffusion-based anomaly detection as a scalable alternative to traditional methods and recent deep-learning techniques for standard unsupervised and semi-supervised anomaly detection settings.
♻ ☆ Non-autoregressive Generative Models for Reranking Recommendation KDD 2024
Contemporary recommendation systems are designed to meet users' needs by delivering tailored lists of items that align with their specific demands or interests. In a multi-stage recommendation system, reranking plays a crucial role by modeling the intra-list correlations among items. The key challenge of reranking lies in the exploration of optimal sequences within the combinatorial space of permutations. Recent research proposes a generator-evaluator learning paradigm, where the generator generates multiple feasible sequences and the evaluator picks out the best sequence based on the estimated listwise score. The generator is of vital importance, and generative models are well-suited for the generator function. Current generative models employ an autoregressive strategy for sequence generation. However, deploying autoregressive models in real-time industrial systems is challenging. To address these issues, we propose a Non-AutoRegressive generative model for reranking Recommendation (NAR4Rec) designed to enhance efficiency and effectiveness. To tackle challenges such as sparse training samples and dynamic candidates, we introduce a matching model. Considering the diverse nature of user feedback, we employ a sequence-level unlikelihood training objective to differentiate feasible sequences from unfeasible ones. Additionally, to overcome the lack of dependency modeling in non-autoregressive models regarding target items, we introduce contrastive decoding to capture correlations among these items. Extensive offline experiments validate the superior performance of NAR4Rec over state-of-the-art reranking methods. Online A/B tests reveal that NAR4Rec significantly enhances the user experience. Furthermore, NAR4Rec has been fully deployed in a popular video app Kuaishou with over 300 million daily active users.
comment: Accepted by KDD 2024
♻ ☆ Swift Hydra: Self-Reinforcing Generative Framework for Anomaly Detection with Multiple Mamba Models
Despite a plethora of anomaly detection models developed over the years, their ability to generalize to unseen anomalies remains an issue, particularly in critical systems. This paper aims to address this challenge by introducing Swift Hydra, a new framework for training an anomaly detection method based on generative AI and reinforcement learning (RL). Through featuring an RL policy that operates on the latent variables of a generative model, the framework synthesizes novel and diverse anomaly samples that are capable of bypassing a detection model. These generated synthetic samples are, in turn, used to augment the detection model, further improving its ability to handle challenging anomalies. Swift Hydra also incorporates Mamba models structured as a Mixture of Experts (MoE) to enable scalable adaptation of the number of Mamba experts based on data complexity, effectively capturing diverse feature distributions without increasing the model's inference time. Empirical evaluations on ADBench benchmark demonstrate that Swift Hydra outperforms other state-of-the-art anomaly detection models while maintaining a relatively short inference time. From these results, our research highlights a new and auspicious paradigm of integrating RL and generative AI for advancing anomaly detection.
♻ ☆ Zeroth-order Informed Fine-Tuning for Diffusion Model: A Recursive Likelihood Ratio Optimizer
The probabilistic diffusion model (DM), generating content by inferencing through a recursive chain structure, has emerged as a powerful framework for visual generation. After pre-training on enormous unlabeled data, the model needs to be properly aligned to meet requirements for downstream applications. How to efficiently align the foundation DM is a crucial task. Contemporary methods are either based on Reinforcement Learning (RL) or truncated Backpropagation (BP). However, RL and truncated BP suffer from low sample efficiency and biased gradient estimation respectively, resulting in limited improvement or, even worse, complete training failure. To overcome the challenges, we propose the Recursive Likelihood Ratio (RLR) optimizer, a zeroth-order informed fine-tuning paradigm for DM. The zeroth-order gradient estimator enables the computation graph rearrangement within the recursive diffusive chain, making the RLR's gradient estimator an unbiased one with the lower variance than other methods. We provide theoretical guarantees for the performance of the RLR. Extensive experiments are conducted on image and video generation tasks to validate the superiority of the RLR. Furthermore, we propose a novel prompt technique that is natural for the RLR to achieve a synergistic effect.
♻ ☆ Mathematics and Machine Creativity: A Survey on Bridging Mathematics with AI
This paper presents a comprehensive overview on the applications of artificial intelligence (AI) in mathematical research, highlighting the transformative role AI has begun to play in this domain. Traditionally, AI advancements have heavily relied on theoretical foundations provided by mathematics and statistics. However, recent developments in AI, particularly in reinforcement learning (RL) and large language models (LLMs), have demonstrated the potential for AI to contribute back to mathematics by offering flexible algorithmic frameworks and powerful inductive reasoning capabilities that support various aspects of mathematical research. This survey aims to establish a bridge between AI and mathematics, providing insights into the mutual benefits and fostering deeper interdisciplinary understanding. In particular, we argue that while current AI and LLMs may struggle with complex deductive reasoning, their "inherent creativity", the ability to generate outputs at high throughput based on recognition of shallow patterns, holds significant potential to support and inspire mathematical research. This creative capability, often overlooked, could be the key to unlocking new perspectives and methodologies in mathematics. Furthermore, we address the lack of cross-disciplinary communication: mathematicians may not fully comprehend the latest advances in AI, while AI researchers frequently prioritize benchmark performance over real-world applications in frontier mathematical research. This paper seeks to close that gap, offering a detailed exploration of AI fundamentals, its strengths, and its emerging applications in the mathematical sciences.
comment: 26 pages, 3 figures
♻ ☆ h4rm3l: A language for Composable Jailbreak Attack Synthesis ICLR 2025
Despite their demonstrated valuable capabilities, state-of-the-art (SOTA) widely deployed large language models (LLMs) still have the potential to cause harm to society due to the ineffectiveness of their safety filters, which can be bypassed by prompt transformations called jailbreak attacks. Current approaches to LLM safety assessment, which employ datasets of templated prompts and benchmarking pipelines, fail to cover sufficiently large and diverse sets of jailbreak attacks, leading to the widespread deployment of unsafe LLMs. Recent research showed that novel jailbreak attacks could be derived by composition; however, a formal composable representation for jailbreak attacks, which, among other benefits, could enable the exploration of a large compositional space of jailbreak attacks through program synthesis methods, has not been previously proposed. We introduce h4rm3l, a novel approach that addresses this gap with a human-readable domain-specific language (DSL). Our framework comprises: (1) The h4rm3l DSL, which formally expresses jailbreak attacks as compositions of parameterized string transformation primitives. (2) A synthesizer with bandit algorithms that efficiently generates jailbreak attacks optimized for a target black box LLM. (3) The h4rm3l red-teaming software toolkit that employs the previous two components and an automated harmful LLM behavior classifier that is strongly aligned with human judgment. We demonstrate h4rm3l's efficacy by synthesizing a dataset of 2656 successful novel jailbreak attacks targeting 6 SOTA open-source and proprietary LLMs, and by benchmarking those models against a subset of these synthesized attacks. Our results show that h4rm3l's synthesized attacks are diverse and more successful than existing jailbreak attacks in literature, with success rates exceeding 90% on SOTA LLMs.
comment: Accepted to the Thirteenth International Conference on Learning Representations (ICLR 2025)
♻ ☆ IPCGRL: Language-Instructed Reinforcement Learning for Procedural Level Generation
Recent research has highlighted the significance of natural language in enhancing the controllability of generative models. While various efforts have been made to leverage natural language for content generation, research on deep reinforcement learning (DRL) agents utilizing text-based instructions for procedural content generation remains limited. In this paper, we propose IPCGRL, an instruction-based procedural content generation method via reinforcement learning, which incorporates a sentence embedding model. IPCGRL fine-tunes task-specific embedding representations to effectively compress game-level conditions. We evaluate IPCGRL in a two-dimensional level generation task and compare its performance with a general-purpose embedding method. The results indicate that IPCGRL achieves up to a 21.4% improvement in controllability and a 17.2% improvement in generalizability for unseen instructions. Furthermore, the proposed method extends the modality of conditional input, enabling a more flexible and expressive interaction framework for procedural content generation.
comment: 9 pages, 9 figures, 3 tables
♻ ☆ Generative Prompt Internalization NAACL 2025
Prompts used in recent large language model based applications are often fixed and lengthy, leading to significant computational overhead. To address this challenge, we propose Generative Prompt Internalization (GenPI), a lightweight method that employs a joint training approach. GenPI not only replicates the behavior of models with prompt inputs but also generates the content of the prompt along with reasons for why the model's behavior should change accordingly. We demonstrate that our approach effectively internalizes complex prompts across various agent-based application scenarios. For effective training without interactions with the dedicated environments, we introduce a data synthesis technique that autonomously collects conversational datasets by swapping the roles of the agent and environment. This method is especially useful in scenarios where only a predefined prompt is available without a corresponding training dataset. By internalizing complex prompts, Generative Prompt Internalization enables high performance and efficient inference without the need for explicit prompts.
comment: NAACL 2025 (Main Conference)
♻ ☆ Cross-modal Information Flow in Multimodal Large Language Models
The recent advancements in auto-regressive multimodal large language models (MLLMs) have demonstrated promising progress for vision-language tasks. While there exists a variety of studies investigating the processing of linguistic information within large language models, little is currently known about the inner working mechanism of MLLMs and how linguistic and visual information interact within these models. In this study, we aim to fill this gap by examining the information flow between different modalities -- language and vision -- in MLLMs, focusing on visual question answering. Specifically, given an image-question pair as input, we investigate where in the model and how the visual and linguistic information are combined to generate the final prediction. Conducting experiments with a series of models from the LLaVA series, we find that there are two distinct stages in the process of integration of the two modalities. In the lower layers, the model first transfers the more general visual features of the whole image into the representations of (linguistic) question tokens. In the middle layers, it once again transfers visual information about specific objects relevant to the question to the respective token positions of the question. Finally, in the higher layers, the resulting multimodal representation is propagated to the last position of the input sequence for the final prediction. Overall, our findings provide a new and comprehensive perspective on the spatial and functional aspects of image and language processing in the MLLMs, thereby facilitating future research into multimodal information localization and editing. Our code and collected dataset are released here: https://github.com/FightingFighting/cross-modal-information-flow-in-MLLM.git.
Computer Vision and Pattern Recognition 128
☆ Zero-Shot Human-Object Interaction Synthesis with Multimodal Priors
Human-object interaction (HOI) synthesis is important for various applications, ranging from virtual reality to robotics. However, acquiring 3D HOI data is challenging due to its complexity and high cost, limiting existing methods to the narrow diversity of object types and interaction patterns in training datasets. This paper proposes a novel zero-shot HOI synthesis framework without relying on end-to-end training on currently limited 3D HOI datasets. The core idea of our method lies in leveraging extensive HOI knowledge from pre-trained Multimodal Models. Given a text description, our system first obtains temporally consistent 2D HOI image sequences using image or video generation models, which are then uplifted to 3D HOI milestones of human and object poses. We employ pre-trained human pose estimation models to extract human poses and introduce a generalizable category-level 6-DoF estimation method to obtain the object poses from 2D HOI images. Our estimation method is adaptive to various object templates obtained from text-to-3D models or online retrieval. A physics-based tracking of the 3D HOI kinematic milestone is further applied to refine both body motions and object poses, yielding more physically plausible HOI generation results. The experimental results demonstrate that our method is capable of generating open-vocabulary HOIs with physical realism and semantic diversity.
☆ Peepers & Pixels: Human Recognition Accuracy on Low Resolution Faces
Automated one-to-many (1:N) face recognition is a powerful investigative tool commonly used by law enforcement agencies. In this context, potential matches resulting from automated 1:N recognition are reviewed by human examiners prior to possible use as investigative leads. While automated 1:N recognition can achieve near-perfect accuracy under ideal imaging conditions, operational scenarios may necessitate the use of surveillance imagery, which is often degraded in various quality dimensions. One important quality dimension is image resolution, typically quantified by the number of pixels on the face. The common metric for this is inter-pupillary distance (IPD), which measures the number of pixels between the pupils. Low IPD is known to degrade the accuracy of automated face recognition. However, the threshold IPD for reliability in human face recognition remains undefined. This study aims to explore the boundaries of human recognition accuracy by systematically testing accuracy across a range of IPD values. We find that at low IPDs (10px, 5px), human accuracy is at or below chance levels (50.7%, 35.9%), even as confidence in decision-making remains relatively high (77%, 70.7%). Our findings indicate that, for low IPD images, human recognition ability could be a limiting factor to overall system accuracy.
comment: 10 pages, 3 figures
☆ EBS-EKF: Accurate and High Frequency Event-based Star Tracking CVPR
Event-based sensors (EBS) are a promising new technology for star tracking due to their low latency and power efficiency, but prior work has thus far been evaluated exclusively in simulation with simplified signal models. We propose a novel algorithm for event-based star tracking, grounded in an analysis of the EBS circuit and an extended Kalman filter (EKF). We quantitatively evaluate our method using real night sky data, comparing its results with those from a space-ready active-pixel sensor (APS) star tracker. We demonstrate that our method is an order-of-magnitude more accurate than existing methods due to improved signal modeling and state estimation, while providing more frequent updates and greater motion tolerance than conventional APS trackers. We provide all code and the first dataset of events synchronized with APS solutions.
comment: Accepted into the proceedings of the Conference on Computer Vision and Pattern Recognition (CVPR) for 2025. Link to code and dataset is https://gitlab.kitware.com/nest-public/kw_ebs_star_tracking#
☆ Can Multi-modal (reasoning) LLMs work as deepfake detectors?
Deepfake detection remains a critical challenge in the era of advanced generative models, particularly as synthetic media becomes more sophisticated. In this study, we explore the potential of state of the art multi-modal (reasoning) large language models (LLMs) for deepfake image detection such as (OpenAI O1/4o, Gemini thinking Flash 2, Deepseek Janus, Grok 3, llama 3.2, Qwen 2/2.5 VL, Mistral Pixtral, Claude 3.5/3.7 sonnet) . We benchmark 12 latest multi-modal LLMs against traditional deepfake detection methods across multiple datasets, including recently published real-world deepfake imagery. To enhance performance, we employ prompt tuning and conduct an in-depth analysis of the models' reasoning pathways to identify key contributing factors in their decision-making process. Our findings indicate that best multi-modal LLMs achieve competitive performance with promising generalization ability with zero shot, even surpass traditional deepfake detection pipelines in out-of-distribution datasets while the rest of the LLM families performs extremely disappointing with some worse than random guess. Furthermore, we found newer model version and reasoning capabilities does not contribute to performance in such niche tasks of deepfake detection while model size do help in some cases. This study highlights the potential of integrating multi-modal reasoning in future deepfake detection frameworks and provides insights into model interpretability for robustness in real-world scenarios.
☆ iNatAg: Multi-Class Classification Models Enabled by a Large-Scale Benchmark Dataset with 4.7M Images of 2,959 Crop and Weed Species
Accurate identification of crop and weed species is critical for precision agriculture and sustainable farming. However, it remains a challenging task due to a variety of factors -- a high degree of visual similarity among species, environmental variability, and a continued lack of large, agriculture-specific image data. We introduce iNatAg, a large-scale image dataset which contains over 4.7 million images of 2,959 distinct crop and weed species, with precise annotations along the taxonomic hierarchy from binary crop/weed labels to specific species labels. Curated from the broader iNaturalist database, iNatAg contains data from every continent and accurately reflects the variability of natural image captures and environments. Enabled by this data, we train benchmark models built upon the Swin Transformer architecture and evaluate the impact of various modifications such as the incorporation of geospatial data and LoRA finetuning. Our best models achieve state-of-the-art performance across all taxonomic classification tasks, achieving 92.38\% on crop and weed classification. Furthermore, the scale of our dataset enables us to explore incorrect misclassifications and unlock new analytic possiblities for plant species. By combining large-scale species coverage, multi-task labels, and geographic diversity, iNatAg provides a new foundation for building robust, geolocation-aware agricultural classification systems. We release the iNatAg dataset publicly through AgML (https://github.com/Project-AgML/AgML), enabling direct access and integration into agricultural machine learning workflows.
☆ Learning Scene-Level Signed Directional Distance Function with Ellipsoidal Priors and Neural Residuals
Dense geometric environment representations are critical for autonomous mobile robot navigation and exploration. Recent work shows that implicit continuous representations of occupancy, signed distance, or radiance learned using neural networks offer advantages in reconstruction fidelity, efficiency, and differentiability over explicit discrete representations based on meshes, point clouds, and voxels. In this work, we explore a directional formulation of signed distance, called signed directional distance function (SDDF). Unlike signed distance function (SDF) and similar to neural radiance fields (NeRF), SDDF has a position and viewing direction as input. Like SDF and unlike NeRF, SDDF directly provides distance to the observed surface along the direction, rather than integrating along the view ray, allowing efficient view synthesis. To learn and predict scene-level SDDF efficiently, we develop a differentiable hybrid representation that combines explicit ellipsoid priors and implicit neural residuals. This approach allows the model to effectively handle large distance discontinuities around obstacle boundaries while preserving the ability for dense high-fidelity prediction. We show that SDDF is competitive with the state-of-the-art neural implicit scene models in terms of reconstruction accuracy and rendering efficiency, while allowing differentiable view prediction for robot trajectory optimization.
☆ Med3DVLM: An Efficient Vision-Language Model for 3D Medical Image Analysis
Vision-language models (VLMs) have shown promise in 2D medical image analysis, but extending them to 3D remains challenging due to the high computational demands of volumetric data and the difficulty of aligning 3D spatial features with clinical text. We present Med3DVLM, a 3D VLM designed to address these challenges through three key innovations: (1) DCFormer, an efficient encoder that uses decomposed 3D convolutions to capture fine-grained spatial features at scale; (2) SigLIP, a contrastive learning strategy with pairwise sigmoid loss that improves image-text alignment without relying on large negative batches; and (3) a dual-stream MLP-Mixer projector that fuses low- and high-level image features with text embeddings for richer multi-modal representations. We evaluate our model on the M3D dataset, which includes radiology reports and VQA data for 120,084 3D medical images. Results show that Med3DVLM achieves superior performance across multiple benchmarks. For image-text retrieval, it reaches 61.00% R@1 on 2,000 samples, significantly outperforming the current state-of-the-art M3D model (19.10%). For report generation, it achieves a METEOR score of 36.42% (vs. 14.38%). In open-ended visual question answering (VQA), it scores 36.76% METEOR (vs. 33.58%), and in closed-ended VQA, it achieves 79.95% accuracy (vs. 75.78%). These results highlight Med3DVLM's ability to bridge the gap between 3D imaging and language, enabling scalable, multi-task reasoning across clinical applications. Our code is publicly available at https://github.com/mirthAI/Med3DVLM.
☆ Hyperdimensional Uncertainty Quantification for Multimodal Uncertainty Fusion in Autonomous Vehicles Perception CVPR 2025
Uncertainty Quantification (UQ) is crucial for ensuring the reliability of machine learning models deployed in real-world autonomous systems. However, existing approaches typically quantify task-level output prediction uncertainty without considering epistemic uncertainty at the multimodal feature fusion level, leading to sub-optimal outcomes. Additionally, popular uncertainty quantification methods, e.g., Bayesian approximations, remain challenging to deploy in practice due to high computational costs in training and inference. In this paper, we propose HyperDUM, a novel deterministic uncertainty method (DUM) that efficiently quantifies feature-level epistemic uncertainty by leveraging hyperdimensional computing. Our method captures the channel and spatial uncertainties through channel and patch -wise projection and bundling techniques respectively. Multimodal sensor features are then adaptively weighted to mitigate uncertainty propagation and improve feature fusion. Our evaluations show that HyperDUM on average outperforms the state-of-the-art (SOTA) algorithms by up to 2.01%/1.27% in 3D Object Detection and up to 1.29% improvement over baselines in semantic segmentation tasks under various types of uncertainties. Notably, HyperDUM requires 2.36x less Floating Point Operations and up to 38.30x less parameters than SOTA methods, providing an efficient solution for real-world autonomous systems.
comment: Accepted at CVPR 2025
☆ The Coralscapes Dataset: Semantic Scene Understanding in Coral Reefs
Coral reefs are declining worldwide due to climate change and local stressors. To inform effective conservation or restoration, monitoring at the highest possible spatial and temporal resolution is necessary. Conventional coral reef surveying methods are limited in scalability due to their reliance on expert labor time, motivating the use of computer vision tools to automate the identification and abundance estimation of live corals from images. However, the design and evaluation of such tools has been impeded by the lack of large high quality datasets. We release the Coralscapes dataset, the first general-purpose dense semantic segmentation dataset for coral reefs, covering 2075 images, 39 benthic classes, and 174k segmentation masks annotated by experts. Coralscapes has a similar scope and the same structure as the widely used Cityscapes dataset for urban scene segmentation, allowing benchmarking of semantic segmentation models in a new challenging domain which requires expert knowledge to annotate. We benchmark a wide range of semantic segmentation models, and find that transfer learning from Coralscapes to existing smaller datasets consistently leads to state-of-the-art performance. Coralscapes will catalyze research on efficient, scalable, and standardized coral reef surveying methods based on computer vision, and holds the potential to streamline the development of underwater ecological robotics.
☆ SLIP: Spoof-Aware One-Class Face Anti-Spoofing with Language Image Pretraining AAAI 2025
Face anti-spoofing (FAS) plays a pivotal role in ensuring the security and reliability of face recognition systems. With advancements in vision-language pretrained (VLP) models, recent two-class FAS techniques have leveraged the advantages of using VLP guidance, while this potential remains unexplored in one-class FAS methods. The one-class FAS focuses on learning intrinsic liveness features solely from live training images to differentiate between live and spoof faces. However, the lack of spoof training data can lead one-class FAS models to inadvertently incorporate domain information irrelevant to the live/spoof distinction (e.g., facial content), causing performance degradation when tested with a new application domain. To address this issue, we propose a novel framework called Spoof-aware one-class face anti-spoofing with Language Image Pretraining (SLIP). Given that live faces should ideally not be obscured by any spoof-attack-related objects (e.g., paper, or masks) and are assumed to yield zero spoof cue maps, we first propose an effective language-guided spoof cue map estimation to enhance one-class FAS models by simulating whether the underlying faces are covered by attack-related objects and generating corresponding nonzero spoof cue maps. Next, we introduce a novel prompt-driven liveness feature disentanglement to alleviate live/spoof-irrelative domain variations by disentangling live/spoof-relevant and domain-dependent information. Finally, we design an effective augmentation strategy by fusing latent features from live images and spoof prompts to generate spoof-like image features and thus diversify latent spoof features to facilitate the learning of one-class FAS. Our extensive experiments and ablation studies support that SLIP consistently outperforms previous one-class FAS methods.
comment: Accepted by AAAI 2025
☆ Thin-Shell-SfT: Fine-Grained Monocular Non-rigid 3D Surface Tracking with Neural Deformation Fields CVPR 2025
3D reconstruction of highly deformable surfaces (e.g. cloths) from monocular RGB videos is a challenging problem, and no solution provides a consistent and accurate recovery of fine-grained surface details. To account for the ill-posed nature of the setting, existing methods use deformation models with statistical, neural, or physical priors. They also predominantly rely on nonadaptive discrete surface representations (e.g. polygonal meshes), perform frame-by-frame optimisation leading to error propagation, and suffer from poor gradients of the mesh-based differentiable renderers. Consequently, fine surface details such as cloth wrinkles are often not recovered with the desired accuracy. In response to these limitations, we propose ThinShell-SfT, a new method for non-rigid 3D tracking that represents a surface as an implicit and continuous spatiotemporal neural field. We incorporate continuous thin shell physics prior based on the Kirchhoff-Love model for spatial regularisation, which starkly contrasts the discretised alternatives of earlier works. Lastly, we leverage 3D Gaussian splatting to differentiably render the surface into image space and optimise the deformations based on analysis-bysynthesis principles. Our Thin-Shell-SfT outperforms prior works qualitatively and quantitatively thanks to our continuous surface formulation in conjunction with a specially tailored simulation prior and surface-induced 3D Gaussians. See our project page at https://4dqv.mpiinf.mpg.de/ThinShellSfT.
comment: 15 pages, 12 figures and 3 tables; project page: https://4dqv.mpiinf.mpg.de/ThinShellSfT; CVPR 2025
☆ EventFly: Event Camera Perception from Ground to the Sky CVPR 2025
Cross-platform adaptation in event-based dense perception is crucial for deploying event cameras across diverse settings, such as vehicles, drones, and quadrupeds, each with unique motion dynamics, viewpoints, and class distributions. In this work, we introduce EventFly, a framework for robust cross-platform adaptation in event camera perception. Our approach comprises three key components: i) Event Activation Prior (EAP), which identifies high-activation regions in the target domain to minimize prediction entropy, fostering confident, domain-adaptive predictions; ii) EventBlend, a data-mixing strategy that integrates source and target event voxel grids based on EAP-driven similarity and density maps, enhancing feature alignment; and iii) EventMatch, a dual-discriminator technique that aligns features from source, target, and blended domains for better domain-invariant learning. To holistically assess cross-platform adaptation abilities, we introduce EXPo, a large-scale benchmark with diverse samples across vehicle, drone, and quadruped platforms. Extensive experiments validate our effectiveness, demonstrating substantial gains over popular adaptation methods. We hope this work can pave the way for more adaptive, high-performing event perception across diverse and complex environments.
comment: CVPR 2025; 30 pages, 8 figures, 16 tables; Project Page at https://event-fly.github.io/
☆ PartRM: Modeling Part-Level Dynamics with Large Cross-State Reconstruction Model CVPR 2025
As interest grows in world models that predict future states from current observations and actions, accurately modeling part-level dynamics has become increasingly relevant for various applications. Existing approaches, such as Puppet-Master, rely on fine-tuning large-scale pre-trained video diffusion models, which are impractical for real-world use due to the limitations of 2D video representation and slow processing times. To overcome these challenges, we present PartRM, a novel 4D reconstruction framework that simultaneously models appearance, geometry, and part-level motion from multi-view images of a static object. PartRM builds upon large 3D Gaussian reconstruction models, leveraging their extensive knowledge of appearance and geometry in static objects. To address data scarcity in 4D, we introduce the PartDrag-4D dataset, providing multi-view observations of part-level dynamics across over 20,000 states. We enhance the model's understanding of interaction conditions with a multi-scale drag embedding module that captures dynamics at varying granularities. To prevent catastrophic forgetting during fine-tuning, we implement a two-stage training process that focuses sequentially on motion and appearance learning. Experimental results show that PartRM establishes a new state-of-the-art in part-level motion learning and can be applied in manipulation tasks in robotics. Our code, data, and models are publicly available to facilitate future research.
comment: Accepted to CVPR 2025. Project Page: https://partrm.c7w.tech/
☆ Learning 3D Object Spatial Relationships from Pre-trained 2D Diffusion Models
We present a method for learning 3D spatial relationships between object pairs, referred to as object-object spatial relationships (OOR), by leveraging synthetically generated 3D samples from pre-trained 2D diffusion models. We hypothesize that images synthesized by 2D diffusion models inherently capture plausible and realistic OOR cues, enabling efficient ways to collect a 3D dataset to learn OOR for various unbounded object categories. Our approach begins by synthesizing diverse images that capture plausible OOR cues, which we then uplift into 3D samples. Leveraging our diverse collection of plausible 3D samples for the object pairs, we train a score-based OOR diffusion model to learn the distribution of their relative spatial relationships. Additionally, we extend our pairwise OOR to multi-object OOR by enforcing consistency across pairwise relations and preventing object collisions. Extensive experiments demonstrate the robustness of our method across various object-object spatial relationships, along with its applicability to real-world 3D scene arrangement tasks using the OOR diffusion model.
comment: Project Page: https://tlb-miss.github.io/oor/
☆ SuperFlow++: Enhanced Spatiotemporal Consistency for Cross-Modal Data Pretraining
LiDAR representation learning has emerged as a promising approach to reducing reliance on costly and labor-intensive human annotations. While existing methods primarily focus on spatial alignment between LiDAR and camera sensors, they often overlook the temporal dynamics critical for capturing motion and scene continuity in driving scenarios. To address this limitation, we propose SuperFlow++, a novel framework that integrates spatiotemporal cues in both pretraining and downstream tasks using consecutive LiDAR-camera pairs. SuperFlow++ introduces four key components: (1) a view consistency alignment module to unify semantic information across camera views, (2) a dense-to-sparse consistency regularization mechanism to enhance feature robustness across varying point cloud densities, (3) a flow-based contrastive learning approach that models temporal relationships for improved scene understanding, and (4) a temporal voting strategy that propagates semantic information across LiDAR scans to improve prediction consistency. Extensive evaluations on 11 heterogeneous LiDAR datasets demonstrate that SuperFlow++ outperforms state-of-the-art methods across diverse tasks and driving conditions. Furthermore, by scaling both 2D and 3D backbones during pretraining, we uncover emergent properties that provide deeper insights into developing scalable 3D foundation models. With strong generalizability and computational efficiency, SuperFlow++ establishes a new benchmark for data-efficient LiDAR-based perception in autonomous driving. The code is publicly available at https://github.com/Xiangxu-0103/SuperFlow
comment: Preprint; 15 pages, 6 figures, 10 tables; Code at https://github.com/Xiangxu-0103/SuperFlow
☆ CoLLM: A Large Language Model for Composed Image Retrieval CVPR 2025
Composed Image Retrieval (CIR) is a complex task that aims to retrieve images based on a multimodal query. Typical training data consists of triplets containing a reference image, a textual description of desired modifications, and the target image, which are expensive and time-consuming to acquire. The scarcity of CIR datasets has led to zero-shot approaches utilizing synthetic triplets or leveraging vision-language models (VLMs) with ubiquitous web-crawled image-caption pairs. However, these methods have significant limitations: synthetic triplets suffer from limited scale, lack of diversity, and unnatural modification text, while image-caption pairs hinder joint embedding learning of the multimodal query due to the absence of triplet data. Moreover, existing approaches struggle with complex and nuanced modification texts that demand sophisticated fusion and understanding of vision and language modalities. We present CoLLM, a one-stop framework that effectively addresses these limitations. Our approach generates triplets on-the-fly from image-caption pairs, enabling supervised training without manual annotation. We leverage Large Language Models (LLMs) to generate joint embeddings of reference images and modification texts, facilitating deeper multimodal fusion. Additionally, we introduce Multi-Text CIR (MTCIR), a large-scale dataset comprising 3.4M samples, and refine existing CIR benchmarks (CIRR and Fashion-IQ) to enhance evaluation reliability. Experimental results demonstrate that CoLLM achieves state-of-the-art performance across multiple CIR benchmarks and settings. MTCIR yields competitive results, with up to 15% performance improvement. Our refined benchmarks provide more reliable evaluation metrics for CIR models, contributing to the advancement of this important field.
comment: CVPR 2025. Project page: https://collm-cvpr25.github.io/
☆ FullDiT: Multi-Task Video Generative Foundation Model with Full Attention
Current video generative foundation models primarily focus on text-to-video tasks, providing limited control for fine-grained video content creation. Although adapter-based approaches (e.g., ControlNet) enable additional controls with minimal fine-tuning, they encounter challenges when integrating multiple conditions, including: branch conflicts between independently trained adapters, parameter redundancy leading to increased computational cost, and suboptimal performance compared to full fine-tuning. To address these challenges, we introduce FullDiT, a unified foundation model for video generation that seamlessly integrates multiple conditions via unified full-attention mechanisms. By fusing multi-task conditions into a unified sequence representation and leveraging the long-context learning ability of full self-attention to capture condition dynamics, FullDiT reduces parameter overhead, avoids conditions conflict, and shows scalability and emergent ability. We further introduce FullBench for multi-task video generation evaluation. Experiments demonstrate that FullDiT achieves state-of-the-art results, highlighting the efficacy of full-attention in complex multi-task video generation.
comment: Project Page: https://fulldit.github.io/
☆ Self-Supervised Learning of Motion Concepts by Optimizing Counterfactuals
Estimating motion in videos is an essential computer vision problem with many downstream applications, including controllable video generation and robotics. Current solutions are primarily trained using synthetic data or require tuning of situation-specific heuristics, which inherently limits these models' capabilities in real-world contexts. Despite recent developments in large-scale self-supervised learning from videos, leveraging such representations for motion estimation remains relatively underexplored. In this work, we develop Opt-CWM, a self-supervised technique for flow and occlusion estimation from a pre-trained next-frame prediction model. Opt-CWM works by learning to optimize counterfactual probes that extract motion information from a base video model, avoiding the need for fixed heuristics while training on unrestricted video inputs. We achieve state-of-the-art performance for motion estimation on real-world videos while requiring no labeled data.
comment: Project webpage: https://neuroailab.github.io/opt_cwm_page/
☆ Tracktention: Leveraging Point Tracking to Attend Videos Faster and Better CVPR 2025
Temporal consistency is critical in video prediction to ensure that outputs are coherent and free of artifacts. Traditional methods, such as temporal attention and 3D convolution, may struggle with significant object motion and may not capture long-range temporal dependencies in dynamic scenes. To address this gap, we propose the Tracktention Layer, a novel architectural component that explicitly integrates motion information using point tracks, i.e., sequences of corresponding points across frames. By incorporating these motion cues, the Tracktention Layer enhances temporal alignment and effectively handles complex object motions, maintaining consistent feature representations over time. Our approach is computationally efficient and can be seamlessly integrated into existing models, such as Vision Transformers, with minimal modification. It can be used to upgrade image-only models to state-of-the-art video ones, sometimes outperforming models natively designed for video prediction. We demonstrate this on video depth prediction and video colorization, where models augmented with the Tracktention Layer exhibit significantly improved temporal consistency compared to baselines.
comment: CVPR 2025. Project website: zlai0.github.io/TrackTention
☆ Scaling Vision Pre-Training to 4K Resolution CVPR 2025
High-resolution perception of visual details is crucial for daily tasks. Current vision pre-training, however, is still limited to low resolutions (e.g., 378 x 378 pixels) due to the quadratic cost of processing larger images. We introduce PS3 that scales CLIP-style vision pre-training to 4K resolution with a near-constant cost. Instead of contrastive learning on global image representation, PS3 is pre-trained by selectively processing local regions and contrasting them with local detailed captions, enabling high-resolution representation learning with greatly reduced computational overhead. The pre-trained PS3 is able to both encode the global image at low resolution and selectively process local high-resolution regions based on their saliency or relevance to a text prompt. When applying PS3 to multi-modal LLM (MLLM), the resulting model, named VILA-HD, significantly improves high-resolution visual perception compared to baselines without high-resolution vision pre-training such as AnyRes and S^2 while using up to 4.3x fewer tokens. PS3 also unlocks appealing scaling properties of VILA-HD, including scaling up resolution for free and scaling up test-time compute for better performance. Compared to state of the arts, VILA-HD outperforms previous MLLMs such as NVILA and Qwen2-VL across multiple benchmarks and achieves better efficiency than latest token pruning approaches. Finally, we find current benchmarks do not require 4K-resolution perception, which motivates us to propose 4KPro, a new benchmark of image QA at 4K resolution, on which VILA-HD outperforms all previous MLLMs, including a 14.5% improvement over GPT-4o, and a 3.2% improvement and 2.96x speedup over Qwen2-VL.
comment: CVPR 2025. Project Page: https://nvlabs.github.io/PS3
☆ ICE: Intrinsic Concept Extraction from a Single Image via Diffusion Models CVPR 2025
The inherent ambiguity in defining visual concepts poses significant challenges for modern generative models, such as the diffusion-based Text-to-Image (T2I) models, in accurately learning concepts from a single image. Existing methods lack a systematic way to reliably extract the interpretable underlying intrinsic concepts. To address this challenge, we present ICE, short for Intrinsic Concept Extraction, a novel framework that exclusively utilizes a T2I model to automatically and systematically extract intrinsic concepts from a single image. ICE consists of two pivotal stages. In the first stage, ICE devises an automatic concept localization module to pinpoint relevant text-based concepts and their corresponding masks within the image. This critical stage streamlines concept initialization and provides precise guidance for subsequent analysis. The second stage delves deeper into each identified mask, decomposing the object-level concepts into intrinsic concepts and general concepts. This decomposition allows for a more granular and interpretable breakdown of visual elements. Our framework demonstrates superior performance on intrinsic concept extraction from a single image in an unsupervised manner. Project page: https://visual-ai.github.io/ice
comment: CVPR 2025, Project page: https://visual-ai.github.io/ice
☆ TokenHSI: Unified Synthesis of Physical Human-Scene Interactions through Task Tokenization CVPR 2025
Synthesizing diverse and physically plausible Human-Scene Interactions (HSI) is pivotal for both computer animation and embodied AI. Despite encouraging progress, current methods mainly focus on developing separate controllers, each specialized for a specific interaction task. This significantly hinders the ability to tackle a wide variety of challenging HSI tasks that require the integration of multiple skills, e.g., sitting down while carrying an object. To address this issue, we present TokenHSI, a single, unified transformer-based policy capable of multi-skill unification and flexible adaptation. The key insight is to model the humanoid proprioception as a separate shared token and combine it with distinct task tokens via a masking mechanism. Such a unified policy enables effective knowledge sharing across skills, thereby facilitating the multi-task training. Moreover, our policy architecture supports variable length inputs, enabling flexible adaptation of learned skills to new scenarios. By training additional task tokenizers, we can not only modify the geometries of interaction targets but also coordinate multiple skills to address complex tasks. The experiments demonstrate that our approach can significantly improve versatility, adaptability, and extensibility in various HSI tasks. Website: https://liangpan99.github.io/TokenHSI/
comment: CVPR 2025
☆ CAFe: Unifying Representation and Generation with Contrastive-Autoregressive Finetuning
The rapid advancement of large vision-language models (LVLMs) has driven significant progress in multimodal tasks, enabling models to interpret, reason, and generate outputs across both visual and textual domains. While excelling in generative tasks, existing LVLMs often face limitations in tasks requiring high-fidelity representation learning, such as generating image or text embeddings for retrieval. Recent work has proposed finetuning LVLMs for representational learning, but the fine-tuned model often loses its generative capabilities due to the representational learning training paradigm. To address this trade-off, we introduce CAFe, a contrastive-autoregressive fine-tuning framework that enhances LVLMs for both representation and generative tasks. By integrating a contrastive objective with autoregressive language modeling, our approach unifies these traditionally separate tasks, achieving state-of-the-art results in both multimodal retrieval and multimodal generative benchmarks, including object hallucination (OH) mitigation. CAFe establishes a novel framework that synergizes embedding and generative functionalities in a single model, setting a foundation for future multimodal models that excel in both retrieval precision and coherent output generation.
☆ Scaling Down Text Encoders of Text-to-Image Diffusion Models CVPR 2025
Text encoders in diffusion models have rapidly evolved, transitioning from CLIP to T5-XXL. Although this evolution has significantly enhanced the models' ability to understand complex prompts and generate text, it also leads to a substantial increase in the number of parameters. Despite T5 series encoders being trained on the C4 natural language corpus, which includes a significant amount of non-visual data, diffusion models with T5 encoder do not respond to those non-visual prompts, indicating redundancy in representational power. Therefore, it raises an important question: "Do we really need such a large text encoder?" In pursuit of an answer, we employ vision-based knowledge distillation to train a series of T5 encoder models. To fully inherit its capabilities, we constructed our dataset based on three criteria: image quality, semantic understanding, and text-rendering. Our results demonstrate the scaling down pattern that the distilled T5-base model can generate images of comparable quality to those produced by T5-XXL, while being 50 times smaller in size. This reduction in model size significantly lowers the GPU requirements for running state-of-the-art models such as FLUX and SD3, making high-quality text-to-image generation more accessible.
comment: accepted by CVPR 2025
☆ Visuo-Tactile Object Pose Estimation for a Multi-Finger Robot Hand with Low-Resolution In-Hand Tactile Sensing ICRA
Accurate 3D pose estimation of grasped objects is an important prerequisite for robots to perform assembly or in-hand manipulation tasks, but object occlusion by the robot's own hand greatly increases the difficulty of this perceptual task. Here, we propose that combining visual information and proprioception with binary, low-resolution tactile contact measurements from across the interior surface of an articulated robotic hand can mitigate this issue. The visuo-tactile object-pose-estimation problem is formulated probabilistically in a factor graph. The pose of the object is optimized to align with the three kinds of measurements using a robust cost function to reduce the influence of visual or tactile outlier readings. The advantages of the proposed approach are first demonstrated in simulation: a custom 15-DoF robot hand with one binary tactile sensor per link grasps 17 YCB objects while observed by an RGB-D camera. This low-resolution in-hand tactile sensing significantly improves object-pose estimates under high occlusion and also high visual noise. We also show these benefits through grasping tests with a preliminary real version of our tactile hand, obtaining reasonable visuo-tactile estimates of object pose at approximately 13.3 Hz on average.
comment: Accepted for publication at the IEEE International Conference on Robotics and Automation (ICRA), 2025
☆ Mask$^2$DiT: Dual Mask-based Diffusion Transformer for Multi-Scene Long Video Generation CVPR 2025
Sora has unveiled the immense potential of the Diffusion Transformer (DiT) architecture in single-scene video generation. However, the more challenging task of multi-scene video generation, which offers broader applications, remains relatively underexplored. To bridge this gap, we propose Mask$^2$DiT, a novel approach that establishes fine-grained, one-to-one alignment between video segments and their corresponding text annotations. Specifically, we introduce a symmetric binary mask at each attention layer within the DiT architecture, ensuring that each text annotation applies exclusively to its respective video segment while preserving temporal coherence across visual tokens. This attention mechanism enables precise segment-level textual-to-visual alignment, allowing the DiT architecture to effectively handle video generation tasks with a fixed number of scenes. To further equip the DiT architecture with the ability to generate additional scenes based on existing ones, we incorporate a segment-level conditional mask, which conditions each newly generated segment on the preceding video segments, thereby enabling auto-regressive scene extension. Both qualitative and quantitative experiments confirm that Mask$^2$DiT excels in maintaining visual consistency across segments while ensuring semantic alignment between each segment and its corresponding text description. Our project page is https://tianhao-qi.github.io/Mask2DiTProject.
comment: Accepted by CVPR 2025
☆ GENIUS: A Generative Framework for Universal Multimodal Search CVPR 2025
Generative retrieval is an emerging approach in information retrieval that generates identifiers (IDs) of target data based on a query, providing an efficient alternative to traditional embedding-based retrieval methods. However, existing models are task-specific and fall short of embedding-based retrieval in performance. This paper proposes GENIUS, a universal generative retrieval framework supporting diverse tasks across multiple modalities and domains. At its core, GENIUS introduces modality-decoupled semantic quantization, transforming multimodal data into discrete IDs encoding both modality and semantics. Moreover, to enhance generalization, we propose a query augmentation that interpolates between a query and its target, allowing GENIUS to adapt to varied query forms. Evaluated on the M-BEIR benchmark, it surpasses prior generative methods by a clear margin. Unlike embedding-based retrieval, GENIUS consistently maintains high retrieval speed across database size, with competitive performance across multiple benchmarks. With additional re-ranking, GENIUS often achieves results close to those of embedding-based methods while preserving efficiency.
comment: Accepted to CVPR 2025
☆ Unpaired Translation of Chest X-ray Images for Lung Opacity Diagnosis via Adaptive Activation Masks and Cross-Domain Alignment
Chest X-ray radiographs (CXRs) play a pivotal role in diagnosing and monitoring cardiopulmonary diseases. However, lung opac- ities in CXRs frequently obscure anatomical structures, impeding clear identification of lung borders and complicating the localization of pathology. This challenge significantly hampers segmentation accuracy and precise lesion identification, which are crucial for diagnosis. To tackle these issues, our study proposes an unpaired CXR translation framework that converts CXRs with lung opacities into counterparts without lung opacities while preserving semantic features. Central to our approach is the use of adaptive activation masks to selectively modify opacity regions in lung CXRs. Cross-domain alignment ensures translated CXRs without opacity issues align with feature maps and prediction labels from a pre-trained CXR lesion classifier, facilitating the interpretability of the translation process. We validate our method using RSNA, MIMIC-CXR-JPG and JSRT datasets, demonstrating superior translation quality through lower Frechet Inception Distance (FID) and Kernel Inception Distance (KID) scores compared to existing meth- ods (FID: 67.18 vs. 210.4, KID: 0.01604 vs. 0.225). Evaluation on RSNA opacity, MIMIC acute respiratory distress syndrome (ARDS) patient CXRs and JSRT CXRs show our method enhances segmentation accuracy of lung borders and improves lesion classification, further underscoring its potential in clinical settings (RSNA: mIoU: 76.58% vs. 62.58%, Sensitivity: 85.58% vs. 77.03%; MIMIC ARDS: mIoU: 86.20% vs. 72.07%, Sensitivity: 92.68% vs. 86.85%; JSRT: mIoU: 91.08% vs. 85.6%, Sensitivity: 97.62% vs. 95.04%). Our approach advances CXR imaging analysis, especially in investigating segmentation impacts through image translation techniques.
☆ FALCONEye: Finding Answers and Localizing Content in ONE-hour-long videos with multi-modal LLMs
Information retrieval in hour-long videos presents a significant challenge, even for state-of-the-art Vision-Language Models (VLMs), particularly when the desired information is localized within a small subset of frames. Long video data presents challenges for VLMs due to context window limitations and the difficulty of pinpointing frames containing the answer. Our novel video agent, FALCONEye, combines a VLM and a Large Language Model (LLM) to search relevant information along the video, and locate the frames with the answer. FALCONEye novelty relies on 1) the proposed meta-architecture, which is better suited to tackle hour-long videos compared to short video approaches in the state-of-the-art; 2) a new efficient exploration algorithm to locate the information using short clips, captions and answer confidence; and 3) our state-of-the-art VLMs calibration analysis for the answer confidence. Our agent is built over a small-size VLM and a medium-size LLM being accessible to run on standard computational resources. We also release FALCON-Bench, a benchmark to evaluate long (average > 1 hour) Video Answer Search challenges, highlighting the need for open-ended question evaluation. Our experiments show FALCONEye's superior performance than the state-of-the-art in FALCON-Bench, and similar or better performance in related benchmarks.
☆ Towards Online Multi-Modal Social Interaction Understanding
Multimodal social interaction understanding (MMSI) is critical in human-robot interaction systems. In real-world scenarios, AI agents are required to provide real-time feedback. However, existing models often depend on both past and future contexts, which hinders them from applying to real-world problems. To bridge this gap, we propose an online MMSI setting, where the model must resolve MMSI tasks using only historical information, such as recorded dialogues and video streams. To address the challenges of missing the useful future context, we develop a novel framework, named Online-MMSI-VLM, that leverages two complementary strategies: multi-party conversation forecasting and social-aware visual prompting with multi-modal large language models. First, to enrich linguistic context, the multi-party conversation forecasting simulates potential future utterances in a coarse-to-fine manner, anticipating upcoming speaker turns and then generating fine-grained conversational details. Second, to effectively incorporate visual social cues like gaze and gesture, social-aware visual prompting highlights the social dynamics in video with bounding boxes and body keypoints for each person and frame. Extensive experiments on three tasks and two datasets demonstrate that our method achieves state-of-the-art performance and significantly outperforms baseline models, indicating its effectiveness on Online-MMSI. The code and pre-trained models will be publicly released at: https://github.com/Sampson-Lee/OnlineMMSI.
☆ FireEdit: Fine-grained Instruction-based Image Editing via Region-aware Vision Language Model CVPR 2025
Currently, instruction-based image editing methods have made significant progress by leveraging the powerful cross-modal understanding capabilities of vision language models (VLMs). However, they still face challenges in three key areas: 1) complex scenarios; 2) semantic consistency; and 3) fine-grained editing. To address these issues, we propose FireEdit, an innovative Fine-grained Instruction-based image editing framework that exploits a REgion-aware VLM. FireEdit is designed to accurately comprehend user instructions and ensure effective control over the editing process. Specifically, we enhance the fine-grained visual perception capabilities of the VLM by introducing additional region tokens. Relying solely on the output of the LLM to guide the diffusion model may lead to suboptimal editing results. Therefore, we propose a Time-Aware Target Injection module and a Hybrid Visual Cross Attention module. The former dynamically adjusts the guidance strength at various denoising stages by integrating timestep embeddings with the text embeddings. The latter enhances visual details for image editing, thereby preserving semantic consistency between the edited result and the source image. By combining the VLM enhanced with fine-grained region tokens and the time-dependent diffusion model, FireEdit demonstrates significant advantages in comprehending editing instructions and maintaining high semantic consistency. Extensive experiments indicate that our approach surpasses the state-of-the-art instruction-based image editing methods. Our project is available at https://zjgans.github.io/fireedit.github.io.
comment: Accepted to CVPR 2025
☆ AudCast: Audio-Driven Human Video Generation by Cascaded Diffusion Transformers CVPR
Despite the recent progress of audio-driven video generation, existing methods mostly focus on driving facial movements, leading to non-coherent head and body dynamics. Moving forward, it is desirable yet challenging to generate holistic human videos with both accurate lip-sync and delicate co-speech gestures w.r.t. given audio. In this work, we propose AudCast, a generalized audio-driven human video generation framework adopting a cascade Diffusion-Transformers (DiTs) paradigm, which synthesizes holistic human videos based on a reference image and a given audio. 1) Firstly, an audio-conditioned Holistic Human DiT architecture is proposed to directly drive the movements of any human body with vivid gesture dynamics. 2) Then to enhance hand and face details that are well-knownly difficult to handle, a Regional Refinement DiT leverages regional 3D fitting as the bridge to reform the signals, producing the final results. Extensive experiments demonstrate that our framework generates high-fidelity audio-driven holistic human videos with temporal coherence and fine facial and hand details. Resources can be found at https://guanjz20.github.io/projects/AudCast.
comment: Accepted to IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2025. Project page: https://guanjz20.github.io/projects/AudCast
☆ GyralNet Subnetwork Partitioning via Differentiable Spectral Modularity Optimization
Understanding the structural and functional organization of the human brain requires a detailed examination of cortical folding patterns, among which the three-hinge gyrus (3HG) has been identified as a key structural landmark. GyralNet, a network representation of cortical folding, models 3HGs as nodes and gyral crests as edges, highlighting their role as critical hubs in cortico-cortical connectivity. However, existing methods for analyzing 3HGs face significant challenges, including the sub-voxel scale of 3HGs at typical neuroimaging resolutions, the computational complexity of establishing cross-subject correspondences, and the oversimplification of treating 3HGs as independent nodes without considering their community-level relationships. To address these limitations, we propose a fully differentiable subnetwork partitioning framework that employs a spectral modularity maximization optimization strategy to modularize the organization of 3HGs within GyralNet. By incorporating topological structural similarity and DTI-derived connectivity patterns as attribute features, our approach provides a biologically meaningful representation of cortical organization. Extensive experiments on the Human Connectome Project (HCP) dataset demonstrate that our method effectively partitions GyralNet at the individual level while preserving the community-level consistency of 3HGs across subjects, offering a robust foundation for understanding brain connectivity.
comment: 10 pages, 3 figures
☆ Domain-incremental White Blood Cell Classification with Privacy-aware Continual Learning
White blood cell (WBC) classification plays a vital role in hematology for diagnosing various medical conditions. However, it faces significant challenges due to domain shifts caused by variations in sample sources (e.g., blood or bone marrow) and differing imaging conditions across hospitals. Traditional deep learning models often suffer from catastrophic forgetting in such dynamic environments, while foundation models, though generally robust, experience performance degradation when the distribution of inference data differs from that of the training data. To address these challenges, we propose a generative replay-based Continual Learning (CL) strategy designed to prevent forgetting in foundation models for WBC classification. Our method employs lightweight generators to mimic past data with a synthetic latent representation to enable privacy-preserving replay. To showcase the effectiveness, we carry out extensive experiments with a total of four datasets with different task ordering and four backbone models including ResNet50, RetCCL, CTransPath, and UNI. Experimental results demonstrate that conventional fine-tuning methods degrade performance on previously learned tasks and struggle with domain shifts. In contrast, our continual learning strategy effectively mitigates catastrophic forgetting, preserving model performance across varying domains. This work presents a practical solution for maintaining reliable WBC classification in real-world clinical settings, where data distributions frequently evolve.
☆ ACVUBench: Audio-Centric Video Understanding Benchmark
Audio often serves as an auxiliary modality in video understanding tasks of audio-visual large language models (LLMs), merely assisting in the comprehension of visual information. However, a thorough understanding of videos significantly depends on auditory information, as audio offers critical context, emotional cues, and semantic meaning that visual data alone often lacks. This paper proposes an audio-centric video understanding benchmark (ACVUBench) to evaluate the video comprehension capabilities of multimodal LLMs with a particular focus on auditory information. Specifically, ACVUBench incorporates 2,662 videos spanning 18 different domains with rich auditory information, together with over 13k high-quality human annotated or validated question-answer pairs. Moreover, ACVUBench introduces a suite of carefully designed audio-centric tasks, holistically testing the understanding of both audio content and audio-visual interactions in videos. A thorough evaluation across a diverse range of open-source and proprietary multimodal LLMs is performed, followed by the analyses of deficiencies in audio-visual LLMs. Demos are available at https://github.com/lark-png/ACVUBench.
☆ LENVIZ: A High-Resolution Low-Exposure Night Vision Benchmark Dataset
Low-light image enhancement is crucial for a myriad of applications, from night vision and surveillance, to autonomous driving. However, due to the inherent limitations that come in hand with capturing images in low-illumination environments, the task of enhancing such scenes still presents a formidable challenge. To advance research in this field, we introduce our Low Exposure Night Vision (LENVIZ) Dataset, a comprehensive multi-exposure benchmark dataset for low-light image enhancement comprising of over 230K frames showcasing 24K real-world indoor and outdoor, with-and without human, scenes. Captured using 3 different camera sensors, LENVIZ offers a wide range of lighting conditions, noise levels, and scene complexities, making it the largest publicly available up-to 4K resolution benchmark in the field. LENVIZ includes high quality human-generated ground truth, for which each multi-exposure low-light scene has been meticulously curated and edited by expert photographers to ensure optimal image quality. Furthermore, we also conduct a comprehensive analysis of current state-of-the-art low-light image enhancement techniques on our dataset and highlight potential areas of improvement.
comment: Dataset will be released upon publication
☆ SeLIP: Similarity Enhanced Contrastive Language Image Pretraining for Multi-modal Head MRI
Despite that deep learning (DL) methods have presented tremendous potential in many medical image analysis tasks, the practical applications of medical DL models are limited due to the lack of enough data samples with manual annotations. By noting that the clinical radiology examinations are associated with radiology reports that describe the images, we propose to develop a foundation model for multi-model head MRI by using contrastive learning on the images and the corresponding radiology findings. In particular, a contrastive learning framework is proposed, where a mixed syntax and semantic similarity matching metric is integrated to reduce the thirst of extreme large dataset in conventional contrastive learning framework. Our proposed similarity enhanced contrastive language image pretraining (SeLIP) is able to effectively extract more useful features. Experiments revealed that our proposed SeLIP performs well in many downstream tasks including image-text retrieval task, classification task, and image segmentation, which highlights the importance of considering the similarities among texts describing different images in developing medical image foundation models.
☆ Unpaired Object-Level SAR-to-Optical Image Translation for Aircraft with Keypoints-Guided Diffusion Models
Synthetic Aperture Radar (SAR) imagery provides all-weather, all-day, and high-resolution imaging capabilities but its unique imaging mechanism makes interpretation heavily reliant on expert knowledge, limiting interpretability, especially in complex target tasks. Translating SAR images into optical images is a promising solution to enhance interpretation and support downstream tasks. Most existing research focuses on scene-level translation, with limited work on object-level translation due to the scarcity of paired data and the challenge of accurately preserving contour and texture details. To address these issues, this study proposes a keypoint-guided diffusion model (KeypointDiff) for SAR-to-optical image translation of unpaired aircraft targets. This framework introduces supervision on target class and azimuth angle via keypoints, along with a training strategy for unpaired data. Based on the classifier-free guidance diffusion architecture, a class-angle guidance module (CAGM) is designed to integrate class and angle information into the diffusion generation process. Furthermore, adversarial loss and consistency loss are employed to improve image fidelity and detail quality, tailored for aircraft targets. During sampling, aided by a pre-trained keypoint detector, the model eliminates the requirement for manually labeled class and azimuth information, enabling automated SAR-to-optical translation. Experimental results demonstrate that the proposed method outperforms existing approaches across multiple metrics, providing an efficient and effective solution for object-level SAR-to-optical translation and downstream tasks. Moreover, the method exhibits strong zero-shot generalization to untrained aircraft types with the assistance of the keypoint detector.
☆ PAVE: Patching and Adapting Video Large Language Models CVPR2025
Pre-trained video large language models (Video LLMs) exhibit remarkable reasoning capabilities, yet adapting these models to new tasks involving additional modalities or data types (e.g., audio or 3D information) remains challenging. In this paper, we present PAVE, a flexible framework for adapting pre-trained Video LLMs to downstream tasks with side-channel signals, such as audio, 3D cues, or multi-view videos. PAVE introduces lightweight adapters, referred to as "patches," which add a small number of parameters and operations to a base model without modifying its architecture or pre-trained weights. In doing so, PAVE can effectively adapt the pre-trained base model to support diverse downstream tasks, including audio-visual question answering, 3D reasoning, multi-view video recognition, and high frame rate video understanding. Across these tasks, PAVE significantly enhances the performance of the base model, surpassing state-of-the-art task-specific models while incurring a minor cost of ~0.1% additional FLOPs and parameters. Further, PAVE supports multi-task learning and generalizes well across different Video LLMs. Our code is available at https://github.com/dragonlzm/PAVE.
comment: CVPR2025 Camera Ready
☆ SITA: Structurally Imperceptible and Transferable Adversarial Attacks for Stylized Image Generation
Image generation technology has brought significant advancements across various fields but has also raised concerns about data misuse and potential rights infringements, particularly with respect to creating visual artworks. Current methods aimed at safeguarding artworks often employ adversarial attacks. However, these methods face challenges such as poor transferability, high computational costs, and the introduction of noticeable noise, which compromises the aesthetic quality of the original artwork. To address these limitations, we propose a Structurally Imperceptible and Transferable Adversarial (SITA) attacks. SITA leverages a CLIP-based destylization loss, which decouples and disrupts the robust style representation of the image. This disruption hinders style extraction during stylized image generation, thereby impairing the overall stylization process. Importantly, SITA eliminates the need for a surrogate diffusion model, leading to significantly reduced computational overhead. The method's robust style feature disruption ensures high transferability across diverse models. Moreover, SITA introduces perturbations by embedding noise within the imperceptible structural details of the image. This approach effectively protects against style extraction without compromising the visual quality of the artwork. Extensive experiments demonstrate that SITA offers superior protection for artworks against unauthorized use in stylized generation. It significantly outperforms existing methods in terms of transferability, computational efficiency, and noise imperceptibility. Code is available at https://github.com/A-raniy-day/SITA.
☆ Fine-Grained Erasure in Text-to-Image Diffusion-based Foundation Models CVPR 2025
Existing unlearning algorithms in text-to-image generative models often fail to preserve the knowledge of semantically related concepts when removing specific target concepts: a challenge known as adjacency. To address this, we propose FADE (Fine grained Attenuation for Diffusion Erasure), introducing adjacency aware unlearning in diffusion models. FADE comprises two components: (1) the Concept Neighborhood, which identifies an adjacency set of related concepts, and (2) Mesh Modules, employing a structured combination of Expungement, Adjacency, and Guidance loss components. These enable precise erasure of target concepts while preserving fidelity across related and unrelated concepts. Evaluated on datasets like Stanford Dogs, Oxford Flowers, CUB, I2P, Imagenette, and ImageNet1k, FADE effectively removes target concepts with minimal impact on correlated concepts, achieving atleast a 12% improvement in retention performance over state-of-the-art methods.
comment: Published in CVPR 2025
☆ LPOSS: Label Propagation Over Patches and Pixels for Open-vocabulary Semantic Segmentation
We propose a training-free method for open-vocabulary semantic segmentation using Vision-and-Language Models (VLMs). Our approach enhances the initial per-patch predictions of VLMs through label propagation, which jointly optimizes predictions by incorporating patch-to-patch relationships. Since VLMs are primarily optimized for cross-modal alignment and not for intra-modal similarity, we use a Vision Model (VM) that is observed to better capture these relationships. We address resolution limitations inherent to patch-based encoders by applying label propagation at the pixel level as a refinement step, significantly improving segmentation accuracy near class boundaries. Our method, called LPOSS+, performs inference over the entire image, avoiding window-based processing and thereby capturing contextual interactions across the full image. LPOSS+ achieves state-of-the-art performance among training-free methods, across a diverse set of datasets. Code: https://github.com/vladan-stojnic/LPOSS
☆ Resilient Sensor Fusion under Adverse Sensor Failures via Multi-Modal Expert Fusion CVPR 2025
Modern autonomous driving perception systems utilize complementary multi-modal sensors, such as LiDAR and cameras. Although sensor fusion architectures enhance performance in challenging environments, they still suffer significant performance drops under severe sensor failures, such as LiDAR beam reduction, LiDAR drop, limited field of view, camera drop, and occlusion. This limitation stems from inter-modality dependencies in current sensor fusion frameworks. In this study, we introduce an efficient and robust LiDAR-camera 3D object detector, referred to as MoME, which can achieve robust performance through a mixture of experts approach. Our MoME fully decouples modality dependencies using three parallel expert decoders, which use camera features, LiDAR features, or a combination of both to decode object queries, respectively. We propose Multi-Expert Decoding (MED) framework, where each query is decoded selectively using one of three expert decoders. MoME utilizes an Adaptive Query Router (AQR) to select the most appropriate expert decoder for each query based on the quality of camera and LiDAR features. This ensures that each query is processed by the best-suited expert, resulting in robust performance across diverse sensor failure scenarios. We evaluated the performance of MoME on the nuScenes-R benchmark. Our MoME achieved state-of-the-art performance in extreme weather and sensor failure conditions, significantly outperforming the existing models across various sensor failure scenarios.
comment: Accepted to CVPR 2025
☆ BiPrompt-SAM: Enhancing Image Segmentation via Explicit Selection between Point and Text Prompts
Segmentation is a fundamental task in computer vision, with prompt-driven methods gaining prominence due to their flexibility. The recent Segment Anything Model (SAM) has demonstrated powerful point-prompt segmentation capabilities, while text-based segmentation models offer rich semantic understanding. However, existing approaches rarely explore how to effectively combine these complementary modalities for optimal segmentation performance. This paper presents BiPrompt-SAM, a novel dual-modal prompt segmentation framework that fuses the advantages of point and text prompts through an explicit selection mechanism. Specifically, we leverage SAM's inherent ability to generate multiple mask candidates, combined with a semantic guidance mask from text prompts, and explicitly select the most suitable candidate based on similarity metrics. This approach can be viewed as a simplified Mixture of Experts (MoE) system, where the point and text modules act as distinct "experts," and the similarity scoring serves as a rudimentary "gating network." We conducted extensive evaluations on both the Endovis17 medical dataset and RefCOCO series natural image datasets. On Endovis17, BiPrompt-SAM achieved 89.55\% mDice and 81.46\% mIoU, comparable to state-of-the-art specialized medical segmentation models. On the RefCOCO series datasets, our method attained 87.1\%, 86.5\%, and 85.8\% IoU, significantly outperforming existing approaches. Experiments demonstrate that our explicit dual-selection method effectively combines the spatial precision of point prompts with the semantic richness of text prompts, particularly excelling in scenarios involving semantically complex objects, multiple similar objects, and partial occlusions. BiPrompt-SAM not only provides a simple yet effective implementation but also offers a new perspective on multi-modal prompt fusion.
☆ Test-Time Reasoning Through Visual Human Preferences with VLMs and Soft Rewards
Can Visual Language Models (VLMs) effectively capture human visual preferences? This work addresses this question by training VLMs to think about preferences at test time, employing reinforcement learning methods inspired by DeepSeek R1 and OpenAI O1. Using datasets such as ImageReward and Human Preference Score v2 (HPSv2), our models achieve accuracies of 64.9% on the ImageReward test set (trained on ImageReward official split) and 65.4% on HPSv2 (trained on approximately 25% of its data). These results match traditional encoder-based models while providing transparent reasoning and enhanced generalization. This approach allows to use not only rich VLM world knowledge, but also its potential to think, yielding interpretable outcomes that help decision-making processes. By demonstrating that human visual preferences reasonable by current VLMs, we introduce efficient soft-reward strategies for image ranking, outperforming simplistic selection or scoring methods. This reasoning capability enables VLMs to rank arbitrary images-regardless of aspect ratio or complexity-thereby potentially amplifying the effectiveness of visual Preference Optimization. By reducing the need for extensive markup while improving reward generalization and explainability, our findings can be a strong mile-stone that will enhance text-to-vision models even further.
☆ OpenLex3D: A New Evaluation Benchmark for Open-Vocabulary 3D Scene Representations
3D scene understanding has been transformed by open-vocabulary language models that enable interaction via natural language. However, the evaluation of these representations is limited to closed-set semantics that do not capture the richness of language. This work presents OpenLex3D, a dedicated benchmark to evaluate 3D open-vocabulary scene representations. OpenLex3D provides entirely new label annotations for 23 scenes from Replica, ScanNet++, and HM3D, which capture real-world linguistic variability by introducing synonymical object categories and additional nuanced descriptions. By introducing an open-set 3D semantic segmentation task and an object retrieval task, we provide insights on feature precision, segmentation, and downstream capabilities. We evaluate various existing 3D open-vocabulary methods on OpenLex3D, showcasing failure cases, and avenues for improvement. The benchmark is publicly available at: https://openlex3d.github.io/.
☆ Dita: Scaling Diffusion Transformer for Generalist Vision-Language-Action Policy
While recent vision-language-action models trained on diverse robot datasets exhibit promising generalization capabilities with limited in-domain data, their reliance on compact action heads to predict discretized or continuous actions constrains adaptability to heterogeneous action spaces. We present Dita, a scalable framework that leverages Transformer architectures to directly denoise continuous action sequences through a unified multimodal diffusion process. Departing from prior methods that condition denoising on fused embeddings via shallow networks, Dita employs in-context conditioning -- enabling fine-grained alignment between denoised actions and raw visual tokens from historical observations. This design explicitly models action deltas and environmental nuances. By scaling the diffusion action denoiser alongside the Transformer's scalability, Dita effectively integrates cross-embodiment datasets across diverse camera perspectives, observation scenes, tasks, and action spaces. Such synergy enhances robustness against various variances and facilitates the successful execution of long-horizon tasks. Evaluations across extensive benchmarks demonstrate state-of-the-art or comparative performance in simulation. Notably, Dita achieves robust real-world adaptation to environmental variances and complex long-horizon tasks through 10-shot finetuning, using only third-person camera inputs. The architecture establishes a versatile, lightweight and open-source baseline for generalist robot policy learning. Project Page: https://robodita.github.io.
comment: Preprint; https://robodita.github.io;
☆ Vanishing Depth: A Depth Adapter with Positional Depth Encoding for Generalized Image Encoders
Generalized metric depth understanding is critical for precise vision-guided robotics, which current state-of-the-art (SOTA) vision-encoders do not support. To address this, we propose Vanishing Depth, a self-supervised training approach that extends pretrained RGB encoders to incorporate and align metric depth into their feature embeddings. Based on our novel positional depth encoding, we enable stable depth density and depth distribution invariant feature extraction. We achieve performance improvements and SOTA results across a spectrum of relevant RGBD downstream tasks - without the necessity of finetuning the encoder. Most notably, we achieve 56.05 mIoU on SUN-RGBD segmentation, 88.3 RMSE on Void's depth completion, and 83.8 Top 1 accuracy on NYUv2 scene classification. In 6D-object pose estimation, we outperform our predecessors of DinoV2, EVA-02, and Omnivore and achieve SOTA results for non-finetuned encoders in several related RGBD downstream tasks.
comment: Preprint
☆ ORION: A Holistic End-to-End Autonomous Driving Framework by Vision-Language Instructed Action Generation
End-to-end (E2E) autonomous driving methods still struggle to make correct decisions in interactive closed-loop evaluation due to limited causal reasoning capability. Current methods attempt to leverage the powerful understanding and reasoning abilities of Vision-Language Models (VLMs) to resolve this dilemma. However, the problem is still open that few VLMs for E2E methods perform well in the closed-loop evaluation due to the gap between the semantic reasoning space and the purely numerical trajectory output in the action space. To tackle this issue, we propose ORION, a holistic E2E autonomous driving framework by vision-language instructed action generation. ORION uniquely combines a QT-Former to aggregate long-term history context, a Large Language Model (LLM) for driving scenario reasoning, and a generative planner for precision trajectory prediction. ORION further aligns the reasoning space and the action space to implement a unified E2E optimization for both visual question-answering (VQA) and planning tasks. Our method achieves an impressive closed-loop performance of 77.74 Driving Score (DS) and 54.62% Success Rate (SR) on the challenge Bench2Drive datasets, which outperforms state-of-the-art (SOTA) methods by a large margin of 14.28 DS and 19.61% SR.
☆ Surg-3M: A Dataset and Foundation Model for Perception in Surgical Settings
Advancements in computer-assisted surgical procedures heavily rely on accurate visual data interpretation from camera systems used during surgeries. Traditional open-access datasets focusing on surgical procedures are often limited by their small size, typically consisting of fewer than 100 videos with less than 100K images. To address these constraints, a new dataset called Surg-3M has been compiled using a novel aggregation pipeline that collects high-resolution videos from online sources. Featuring an extensive collection of over 4K surgical videos and more than 3 million high-quality images from multiple procedure types, Surg-3M offers a comprehensive resource surpassing existing alternatives in size and scope, including two novel tasks. To demonstrate the effectiveness of this dataset, we present SurgFM, a self-supervised foundation model pretrained on Surg-3M that achieves impressive results in downstream tasks such as surgical phase recognition, action recognition, and tool presence detection. Combining key components from ConvNeXt, DINO, and an innovative augmented distillation method, SurgFM exhibits exceptional performance compared to specialist architectures across various benchmarks. Our experimental results show that SurgFM outperforms state-of-the-art models in multiple downstream tasks, including significant gains in surgical phase recognition (+8.9pp, +4.7pp, and +3.9pp of Jaccard in AutoLaparo, M2CAI16, and Cholec80), action recognition (+3.1pp of mAP in CholecT50) and tool presence detection (+4.6pp of mAP in Cholec80). Moreover, even when using only half of the data, SurgFM outperforms state-of-the-art models in AutoLaparo and achieves state-of-the-art performance in Cholec80. Both Surg-3M and SurgFM have significant potential to accelerate progress towards developing autonomous robotic surgery systems.
comment: 15 pages
☆ GRN+: A Simplified Generative Reinforcement Network for Tissue Layer Analysis in 3D Ultrasound Images for Chronic Low-back Pain
3D ultrasound delivers high-resolution, real-time images of soft tissues, which is essential for pain research. However, manually distinguishing various tissues for quantitative analysis is labor-intensive. To streamline this process, we developed and validated GRN+, a novel multi-model framework that automates layer segmentation with minimal annotated data. GRN+ combines a ResNet-based generator and a U-Net segmentation model. Through a method called Segmentation-guided Enhancement (SGE), the generator produces new images and matching masks under the guidance of the segmentation model, with its weights adjusted according to the segmentation loss gradient. To prevent gradient explosion and secure stable training, a two-stage backpropagation strategy was implemented: the first stage propagates the segmentation loss through both the generator and segmentation model, while the second stage concentrates on optimizing the segmentation model alone, thereby refining mask prediction using the generated images. Tested on 69 fully annotated 3D ultrasound scans from 29 subjects with six manually labeled tissue layers, GRN+ outperformed all other semi-supervised methods in terms of the Dice coefficient using only 5% labeled data, despite not using unlabeled data for unsupervised training. Additionally, when applied to fully annotated datasets, GRN+ with SGE achieved a 2.16% higher Dice coefficient while incurring lower computational costs compared to other models. Overall, GRN+ provides accurate tissue segmentation while reducing both computational expenses and the dependency on extensive annotations, making it an effective tool for 3D ultrasound analysis in cLBP patients.
☆ InterSliceBoost: Identifying Tissue Layers in Three-dimensional Ultrasound Images for Chronic Lower Back Pain (cLBP) Assessment
Available studies on chronic lower back pain (cLBP) typically focus on one or a few specific tissues rather than conducting a comprehensive layer-by-layer analysis. Since three-dimensional (3-D) images often contain hundreds of slices, manual annotation of these anatomical structures is both time-consuming and error-prone. We aim to develop and validate a novel approach called InterSliceBoost to enable the training of a segmentation model on a partially annotated dataset without compromising segmentation performance. The architecture of InterSliceBoost includes two components: an inter-slice generator and a segmentation model. The generator utilizes residual block-based encoders to extract features from adjacent image-mask pairs (IMPs). Differential features are calculated and input into a decoder to generate inter-slice IMPs. The segmentation model is trained on partially annotated datasets (e.g., skipping 1, 2, 3, or 7 images) and the generated inter-slice IMPs. To validate the performance of InterSliceBoost, we utilized a dataset of 76 B-mode ultrasound scans acquired on 29 subjects enrolled in an ongoing cLBP study. InterSliceBoost, trained on only 33% of the image slices, achieved a mean Dice coefficient of 80.84% across all six layers on the independent test set, with Dice coefficients of 73.48%, 61.11%, 81.87%, 95.74%, 83.52% and 88.74% for segmenting dermis, superficial fat, superficial fascial membrane, deep fat, deep fascial membrane, and muscle. This performance is significantly higher than the conventional model trained on fully annotated images (p<0.05). InterSliceBoost can effectively segment the six tissue layers depicted on 3-D B-model ultrasound images in settings with partial annotations.
☆ PCM : Picard Consistency Model for Fast Parallel Sampling of Diffusion Models CVPR 2025
Recently, diffusion models have achieved significant advances in vision, text, and robotics. However, they still face slow generation speeds due to sequential denoising processes. To address this, a parallel sampling method based on Picard iteration was introduced, effectively reducing sequential steps while ensuring exact convergence to the original output. Nonetheless, Picard iteration does not guarantee faster convergence, which can still result in slow generation in practice. In this work, we propose a new parallelization scheme, the Picard Consistency Model (PCM), which significantly reduces the number of generation steps in Picard iteration. Inspired by the consistency model, PCM is directly trained to predict the fixed-point solution, or the final output, at any stage of the convergence trajectory. Additionally, we introduce a new concept called model switching, which addresses PCM's limitations and ensures exact convergence. Extensive experiments demonstrate that PCM achieves up to a 2.71x speedup over sequential sampling and a 1.77x speedup over Picard iteration across various tasks, including image generation and robotic control.
comment: Accepted to the CVPR 2025
☆ MindfulLIME: A Stable Solution for Explanations of Machine Learning Models with Enhanced Localization Precision -- A Medical Image Case Study
Ensuring transparency in machine learning decisions is critically important, especially in sensitive sectors such as healthcare, finance, and justice. Despite this, some popular explainable algorithms, such as Local Interpretable Model-agnostic Explanations (LIME), often produce unstable explanations due to the random generation of perturbed samples. Random perturbation introduces small changes or noise to modified instances of the original data, leading to inconsistent explanations. Even slight variations in the generated samples significantly affect the explanations provided by such models, undermining trust and hindering the adoption of interpretable models. To address this challenge, we propose MindfulLIME, a novel algorithm that intelligently generates purposive samples using a graph-based pruning algorithm and uncertainty sampling. MindfulLIME substantially improves the consistency of visual explanations compared to random sampling approaches. Our experimental evaluation, conducted on a widely recognized chest X-ray dataset, confirms MindfulLIME's stability with a 100% success rate in delivering reliable explanations under identical conditions. Additionally, MindfulLIME improves the localization precision of visual explanations by reducing the distance between the generated explanations and the actual local annotations compared to LIME. We also performed comprehensive experiments considering various segmentation algorithms and sample numbers, focusing on stability, quality, and efficiency. The results demonstrate the outstanding performance of MindfulLIME across different segmentation settings, generating fewer high-quality samples within a reasonable processing time. By addressing the stability limitations of LIME in image data, MindfulLIME enhances the trustworthiness and interpretability of machine learning models in specific medical imaging applications, a critical domain.
☆ EventMamba: Enhancing Spatio-Temporal Locality with State Space Models for Event-Based Video Reconstruction
Leveraging its robust linear global modeling capability, Mamba has notably excelled in computer vision. Despite its success, existing Mamba-based vision models have overlooked the nuances of event-driven tasks, especially in video reconstruction. Event-based video reconstruction (EBVR) demands spatial translation invariance and close attention to local event relationships in the spatio-temporal domain. Unfortunately, conventional Mamba algorithms apply static window partitions and standard reshape scanning methods, leading to significant losses in local connectivity. To overcome these limitations, we introduce EventMamba--a specialized model designed for EBVR tasks. EventMamba innovates by incorporating random window offset (RWO) in the spatial domain, moving away from the restrictive fixed partitioning. Additionally, it features a new consistent traversal serialization approach in the spatio-temporal domain, which maintains the proximity of adjacent events both spatially and temporally. These enhancements enable EventMamba to retain Mamba's robust modeling capabilities while significantly preserving the spatio-temporal locality of event data. Comprehensive testing on multiple datasets shows that EventMamba markedly enhances video reconstruction, drastically improving computation speed while delivering superior visual quality compared to Transformer-based methods.
☆ On What Depends the Robustness of Multi-source Models to Missing Data in Earth Observation?
In recent years, the development of robust multi-source models has emerged in the Earth Observation (EO) field. These are models that leverage data from diverse sources to improve predictive accuracy when there is missing data. Despite these advancements, the factors influencing the varying effectiveness of such models remain poorly understood. In this study, we evaluate the predictive performance of six state-of-the-art multi-source models in predicting scenarios where either a single data source is missing or only a single source is available. Our analysis reveals that the efficacy of these models is intricately tied to the nature of the task, the complementarity among data sources, and the model design. Surprisingly, we observe instances where the removal of certain data sources leads to improved predictive performance, challenging the assumption that incorporating all available data is always beneficial. These findings prompt critical reflections on model complexity and the necessity of all collected data sources, potentially shaping the way for more streamlined approaches in EO applications.
comment: Accepted at IEEE International Geoscience and Remote Sensing Symposium 2025
☆ Semi-SD: Semi-Supervised Metric Depth Estimation via Surrounding Cameras for Autonomous Driving
In this paper, we introduce Semi-SD, a novel metric depth estimation framework tailored for surrounding cameras equipment in autonomous driving. In this work, the input data consists of adjacent surrounding frames and camera parameters. We propose a unified spatial-temporal-semantic fusion module to construct the visual fused features. Cross-attention components for surrounding cameras and adjacent frames are utilized to focus on metric scale information refinement and temporal feature matching. Building on this, we propose a pose estimation framework using surrounding cameras, their corresponding estimated depths, and extrinsic parameters, which effectively address the scale ambiguity in multi-camera setups. Moreover, semantic world model and monocular depth estimation world model are integrated to supervised the depth estimation, which improve the quality of depth estimation. We evaluate our algorithm on DDAD and nuScenes datasets, and the results demonstrate that our method achieves state-of-the-art performance in terms of surrounding camera based depth estimation quality. The source code will be available on https://github.com/xieyuser/Semi-SD.
☆ Mind the Gap: Benchmarking Spatial Reasoning in Vision-Language Models
Vision-Language Models (VLMs) have recently emerged as powerful tools, excelling in tasks that integrate visual and textual comprehension, such as image captioning, visual question answering, and image-text retrieval. However, existing benchmarks for VLMs include spatial components, which often fail to isolate spatial reasoning from related tasks such as object detection or semantic comprehension. In this paper, we address these deficiencies with a multi-faceted approach towards understanding spatial reasoning. Informed by the diverse and multi-dimensional nature of human spatial reasoning abilities, we present a detailed analysis that first delineates the core elements of spatial reasoning: spatial relations, orientation and navigation, mental rotation, and spatial visualization, and then assesses the performance of these models in both synthetic and real-world images, bridging controlled and naturalistic contexts. We analyze 13 state-of-the-art Vision-Language Models, uncovering pivotal insights into their spatial reasoning performance. Our results reveal profound shortcomings in current VLMs, with average accuracy across the 13 models approximating random chance, highlighting spatial reasoning as a persistent obstacle. This work not only exposes the pressing need to advance spatial reasoning within VLMs but also establishes a solid platform for future exploration. Code available on GitHub (https://github.com/stogiannidis/srbench) and dataset available on HuggingFace (https://huggingface.co/datasets/stogiannidis/srbench).
comment: 8 main pages, 4 pages Appendix, 5 figures
☆ Bootstrap Your Own Views: Masked Ego-Exo Modeling for Fine-grained View-invariant Video Representations CVPR 2025
View-invariant representation learning from egocentric (first-person, ego) and exocentric (third-person, exo) videos is a promising approach toward generalizing video understanding systems across multiple viewpoints. However, this area has been underexplored due to the substantial differences in perspective, motion patterns, and context between ego and exo views. In this paper, we propose a novel masked ego-exo modeling that promotes both causal temporal dynamics and cross-view alignment, called Bootstrap Your Own Views (BYOV), for fine-grained view-invariant video representation learning from unpaired ego-exo videos. We highlight the importance of capturing the compositional nature of human actions as a basis for robust cross-view understanding. Specifically, self-view masking and cross-view masking predictions are designed to learn view-invariant and powerful representations concurrently. Experimental results demonstrate that our BYOV significantly surpasses existing approaches with notable gains across all metrics in four downstream ego-exo video tasks. The code is available at https://github.com/park-jungin/byov.
comment: CVPR 2025 Camera-ready
☆ High-Quality Spatial Reconstruction and Orthoimage Generation Using Efficient 2D Gaussian Splatting
Highly accurate geometric precision and dense image features characterize True Digital Orthophoto Maps (TDOMs), which are in great demand for applications such as urban planning, infrastructure management, and environmental monitoring. Traditional TDOM generation methods need sophisticated processes, such as Digital Surface Models (DSM) and occlusion detection, which are computationally expensive and prone to errors. This work presents an alternative technique rooted in 2D Gaussian Splatting (2DGS), free of explicit DSM and occlusion detection. With depth map generation, spatial information for every pixel within the TDOM is retrieved and can reconstruct the scene with high precision. Divide-and-conquer strategy achieves excellent GS training and rendering with high-resolution TDOMs at a lower resource cost, which preserves higher quality of rendering on complex terrain and thin structure without a decrease in efficiency. Experimental results demonstrate the efficiency of large-scale scene reconstruction and high-precision terrain modeling. This approach provides accurate spatial data, which assists users in better planning and decision-making based on maps.
☆ Optimization of MedSAM model based on bounding box adaptive perturbation algorithm
The MedSAM model, built upon the SAM framework, enhances medical image segmentation through generalizable training but still exhibits notable limitations. First, constraints in the perturbation window settings during training can cause MedSAM to incorrectly segment small tissues or organs together with adjacent structures, leading to segmentation errors. Second, when dealing with medical image targets characterized by irregular shapes and complex structures, segmentation often relies on narrowing the bounding box to refine segmentation intent. However, MedSAM's performance under reduced bounding box prompts remains suboptimal. To address these challenges, this study proposes a bounding box adaptive perturbation algorithm to optimize the training process. The proposed approach aims to reduce segmentation errors for small targets and enhance the model's accuracy when processing reduced bounding box prompts, ultimately improving the robustness and reliability of the MedSAM model for complex medical imaging tasks.
comment: 6 pages, 6 figures, 3 Tables
♻ ☆ Aether: Geometric-Aware Unified World Modeling
The integration of geometric reconstruction and generative modeling remains a critical challenge in developing AI systems capable of human-like spatial reasoning. This paper proposes Aether, a unified framework that enables geometry-aware reasoning in world models by jointly optimizing three core capabilities: (1) 4D dynamic reconstruction, (2) action-conditioned video prediction, and (3) goal-conditioned visual planning. Through task-interleaved feature learning, Aether achieves synergistic knowledge sharing across reconstruction, prediction, and planning objectives. Building upon video generation models, our framework demonstrates unprecedented synthetic-to-real generalization despite never observing real-world data during training. Furthermore, our approach achieves zero-shot generalization in both action following and reconstruction tasks, thanks to its intrinsic geometric modeling. Remarkably, even without real-world data, its reconstruction performance is comparable with or even better than that of domain-specific models. Additionally, Aether employs camera trajectories as geometry-informed action spaces, enabling effective action-conditioned prediction and visual planning. We hope our work inspires the community to explore new frontiers in physically-reasonable world modeling and its applications.
comment: Project Page: https://aether-world.github.io/
♻ ☆ HunyuanPortrait: Implicit Condition Control for Enhanced Portrait Animation CVPR 2025
We introduce HunyuanPortrait, a diffusion-based condition control method that employs implicit representations for highly controllable and lifelike portrait animation. Given a single portrait image as an appearance reference and video clips as driving templates, HunyuanPortrait can animate the character in the reference image by the facial expression and head pose of the driving videos. In our framework, we utilize pre-trained encoders to achieve the decoupling of portrait motion information and identity in videos. To do so, implicit representation is adopted to encode motion information and is employed as control signals in the animation phase. By leveraging the power of stable video diffusion as the main building block, we carefully design adapter layers to inject control signals into the denoising unet through attention mechanisms. These bring spatial richness of details and temporal consistency. HunyuanPortrait also exhibits strong generalization performance, which can effectively disentangle appearance and motion under different image styles. Our framework outperforms existing methods, demonstrating superior temporal consistency and controllability. Our project is available at https://kkakkkka.github.io/HunyuanPortrait.
comment: Accepted to CVPR 2025
♻ ☆ MC-LLaVA: Multi-Concept Personalized Vision-Language Model
Current vision-language models (VLMs) show exceptional abilities across diverse tasks, such as visual question answering. To enhance user experience, recent studies investigate VLM personalization to understand user-provided concepts. However, they mainly focus on single-concept personalization, neglecting the existence and interplay of multiple concepts, which limits real-world applicability. This paper proposes the first multi-concept personalization paradigm, MC-LLaVA. Specifically, MC-LLaVA employs a multi-concept instruction tuning strategy, effectively integrating multiple concepts in a single training step. To reduce the costs related to joint training, we propose a personalized textual prompt that uses visual token information to initialize concept tokens. Additionally, we introduce a personalized visual prompt during inference, aggregating location confidence maps for enhanced recognition and grounding capabilities. To advance multi-concept personalization research, we further contribute a high-quality instruction tuning dataset. We carefully collect images with multiple characters and objects from movies and manually generate question-answer samples for multi-concept scenarios, featuring superior diversity. Comprehensive qualitative and quantitative experiments demonstrate that MC-LLaVA can achieve impressive multi-concept personalized responses, paving the way for VLMs to become better user-specific assistants. The code and dataset will be publicly available at https://github.com/arctanxarc/MC-LLaVA}.
comment: I sincerely apologize for any inconvenience caused. We actually uploaded this paper to arXiv in November 2024, as arXiv:2411.11706. During this update, we did not consider the replacement operation of arXiv, which led to duplicate submissions. We have made modifications at the original address arXiv:2411.11706
♻ ☆ Learning to segment anatomy and lesions from disparately labeled sources in brain MRI
Segmenting healthy tissue structures alongside lesions in brain Magnetic Resonance Images (MRI) remains a challenge for today's algorithms due to lesion-caused disruption of the anatomy and lack of jointly labeled training datasets, where both healthy tissues and lesions are labeled on the same images. In this paper, we propose a method that is robust to lesion-caused disruptions and can be trained from disparately labeled training sets, i.e., without requiring jointly labeled samples, to automatically segment both. In contrast to prior work, we decouple healthy tissue and lesion segmentation in two paths to leverage multi-sequence acquisitions and merge information with an attention mechanism. During inference, an image-specific adaptation reduces adverse influences of lesion regions on healthy tissue predictions. During training, the adaptation is taken into account through meta-learning and co-training is used to learn from disparately labeled training images. Our model shows an improved performance on several anatomical structures and lesions on a publicly available brain glioblastoma dataset compared to the state-of-the-art segmentation methods.
♻ ☆ Lightweight Embedded FPGA Deployment of Learned Image Compression with Knowledge Distillation and Hybrid Quantization
Learnable Image Compression (LIC) has shown the potential to outperform standardized video codecs in RD efficiency, prompting the research for hardware-friendly implementations. Most existing LIC hardware implementations prioritize latency to RD-efficiency and through an extensive exploration of the hardware design space. We present a novel design paradigm where the burden of tuning the design for a specific hardware platform is shifted towards model dimensioning and without compromising on RD-efficiency. First, we design a framework for distilling a leaner student LIC model from a reference teacher: by tuning a single model hyperparameters, we can meet the constraints of different hardware platforms without a complex hardware design exploration. Second, we propose a hardware-friendly implementation of the Generalized Divisive Normalization - GDN activation that preserves RD efficiency even post parameter quantization. Third, we design a pipelined FPGA configuration which takes full advantage of available FPGA resources by leveraging parallel processing and optimizing resource allocation. Our experiments with a state of the art LIC model show that we outperform all existing FPGA implementations while performing very close to the original model.
comment: 1. Submitted to IEEE Transactions on Circuits and Systems for Video Technology in March 2025. 2. Corrected numerous mistakes from previous versions in results, citations and metrics numbers in figures
♻ ☆ Frequency Dynamic Convolution for Dense Image Prediction CVPR 2025
While Dynamic Convolution (DY-Conv) has shown promising performance by enabling adaptive weight selection through multiple parallel weights combined with an attention mechanism, the frequency response of these weights tends to exhibit high similarity, resulting in high parameter costs but limited adaptability. In this work, we introduce Frequency Dynamic Convolution (FDConv), a novel approach that mitigates these limitations by learning a fixed parameter budget in the Fourier domain. FDConv divides this budget into frequency-based groups with disjoint Fourier indices, enabling the construction of frequency-diverse weights without increasing the parameter cost. To further enhance adaptability, we propose Kernel Spatial Modulation (KSM) and Frequency Band Modulation (FBM). KSM dynamically adjusts the frequency response of each filter at the spatial level, while FBM decomposes weights into distinct frequency bands in the frequency domain and modulates them dynamically based on local content. Extensive experiments on object detection, segmentation, and classification validate the effectiveness of FDConv. We demonstrate that when applied to ResNet-50, FDConv achieves superior performance with a modest increase of +3.6M parameters, outperforming previous methods that require substantial increases in parameter budgets (e.g., CondConv +90M, KW +76.5M). Moreover, FDConv seamlessly integrates into a variety of architectures, including ConvNeXt, Swin-Transformer, offering a flexible and efficient solution for modern vision tasks. The code is made publicly available at https://github.com/Linwei-Chen/FDConv.
comment: Accepted by CVPR 2025
♻ ☆ Re-HOLD: Video Hand Object Interaction Reenactment via adaptive Layout-instructed Diffusion Model CVPR 2025
Current digital human studies focusing on lip-syncing and body movement are no longer sufficient to meet the growing industrial demand, while human video generation techniques that support interacting with real-world environments (e.g., objects) have not been well investigated. Despite human hand synthesis already being an intricate problem, generating objects in contact with hands and their interactions presents an even more challenging task, especially when the objects exhibit obvious variations in size and shape. To tackle these issues, we present a novel video Reenactment framework focusing on Human-Object Interaction (HOI) via an adaptive Layout-instructed Diffusion model (Re-HOLD). Our key insight is to employ specialized layout representation for hands and objects, respectively. Such representations enable effective disentanglement of hand modeling and object adaptation to diverse motion sequences. To further improve the generation quality of HOI, we design an interactive textural enhancement module for both hands and objects by introducing two independent memory banks. We also propose a layout adjustment strategy for the cross-object reenactment scenario to adaptively adjust unreasonable layouts caused by diverse object sizes during inference. Comprehensive qualitative and quantitative evaluations demonstrate that our proposed framework significantly outperforms existing methods. Project page: https://fyycs.github.io/Re-HOLD.
comment: Accepted to CVPR 2025
♻ ☆ IncEventGS: Pose-Free Gaussian Splatting from a Single Event Camera
Implicit neural representation and explicit 3D Gaussian Splatting (3D-GS) for novel view synthesis have achieved remarkable progress with frame-based camera (e.g. RGB and RGB-D cameras) recently. Compared to frame-based camera, a novel type of bio-inspired visual sensor, i.e. event camera, has demonstrated advantages in high temporal resolution, high dynamic range, low power consumption and low latency. Due to its unique asynchronous and irregular data capturing process, limited work has been proposed to apply neural representation or 3D Gaussian splatting for an event camera. In this work, we present IncEventGS, an incremental 3D Gaussian Splatting reconstruction algorithm with a single event camera. To recover the 3D scene representation incrementally, we exploit the tracking and mapping paradigm of conventional SLAM pipelines for IncEventGS. Given the incoming event stream, the tracker firstly estimates an initial camera motion based on prior reconstructed 3D-GS scene representation. The mapper then jointly refines both the 3D scene representation and camera motion based on the previously estimated motion trajectory from the tracker. The experimental results demonstrate that IncEventGS delivers superior performance compared to prior NeRF-based methods and other related baselines, even we do not have the ground-truth camera poses. Furthermore, our method can also deliver better performance compared to state-of-the-art event visual odometry methods in terms of camera motion estimation. Code is publicly available at: https://github.com/wu-cvgl/IncEventGS.
comment: Code Page: https://github.com/wu-cvgl/IncEventGS
♻ ☆ Any6D: Model-free 6D Pose Estimation of Novel Objects CVPR 2025
We introduce Any6D, a model-free framework for 6D object pose estimation that requires only a single RGB-D anchor image to estimate both the 6D pose and size of unknown objects in novel scenes. Unlike existing methods that rely on textured 3D models or multiple viewpoints, Any6D leverages a joint object alignment process to enhance 2D-3D alignment and metric scale estimation for improved pose accuracy. Our approach integrates a render-and-compare strategy to generate and refine pose hypotheses, enabling robust performance in scenarios with occlusions, non-overlapping views, diverse lighting conditions, and large cross-environment variations. We evaluate our method on five challenging datasets: REAL275, Toyota-Light, HO3D, YCBINEOAT, and LM-O, demonstrating its effectiveness in significantly outperforming state-of-the-art methods for novel object pose estimation. Project page: https://taeyeop.com/any6d
comment: CVPR 2025, Project Page: https://taeyeop.com/any6d
♻ ☆ Feature Calibration enhanced Parameter Synthesis for CLIP-based Class-incremental Learning
Class-incremental Learning (CIL) enables models to continuously learn new class knowledge while memorizing previous classes, facilitating their adaptation and evolution in dynamic environments. Traditional CIL methods are mainly based on visual features, which limits their ability to handle complex scenarios. In contrast, Vision-Language Models (VLMs) show promising potential to promote CIL by integrating pretrained knowledge with textual features. However, previous methods make it difficult to overcome catastrophic forgetting while preserving the generalization capabilities of VLMs. To tackle these challenges, we propose Feature Calibration enhanced Parameter Synthesis (FCPS) in this paper. Specifically, our FCPS employs a specific parameter adjustment mechanism to iteratively refine the proportion of original visual features participating in the final class determination, ensuring the model's foundational generalization capabilities. Meanwhile, parameter integration across different tasks achieves a balance between learning new class knowledge and retaining old knowledge. Experimental results on popular benchmarks (e.g., CIFAR100 and ImageNet100) validate the superiority of the proposed method.
♻ ☆ Superpixel Tokenization for Vision Transformers: Preserving Semantic Integrity in Visual Tokens
Transformers, a groundbreaking architecture proposed for Natural Language Processing (NLP), have also achieved remarkable success in Computer Vision. A cornerstone of their success lies in the attention mechanism, which models relationships among tokens. While the tokenization process in NLP inherently ensures that a single token does not contain multiple semantics, the tokenization of Vision Transformer (ViT) utilizes tokens from uniformly partitioned square image patches, which may result in an arbitrary mixing of visual concepts in a token. In this work, we propose to substitute the grid-based tokenization in ViT with superpixel tokenization, which employs superpixels to generate a token that encapsulates a sole visual concept. Unfortunately, the diverse shapes, sizes, and locations of superpixels make integrating superpixels into ViT tokenization rather challenging. Our tokenization pipeline, comprised of pre-aggregate extraction and superpixel-aware aggregation, overcomes the challenges that arise in superpixel tokenization. Extensive experiments demonstrate that our approach, which exhibits strong compatibility with existing frameworks, enhances the accuracy and robustness of ViT on various downstream tasks.
comment: Project page: https://github.com/jangsoohyuk/SuiT
♻ ☆ CompMarkGS: Robust Watermarking for Compressed 3D Gaussian Splatting
3D Gaussian Splatting (3DGS) enables rapid differentiable rendering for 3D reconstruction and novel view synthesis, leading to its widespread commercial use. Consequently, copyright protection via watermarking has become critical. However, because 3DGS relies on millions of Gaussians, which require gigabytes of storage, efficient transfer and storage require compression. Existing 3DGS watermarking methods are vulnerable to quantization-based compression, often resulting in the loss of the embedded watermark. To address this challenge, we propose a novel watermarking method that ensures watermark robustness after model compression while maintaining high rendering quality. In detail, we incorporate a quantization distortion layer that simulates compression during training, preserving the watermark under quantization-based compression. Also, we propose a learnable watermark embedding feature that embeds the watermark into the anchor feature, ensuring structural consistency and seamless integration into the 3D scene. Furthermore, we present a frequency-aware anchor growing mechanism to enhance image quality in high-frequency regions by effectively identifying Guassians within these regions. Experimental results confirm that our method preserves the watermark and maintains superior image quality under high compression, validating it as a promising approach for a secure 3DGS model.
comment: 23 pages, 17 figures
♻ ☆ AMD-Hummingbird: Towards an Efficient Text-to-Video Model
Text-to-Video (T2V) generation has attracted significant attention for its ability to synthesize realistic videos from textual descriptions. However, existing models struggle to balance computational efficiency and high visual quality, particularly on resource-limited devices, e.g.,iGPUs and mobile phones. Most prior work prioritizes visual fidelity while overlooking the need for smaller, more efficient models suitable for real-world deployment. To address this challenge, we propose a lightweight T2V framework, termed Hummingbird, which prunes existing models and enhances visual quality through visual feedback learning. Our approach reduces the size of the U-Net from 1.4 billion to 0.7 billion parameters, significantly improving efficiency while preserving high-quality video generation. Additionally, we introduce a novel data processing pipeline that leverages Large Language Models (LLMs) and Video Quality Assessment (VQA) models to enhance the quality of both text prompts and video data. To support user-driven training and style customization, we publicly release the full training code, including data processing and model training. Extensive experiments show that our method achieves a 31X speedup compared to state-of-the-art models such as VideoCrafter2, while also attaining the highest overall score on VBench. Moreover, our method supports the generation of videos with up to 26 frames, addressing the limitations of existing U-Net-based methods in long video generation. Notably, the entire training process requires only four GPUs, yet delivers performance competitive with existing leading methods. Hummingbird presents a practical and efficient solution for T2V generation, combining high performance, scalability, and flexibility for real-world applications.
comment: Homepage: https://www.amd.com/en/developer/resources/technical-articles/amd-hummingbird-0-9b-text-to-video-diffusion-model-with-4-step-inferencing.html| GitHub: https://github.com/AMD-AIG-AIMA/AMD-Hummingbird-T2V
♻ ☆ AIM2PC: Aerial Image to 3D Building Point Cloud Reconstruction SP
Three-dimensional urban reconstruction of buildings from single-view images has attracted significant attention over the past two decades. However, recent methods primarily focus on rooftops from aerial images, often overlooking essential geometrical details. Additionally, there is a notable lack of datasets containing complete 3D point clouds for entire buildings, along with challenges in obtaining reliable camera pose information for aerial images. This paper addresses these challenges by presenting a novel methodology, AIM2PC , which utilizes our generated dataset that includes complete 3D point clouds and determined camera poses. Our approach takes features from a single aerial image as input and concatenates them with essential additional conditions, such as binary masks and Sobel edge maps, to enable more edge-aware reconstruction. By incorporating a point cloud diffusion model based on Centered denoising Diffusion Probabilistic Models (CDPM), we project these concatenated features onto the partially denoised point cloud using our camera poses at each diffusion step. The proposed method is able to reconstruct the complete 3D building point cloud, including wall information and demonstrates superior performance compared to existing baseline techniques. To allow further comparisons with our methodology the dataset has been made available at https://github.com/Soulaimene/AIM2PCDataset
comment: Accepted to ISPRS Geospatial Week 2025
♻ ☆ LookCloser: Frequency-aware Radiance Field for Tiny-Detail Scene CVPR 2025
Humans perceive and comprehend their surroundings through information spanning multiple frequencies. In immersive scenes, people naturally scan their environment to grasp its overall structure while examining fine details of objects that capture their attention. However, current NeRF frameworks primarily focus on modeling either high-frequency local views or the broad structure of scenes with low-frequency information, which is limited to balancing both. We introduce FA-NeRF, a novel frequency-aware framework for view synthesis that simultaneously captures the overall scene structure and high-definition details within a single NeRF model. To achieve this, we propose a 3D frequency quantification method that analyzes the scene's frequency distribution, enabling frequency-aware rendering. Our framework incorporates a frequency grid for fast convergence and querying, a frequency-aware feature re-weighting strategy to balance features across different frequency contents. Extensive experiments show that our method significantly outperforms existing approaches in modeling entire scenes while preserving fine details. Project page: https://coscatter.github.io/LookCloser/
comment: CVPR 2025. Project page: https://coscatter.github.io/LookCloser
♻ ☆ Global-Local Tree Search in VLMs for 3D Indoor Scene Generation CVPR 2025
Large Vision-Language Models (VLMs), such as GPT-4, have achieved remarkable success across various fields. However, there are few studies on 3D indoor scene generation with VLMs. This paper considers this task as a planning problem subject to spatial and layout common sense constraints. To solve the problem with a VLM, we propose a new global-local tree search algorithm. Globally, the method places each object sequentially and explores multiple placements during each placement process, where the problem space is represented as a tree. To reduce the depth of the tree, we decompose the scene structure hierarchically, i.e. room level, region level, floor object level, and supported object level. The algorithm independently generates the floor objects in different regions and supported objects placed on different floor objects. Locally, we also decompose the sub-task, the placement of each object, into multiple steps. The algorithm searches the tree of problem space. To leverage the VLM model to produce positions of objects, we discretize the top-down view space as a dense grid and fill each cell with diverse emojis to make to cells distinct. We prompt the VLM with the emoji grid and the VLM produces a reasonable location for the object by describing the position with the name of emojis. The quantitative and qualitative experimental results illustrate our approach generates more plausible 3D scenes than state-of-the-art approaches. Our source code is available at https://github.com/dw-dengwei/TreeSearchGen .
comment: Accepted by CVPR 2025
♻ ☆ StableGS: A Floater-Free Framework for 3D Gaussian Splatting
Recent years have witnessed remarkable success of 3D Gaussian Splatting (3DGS) in novel view synthesis, surpassing prior differentiable rendering methods in both quality and efficiency. However, its training process suffers from coupled opacity-color optimization that frequently converges to local minima, producing floater artifacts that degrade visual fidelity. We present StableGS, a framework that eliminates floaters through cross-view depth consistency constraints while introducing a dual-opacity GS model to decouple geometry and material properties of translucent objects. To further enhance reconstruction quality in weakly-textured regions, we integrate DUSt3R depth estimation, significantly improving geometric stability. Our method fundamentally addresses 3DGS training instabilities, outperforming existing state-of-the-art methods across open-source datasets.
♻ ☆ Latent Space Super-Resolution for Higher-Resolution Image Generation with Diffusion Models CVPR 2025
In this paper, we propose LSRNA, a novel framework for higher-resolution (exceeding 1K) image generation using diffusion models by leveraging super-resolution directly in the latent space. Existing diffusion models struggle with scaling beyond their training resolutions, often leading to structural distortions or content repetition. Reference-based methods address the issues by upsampling a low-resolution reference to guide higher-resolution generation. However, they face significant challenges: upsampling in latent space often causes manifold deviation, which degrades output quality. On the other hand, upsampling in RGB space tends to produce overly smoothed outputs. To overcome these limitations, LSRNA combines Latent space Super-Resolution (LSR) for manifold alignment and Region-wise Noise Addition (RNA) to enhance high-frequency details. Our extensive experiments demonstrate that integrating LSRNA outperforms state-of-the-art reference-based methods across various resolutions and metrics, while showing the critical role of latent space upsampling in preserving detail and sharpness. The code is available at https://github.com/3587jjh/LSRNA.
comment: Accepted by CVPR 2025
♻ ☆ CQ-DINO: Mitigating Gradient Dilution via Category Queries for Vast Vocabulary Object Detection
With the exponential growth of data, traditional object detection methods are increasingly struggling to handle vast vocabulary object detection tasks effectively. We analyze two key limitations of classification-based detectors: positive gradient dilution, where rare positive categories receive insufficient learning signals, and hard negative gradient dilution, where discriminative gradients are overwhelmed by numerous easy negatives. To address these challenges, we propose CQ-DINO, a category query-based object detection framework that reformulates classification as a contrastive task between object queries and learnable category queries. Our method introduces image-guided query selection, which reduces the negative space by adaptively retrieving top-K relevant categories per image via cross-attention, thereby rebalancing gradient distributions and facilitating implicit hard example mining. Furthermore, CQ-DINO flexibly integrates explicit hierarchical category relationships in structured datasets (e.g., V3Det) or learns implicit category correlations via self-attention in generic datasets (e.g., COCO). Experiments demonstrate that CQ-DINO achieves superior performance on the challenging V3Det benchmark (surpassing previous methods by 2.1% AP) while maintaining competitiveness in COCO. Our work provides a scalable solution for real-world detection systems requiring wide category coverage. The dataset and code will be publicly at https://github.com/RedAIGC/CQ-DINO.
♻ ☆ Using deep neural networks to detect non-analytically defined expert event labels in canoe sprint force sensor signals
Assessing an athlete's performance in canoe sprint is often established by measuring a variety of kinematic parameters during training sessions. Many of these parameters are related to single or multiple paddle stroke cycles. Determining on- and offset of these cycles in force sensor signals is usually not straightforward and requires human interaction. This paper explores convolutional neural networks (CNNs) and recurrent neural networks (RNNs) in terms of their ability to automatically predict these events. In addition, our work proposes an extension to the recently published SoftED metric for event detection in order to properly assess the model performance on time windows. In our results, an RNN based on bidirectional gated recurrent units (BGRUs) turned out to be the most suitable model for paddle stroke detection.
♻ ☆ VTD-CLIP: Video-to-Text Discretization via Prompting CLIP
Vision-language models bridge visual and linguistic understanding and have proven to be powerful for video recognition tasks. Existing approaches primarily rely on parameter-efficient fine-tuning of image-text pre-trained models, yet they often suffer from limited interpretability and poor generalization due to inadequate temporal modeling. To address these, we propose a simple yet effective video-to-text discretization framework. Our method repurposes the frozen text encoder to construct a visual codebook from video class labels due to the many-to-one contrastive alignment between visual and textual embeddings in multimodal pretraining. This codebook effectively transforms temporal visual data into textual tokens via feature lookups and offers interpretable video representations through explicit video modeling. Then, to enhance robustness against irrelevant or noisy frames, we introduce a confidence-aware fusion module that dynamically weights keyframes by assessing their semantic relevance via the codebook. Furthermore, our method incorporates learnable text prompts to conduct adaptive codebook updates. Extensive experiments on HMDB-51, UCF-101, SSv2, and Kinetics-400 have validated the superiority of our approach, achieving more competitive improvements over state-of-the-art methods. The code will be publicly available at https://github.com/isxinxin/VTD-CLIP.
♻ ☆ Instruct-CLIP: Improving Instruction-Guided Image Editing with Automated Data Refinement Using Contrastive Learning
Although natural language instructions offer an intuitive way to guide automated image editing, deep-learning models often struggle to achieve high-quality results, largely due to the difficulty of creating large, high-quality training datasets. To do this, previous approaches have typically relied on text-to-image (T2I) generative models to produce pairs of original and edited images that simulate the input/output of an instruction-guided image-editing model. However, these image pairs often fail to align with the specified edit instructions due to the limitations of T2I models, which negatively impacts models trained on such datasets. To address this, we present Instruct-CLIP (I-CLIP), a selfsupervised method that learns the semantic changes between original and edited images to refine and better align the instructions in existing datasets. Furthermore, we adapt Instruct-CLIP to handle noisy latent images and diffusion timesteps so that it can be used to train latent diffusion models (LDMs) and efficiently enforce alignment between the edit instruction and the image changes in latent space at any step of the diffusion pipeline. We use Instruct-CLIP to correct the InstructPix2Pix dataset and get over 120K refined samples we then use to fine-tune their model, guided by our novel I-CLIP-based loss function. The resulting model can produce edits that are more aligned with the given instructions. Our code and dataset are available at https://github.com/SherryXTChen/Instruct-CLIP.git.
comment: Computer Vision and Pattern Recognition 2025
♻ ☆ Lessons and Insights from a Unifying Study of Parameter-Efficient Fine-Tuning (PEFT) in Visual Recognition CVPR 2025
Parameter-efficient fine-tuning (PEFT) has attracted significant attention due to the growth of pre-trained model sizes and the need to fine-tune (FT) them for superior downstream performance. Despite a surge in new PEFT methods, a systematic study to understand their performance and suitable application scenarios is lacking, leaving questions like "when to apply PEFT" and "which method to use" largely unanswered, especially in visual recognition. In this paper, we conduct a unifying empirical study of representative PEFT methods with Vision Transformers. We systematically tune their hyperparameters to fairly compare their accuracy on downstream tasks. Our study offers a practical user guide and unveils several new insights. First, if tuned carefully, different PEFT methods achieve similar accuracy in the low-shot benchmark VTAB-1K. This includes simple approaches like FT the bias terms that were reported inferior. Second, despite similar accuracy, we find that PEFT methods make different mistakes and high-confidence predictions, likely due to their different inductive biases. Such an inconsistency (or complementarity) opens up the opportunity for ensemble methods, and we make preliminary attempts at this. Third, going beyond the commonly used low-shot tasks, we find that PEFT is also useful in many-shot regimes, achieving comparable or better accuracy than full FT while using significantly fewer parameters. Lastly, we investigate PEFT's ability to preserve a pre-trained model's robustness to distribution shifts (e.g., CLIP). Perhaps not surprisingly, PEFT approaches outperform full FT alone. However, with weight-space ensembles, full FT can better balance target distribution and distribution shift performance, suggesting a future research direction for robust PEFT.
comment: CVPR 2025. The code is available at https://github.com/OSU-MLB/ViT_PEFT_Vision
♻ ☆ PG-SAM: Prior-Guided SAM with Medical for Multi-organ Segmentation
Segment Anything Model (SAM) demonstrates powerful zero-shot capabilities; however, its accuracy and robustness significantly decrease when applied to medical image segmentation. Existing methods address this issue through modality fusion, integrating textual and image information to provide more detailed priors. In this study, we argue that the granularity of text and the domain gap affect the accuracy of the priors. Furthermore, the discrepancy between high-level abstract semantics and pixel-level boundary details in images can introduce noise into the fusion process. To address this, we propose Prior-Guided SAM (PG-SAM), which employs a fine-grained modality prior aligner to leverage specialized medical knowledge for better modality alignment. The core of our method lies in efficiently addressing the domain gap with fine-grained text from a medical LLM. Meanwhile, it also enhances the priors' quality after modality alignment, ensuring more accurate segmentation. In addition, our decoder enhances the model's expressive capabilities through multi-level feature fusion and iterative mask optimizer operations, supporting unprompted learning. We also propose a unified pipeline that effectively supplies high-quality semantic information to SAM. Extensive experiments on the Synapse dataset demonstrate that the proposed PG-SAM achieves state-of-the-art performance. Our anonymous code is released at https://github.com/logan-0623/PG-SAM.
♻ ☆ Decorum: A Language-Based Approach For Style-Conditioned Synthesis of Indoor 3D Scenes
3D indoor scene generation is an important problem for the design of digital and real-world environments. To automate this process, a scene generation model should be able to not only generate plausible scene layouts, but also take into consideration visual features and style preferences. Existing methods for this task exhibit very limited control over these attributes, only allowing text inputs in the form of simple object-level descriptions or pairwise spatial relationships. Our proposed method Decorum enables users to control the scene generation process with natural language by adopting language-based representations at each stage. This enables us to harness recent advancements in Large Language Models (LLMs) to model language-to-language mappings. In addition, we show that using a text-based representation allows us to select furniture for our scenes using a novel object retrieval method based on multimodal LLMs. Evaluations on the benchmark 3D-FRONT dataset show that our methods achieve improvements over existing work in text-conditioned scene synthesis and object retrieval.
♻ ☆ Medical Report Generation Is A Multi-label Classification Problem
Medical report generation is a critical task in healthcare that involves the automatic creation of detailed and accurate descriptions from medical images. Traditionally, this task has been approached as a sequence generation problem, relying on vision-and-language techniques to generate coherent and contextually relevant reports. However, in this paper, we propose a novel perspective: rethinking medical report generation as a multi-label classification problem. By framing the task this way, we leverage the radiology nodes from the commonly used knowledge graph, which can be better captured through classification techniques. To verify our argument, we introduce a novel report generation framework based on BLIP integrated with classified key nodes, which allows for effective report generation with accurate classification of multiple key aspects within the medical images. This approach not only simplifies the report generation process but also significantly enhances performance metrics. Our extensive experiments demonstrate that leveraging key nodes can achieve state-of-the-art (SOTA) performance, surpassing existing approaches across two benchmark datasets. The results underscore the potential of re-envisioning traditional tasks with innovative methodologies, paving the way for more efficient and accurate medical report generation.
comment: Accepted to 2024 IEEE International Conference on Medical Artificial Intelligence
♻ ☆ Mitigating Data Redundancy to Revitalize Transformer-based Long-Term Time Series Forecasting System
Long-term time-series forecasting (LTSF) is fundamental to various real-world applications, where Transformer-based models have become the dominant framework due to their ability to capture long-range dependencies. However, these models often experience overfitting due to data redundancy in rolling forecasting settings, limiting their generalization ability particularly evident in longer sequences with highly similar adjacent data. In this work, we introduce CLMFormer, a novel framework that mitigates redundancy through curriculum learning and a memory-driven decoder. Specifically, we progressively introduce Bernoulli noise to the training samples, which effectively breaks the high similarity between adjacent data points. This curriculum-driven noise introduction aids the memory-driven decoder by supplying more diverse and representative training data, enhancing the decoder's ability to model seasonal tendencies and dependencies in the time-series data. To further enhance forecasting accuracy, we introduce a memory-driven decoder. This component enables the model to capture seasonal tendencies and dependencies in the time-series data and leverages temporal relationships to facilitate the forecasting process. Extensive experiments on six real-world LTSF benchmarks show that CLMFormer consistently improves Transformer-based models by up to 30%, demonstrating its effectiveness in long-horizon forecasting.
comment: ACM TIST
♻ ☆ Identity-Preserving Text-to-Video Generation by Frequency Decomposition CVPR 2025
Identity-preserving text-to-video (IPT2V) generation aims to create high-fidelity videos with consistent human identity. It is an important task in video generation but remains an open problem for generative models. This paper pushes the technical frontier of IPT2V in two directions that have not been resolved in literature: (1) A tuning-free pipeline without tedious case-by-case finetuning, and (2) A frequency-aware heuristic identity-preserving DiT-based control scheme. We propose ConsisID, a tuning-free DiT-based controllable IPT2V model to keep human identity consistent in the generated video. Inspired by prior findings in frequency analysis of diffusion transformers, it employs identity-control signals in the frequency domain, where facial features can be decomposed into low-frequency global features and high-frequency intrinsic features. First, from a low-frequency perspective, we introduce a global facial extractor, which encodes reference images and facial key points into a latent space, generating features enriched with low-frequency information. These features are then integrated into shallow layers of the network to alleviate training challenges associated with DiT. Second, from a high-frequency perspective, we design a local facial extractor to capture high-frequency details and inject them into transformer blocks, enhancing the model's ability to preserve fine-grained features. We propose a hierarchical training strategy to leverage frequency information for identity preservation, transforming a vanilla pre-trained video generation model into an IPT2V model. Extensive experiments demonstrate that our frequency-aware heuristic scheme provides an optimal control solution for DiT-based models. Thanks to this scheme, our ConsisID generates high-quality, identity-preserving videos, making strides towards more effective IPT2V. Code: https://github.com/PKU-YuanGroup/ConsisID.
comment: CVPR 2025
♻ ☆ Simulator HC: Regression-based Online Simulation of Starting Problem-Solution Pairs for Homotopy Continuation in Geometric Vision
While automatically generated polynomial elimination templates have sparked great progress in the field of 3D computer vision, there remain many problems for which the degree of the constraints or the number of unknowns leads to intractability. In recent years, homotopy continuation has been introduced as a plausible alternative. However, the method currently depends on expensive parallel tracking of all possible solutions in the complex domain, or a classification network for starting problem-solution pairs trained over a limited set of real-world examples. Our innovation lies in a novel approach to finding solution-problem pairs, where we only need to predict a rough initial solution, with the corresponding problem generated by an online simulator. Subsequently, homotopy continuation is applied to track that single solution back to the original problem. We apply this elegant combination to generalized camera resectioning, and also introduce a new solution to the challenging generalized relative pose and scale problem. As demonstrated, the proposed method successfully compensates the raw error committed by the regressor alone, and leads to state-of-the-art efficiency and success rates.
♻ ☆ Bézier Splatting for Fast and Differentiable Vector Graphics
Differentiable vector graphics (VGs) are widely used in image vectorization and vector synthesis, while existing representations are costly to optimize and struggle to achieve high-quality rendering results for high-resolution images. This work introduces a new differentiable VG representation, dubbed B\'ezier splatting, that enables fast yet high-fidelity VG rasterization. B\'ezier splatting samples 2D Gaussians along B\'ezier curves, which naturally provide positional gradients at object boundaries. Thanks to the efficient splatting-based differentiable rasterizer, B\'ezier splatting achieves over 20x and 150x faster per forward and backward rasterization step for open curves compared to DiffVG. Additionally, we introduce an adaptive pruning and densification strategy that dynamically adjusts the spatial distribution of curves to escape local minima, further improving VG quality. Experimental results show that B\'ezier splatting significantly outperforms existing methods with better visual fidelity and 10x faster optimization speed.
comment: Project page: https://xiliu8006.github.io/Bezier_splatting_project/
♻ ☆ VisionArena: 230K Real World User-VLM Conversations with Preference Labels CVPR
With the growing adoption and capabilities of vision-language models (VLMs) comes the need for benchmarks that capture authentic user-VLM interactions. In response, we create VisionArena, a dataset of 230K real-world conversations between users and VLMs. Collected from Chatbot Arena - an open-source platform where users interact with VLMs and submit preference votes - VisionArena spans 73K unique users, 45 VLMs, and 138 languages. Our dataset contains three subsets: VisionArena-Chat, 200k single and multi-turn conversations between a user and a VLM; VisionArena-Battle, 30K conversations comparing two anonymous VLMs with user preference votes; and VisionArena-Bench, an automatic benchmark of 500 diverse user prompts that efficiently approximate the live Chatbot Arena model rankings. Additionally, we highlight the types of question asked by users, the influence of response style on preference, and areas where models often fail. We find open-ended tasks like captioning and humor are highly style-dependent, and current VLMs struggle with spatial reasoning and planning tasks. Lastly, we show finetuning the same base model on VisionArena-Chat outperforms Llava-Instruct-158K, with a 17-point gain on MMMU and a 46-point gain on the WildVision benchmark. Dataset at https://huggingface.co/lmarena-ai
comment: updated for CVPR Camera Ready
♻ ☆ Similarity-Dissimilarity Loss for Multi-label Supervised Contrastive Learning
Supervised contrastive learning has achieved remarkable success by leveraging label information; however, determining positive samples in multi-label scenarios remains a critical challenge. In multi-label supervised contrastive learning (MSCL), relations among multi-label samples are not yet fully defined, leading to ambiguity in identifying positive samples and formulating contrastive loss functions to construct the representation space. To address these challenges, we: (i) first define five distinct multi-label relations in MSCL to systematically identify positive samples, (ii) introduce a novel Similarity-Dissimilarity Loss that dynamically re-weights samples through computing the similarity and dissimilarity factors between positive samples and given anchors based on multi-label relations, and (iii) further provide theoretical grounded proof for our method through rigorous mathematical analysis that supports the formulation and effectiveness of the proposed loss function. We conduct the experiments across both image and text modalities, and extend the evaluation to medical domain. The results demonstrate that our method consistently outperforms baselines in a comprehensive evaluation, confirming its effectiveness and robustness. Code is available at: https://github.com/guangminghuang/similarity-dissimilarity-loss.
♻ ☆ Simulation of prosthetic vision with PRIMA system and enhancement of face representation
Objective. Patients implanted with the PRIMA photovoltaic subretinal prosthesis in geographic atrophy report form vision with the average acuity matching the 100um pixel size. Although this remarkable outcome enables them to read and write, they report difficulty with perceiving faces. This paper provides a novel, non-pixelated algorithm for simulating prosthetic vision the way it is experienced by PRIMA patients, compares the algorithm's predictions to clinical perceptual outcomes, and offers computer vision and machine learning (ML) methods to improve face representation. Approach. Our simulation algorithm integrates a grayscale filter, spatial resolution filter, and contrast filter. This accounts for the limited sampling density of the retinal implant, as well as the reduced contrast sensitivity of prosthetic vision. Patterns of Landolt C and faces created using this simulation algorithm are compared to reports from actual PRIMA users. To recover the facial features lost in prosthetic vision, we apply an ML facial landmarking model as well as contrast adjusting tone curves to the face image prior to its projection onto the implant. Main results. Simulated prosthetic vision matches the maximum letter acuity observed in clinical studies as well as patients' subjective descriptions. Application of the inversed contrast filter helps preserve the contrast in prosthetic vision. Identification of the facial features using an ML facial landmarking model and accentuating them further improve face representation. Significance. Spatial and contrast constraints of prosthetic vision limit resolvable features and degrade natural images. ML based methods and contrast adjustments mitigate some limitations and improve face representation. Even though higher spatial resolution can be expected with implants having smaller pixels, contrast enhancement still remains essential for face recognition.
♻ ☆ Fidelity-Imposed Displacement Editing for the Learn2Reg 2024 SHG-BF Challenge
Co-examination of second-harmonic generation (SHG) and bright-field (BF) microscopy enables the differentiation of tissue components and collagen fibers, aiding the analysis of human breast and pancreatic cancer tissues. However, large discrepancies between SHG and BF images pose challenges for current learning-based registration models in aligning SHG to BF. In this paper, we propose a novel multi-modal registration framework that employs fidelity-imposed displacement editing to address these challenges. The framework integrates batch-wise contrastive learning, feature-based pre-alignment, and instance-level optimization. Experimental results from the Learn2Reg COMULISglobe SHG-BF Challenge validate the effectiveness of our method, securing the 1st place on the online leaderboard.
comment: Due to the instruction and conflict of co-author
♻ ☆ SimMotionEdit: Text-Based Human Motion Editing with Motion Similarity Prediction
Text-based 3D human motion editing is a critical yet challenging task in computer vision and graphics. While training-free approaches have been explored, the recent release of the MotionFix dataset, which includes source-text-motion triplets, has opened new avenues for training, yielding promising results. However, existing methods struggle with precise control, often leading to misalignment between motion semantics and language instructions. In this paper, we introduce a related task, motion similarity prediction, and propose a multi-task training paradigm, where we train the model jointly on motion editing and motion similarity prediction to foster the learning of semantically meaningful representations. To complement this task, we design an advanced Diffusion-Transformer-based architecture that separately handles motion similarity prediction and motion editing. Extensive experiments demonstrate the state-of-the-art performance of our approach in both editing alignment and fidelity.
comment: Project URL: https://github.com/lzhyu/SimMotionEdit
♻ ☆ Motion-Boundary-Driven Unsupervised Surgical Instrument Segmentation in Low-Quality Optical Flow
Unsupervised video-based surgical instrument segmentation has the potential to accelerate the adoption of robot-assisted procedures by reducing the reliance on manual annotations. However, the generally low quality of optical flow in endoscopic footage poses a great challenge for unsupervised methods that rely heavily on motion cues. To overcome this limitation, we propose a novel approach that pinpoints motion boundaries, regions with abrupt flow changes, while selectively discarding frames with globally low-quality flow and adapting to varying motion patterns. Experiments on the EndoVis2017 VOS and EndoVis2017 Challenge datasets show that our method achieves mean Intersection-over-Union (mIoU) scores of 0.75 and 0.72, respectively, effectively alleviating the constraints imposed by suboptimal optical flow. This enables a more scalable and robust surgical instrument segmentation solution in clinical settings. The code will be publicly released.
♻ ☆ VisFocus: Prompt-Guided Vision Encoders for OCR-Free Dense Document Understanding ECCV 2024
In recent years, notable advancements have been made in the domain of visual document understanding, with the prevailing architecture comprising a cascade of vision and language models. The text component can either be extracted explicitly with the use of external OCR models in OCR-based approaches, or alternatively, the vision model can be endowed with reading capabilities in OCR-free approaches. Typically, the queries to the model are input exclusively to the language component, necessitating the visual features to encompass the entire document. In this paper, we present VisFocus, an OCR-free method designed to better exploit the vision encoder's capacity by coupling it directly with the language prompt. To do so, we replace the down-sampling layers with layers that receive the input prompt and allow highlighting relevant parts of the document, while disregarding others. We pair the architecture enhancements with a novel pre-training task, using language masking on a snippet of the document text fed to the visual encoder in place of the prompt, to empower the model with focusing capabilities. Consequently, VisFocus learns to allocate its attention to text patches pertinent to the provided prompt. Our experiments demonstrate that this prompt-guided visual encoding approach significantly improves performance, achieving state-of-the-art results on various benchmarks.
comment: ECCV 2024, official code at https://github.com/amazon-science/visfocus
♻ ☆ LLAVIDAL: A Large LAnguage VIsion Model for Daily Activities of Living CVPR 2025
Current Large Language Vision Models (LLVMs) trained on web videos perform well in general video understanding but struggle with fine-grained details, complex human-object interactions (HOI), and view-invariant representation learning essential for Activities of Daily Living (ADL). This limitation stems from a lack of specialized ADL video instruction-tuning datasets and insufficient modality integration to capture discriminative action representations. To address this, we propose a semi-automated framework for curating ADL datasets, creating ADL-X, a multiview, multimodal RGBS instruction-tuning dataset. Additionally, we introduce LLAVIDAL, an LLVM integrating videos, 3D skeletons, and HOIs to model ADL's complex spatiotemporal relationships. For training LLAVIDAL a simple joint alignment of all modalities yields suboptimal results; thus, we propose a Multimodal Progressive (MMPro) training strategy, incorporating modalities in stages following a curriculum. We also establish ADL MCQ and video description benchmarks to assess LLVM performance in ADL tasks. Trained on ADL-X, LLAVIDAL achieves state-of-the-art performance across ADL benchmarks. Code and data will be made publicly available at: https://adl-x.github.io/.
comment: CVPR 2025 Camera Ready
♻ ☆ POp-GS: Next Best View in 3D-Gaussian Splatting with P-Optimality
In this paper, we present a novel algorithm for quantifying uncertainty and information gained within 3D Gaussian Splatting (3D-GS) through P-Optimality. While 3D-GS has proven to be a useful world model with high-quality rasterizations, it does not natively quantify uncertainty or information, posing a challenge for real-world applications such as 3D-GS SLAM. We propose to quantify information gain in 3D-GS by reformulating the problem through the lens of optimal experimental design, which is a classical solution widely used in literature. By restructuring information quantification of 3D-GS through optimal experimental design, we arrive at multiple solutions, of which T-Optimality and D-Optimality perform the best quantitatively and qualitatively as measured on two popular datasets. Additionally, we propose a block diagonal covariance approximation which provides a measure of correlation at the expense of a greater computation cost.
♻ ☆ From My View to Yours: Ego-Augmented Learning in Large Vision Language Models for Understanding Exocentric Daily Living Activities
Large Vision Language Models (LVLMs) have demonstrated impressive capabilities in video understanding, yet their adoption for Activities of Daily Living (ADL) remains limited by their inability to capture fine-grained interactions and spatial relationships. To address this, we aim to leverage the complementary nature of egocentric views to enhance LVLM's understanding of exocentric ADL videos. Consequently, we propose ego2exo knowledge distillation to learn ego-augmented exp representations. While effective, this approach requires paired ego-exo videos, which are impractical to collect at scale. To address this, we propose Skeleton-guided Synthetic Ego Generation (SK-EGO), which leverages human skeleton motion to generate synthetic ego views from exocentric videos. To enhance the ego representation of LVLMs trained on synthetic data, we develop a domain-agnostic bootstrapped ego2exo strategy that effectively transfers knowledge from real ego-exo pairs to synthetic ego-exo pairs, while mitigating domain misalignment. We find that the exo representations of our ego-augmented LVLMs successfully learn to extract ego-perspective cues, demonstrated through comprehensive evaluation on six ADL benchmarks and our proposed Ego-in-Exo PerceptionMCQ benchmark designed specifically to assess egocentric understanding from exocentric videos. Code, models, and data will be open-sourced at https://github.com/dominickrei/EgoExo4ADL.
♻ ☆ Reanimating Images using Neural Representations of Dynamic Stimuli
While computer vision models have made incredible strides in static image recognition, they still do not match human performance in tasks that require the understanding of complex, dynamic motion. This is notably true for real-world scenarios where embodied agents face complex and motion-rich environments. Our approach, BrainNRDS (Brain-Neural Representations of Dynamic Stimuli), leverages state-of-the-art video diffusion models to decouple static image representation from motion generation, enabling us to utilize fMRI brain activity for a deeper understanding of human responses to dynamic visual stimuli. Conversely, we also demonstrate that information about the brain's representation of motion can enhance the prediction of optical flow in artificial systems. Our novel approach leads to four main findings: (1) Visual motion, represented as fine-grained, object-level resolution optical flow, can be decoded from brain activity generated by participants viewing video stimuli; (2) Video encoders outperform image-based models in predicting video-driven brain activity; (3) Brain-decoded motion signals enable realistic video reanimation based only on the initial frame of the video; and (4) We extend prior work to achieve full video decoding from video-driven brain activity. BrainNRDS advances our understanding of how the brain represents spatial and temporal information in dynamic visual scenes. Our findings demonstrate the potential of combining brain imaging with video diffusion models for developing more robust and biologically-inspired computer vision systems. We show additional decoding and encoding examples on this site: https://brain-nrds.github.io/.
comment: Project Page: https://brain-nrds.github.io
♻ ☆ Repurposing Pre-trained Video Diffusion Models for Event-based Video Interpolation CVPR 2025
Video Frame Interpolation aims to recover realistic missing frames between observed frames, generating a high-frame-rate video from a low-frame-rate video. However, without additional guidance, the large motion between frames makes this problem ill-posed. Event-based Video Frame Interpolation (EVFI) addresses this challenge by using sparse, high-temporal-resolution event measurements as motion guidance. This guidance allows EVFI methods to significantly outperform frame-only methods. However, to date, EVFI methods have relied on a limited set of paired event-frame training data, severely limiting their performance and generalization capabilities. In this work, we overcome the limited data challenge by adapting pre-trained video diffusion models trained on internet-scale datasets to EVFI. We experimentally validate our approach on real-world EVFI datasets, including a new one that we introduce. Our method outperforms existing methods and generalizes across cameras far better than existing approaches.
comment: Accepted to CVPR 2025
♻ ☆ MambaVision: A Hybrid Mamba-Transformer Vision Backbone CVPR'25
We propose a novel hybrid Mamba-Transformer backbone, MambaVision, specifically tailored for vision applications. Our core contribution includes redesigning the Mamba formulation to enhance its capability for efficient modeling of visual features. Through a comprehensive ablation study, we demonstrate the feasibility of integrating Vision Transformers (ViT) with Mamba. Our results show that equipping the Mamba architecture with self-attention blocks in the final layers greatly improves its capacity to capture long-range spatial dependencies. Based on these findings, we introduce a family of MambaVision models with a hierarchical architecture to meet various design criteria. For classification on the ImageNet-1K dataset, MambaVision variants achieve state-of-the-art (SOTA) performance in terms of both Top-1 accuracy and throughput. In downstream tasks such as object detection, instance segmentation, and semantic segmentation on MS COCO and ADE20K datasets, MambaVision outperforms comparably sized backbones while demonstrating favorable performance. Code: https://github.com/NVlabs/MambaVision
comment: Accepted to CVPR'25
♻ ☆ SAMWISE: Infusing Wisdom in SAM2 for Text-Driven Video Segmentation CVPR 2025
Referring Video Object Segmentation (RVOS) relies on natural language expressions to segment an object in a video clip. Existing methods restrict reasoning either to independent short clips, losing global context, or process the entire video offline, impairing their application in a streaming fashion. In this work, we aim to surpass these limitations and design an RVOS method capable of effectively operating in streaming-like scenarios while retaining contextual information from past frames. We build upon the Segment-Anything 2 (SAM2) model, that provides robust segmentation and tracking capabilities and is naturally suited for streaming processing. We make SAM2 wiser, by empowering it with natural language understanding and explicit temporal modeling at the feature extraction stage, without fine-tuning its weights, and without outsourcing modality interaction to external models. To this end, we introduce a novel adapter module that injects temporal information and multi-modal cues in the feature extraction process. We further reveal the phenomenon of tracking bias in SAM2 and propose a learnable module to adjust its tracking focus when the current frame features suggest a new object more aligned with the caption. Our proposed method, SAMWISE, achieves state-of-the-art across various benchmarks, by adding a negligible overhead of less than 5 M parameters. Code is available at https://github.com/ClaudiaCuttano/SAMWISE .
comment: CVPR 2025. Project page: https://claudiacuttano.github.io/SAMWISE
♻ ☆ Aesthetic Post-Training Diffusion Models from Generic Preferences with Step-by-step Preference Optimization CVPR 2025
Generating visually appealing images is fundamental to modern text-to-image generation models. A potential solution to better aesthetics is direct preference optimization (DPO), which has been applied to diffusion models to improve general image quality including prompt alignment and aesthetics. Popular DPO methods propagate preference labels from clean image pairs to all the intermediate steps along the two generation trajectories. However, preference labels provided in existing datasets are blended with layout and aesthetic opinions, which would disagree with aesthetic preference. Even if aesthetic labels were provided (at substantial cost), it would be hard for the two-trajectory methods to capture nuanced visual differences at different steps. To improve aesthetics economically, this paper uses existing generic preference data and introduces step-by-step preference optimization (SPO) that discards the propagation strategy and allows fine-grained image details to be assessed. Specifically, at each denoising step, we 1) sample a pool of candidates by denoising from a shared noise latent, 2) use a step-aware preference model to find a suitable win-lose pair to supervise the diffusion model, and 3) randomly select one from the pool to initialize the next denoising step. This strategy ensures that diffusion models focus on the subtle, fine-grained visual differences instead of layout aspect. We find that aesthetics can be significantly enhanced by accumulating these improved minor differences. When fine-tuning Stable Diffusion v1.5 and SDXL, SPO yields significant improvements in aesthetics compared with existing DPO methods while not sacrificing image-text alignment compared with vanilla models. Moreover, SPO converges much faster than DPO methods due to the use of more correct preference labels provided by the step-aware preference model.
comment: CVPR 2025. Project Page: https://rockeycoss.github.io/spo.github.io/
♻ ☆ CLIP-EBC: CLIP Can Count Accurately through Enhanced Blockwise Classification ICME 2025
We propose CLIP-EBC, the first fully CLIP-based model for accurate crowd density estimation. While the CLIP model has demonstrated remarkable success in addressing recognition tasks such as zero-shot image classification, its potential for counting has been largely unexplored due to the inherent challenges in transforming a regression problem, such as counting, into a recognition task. In this work, we investigate and enhance CLIP's ability to count, focusing specifically on the task of estimating crowd sizes from images. Existing classification-based crowd-counting frameworks have significant limitations, including the quantization of count values into bordering real-valued bins and the sole focus on classification errors. These practices result in label ambiguity near the shared borders and inaccurate prediction of count values. Hence, directly applying CLIP within these frameworks may yield suboptimal performance. To address these challenges, we first propose the Enhanced Blockwise Classification (EBC) framework. Unlike previous methods, EBC utilizes integer-valued bins, effectively reducing ambiguity near bin boundaries. Additionally, it incorporates a regression loss based on density maps to improve the prediction of count values. Within our backbone-agnostic EBC framework, we then introduce CLIP-EBC to fully leverage CLIP's recognition capabilities for this task. Extensive experiments demonstrate the effectiveness of EBC and the competitive performance of CLIP-EBC. Specifically, our EBC framework can improve existing classification-based methods by up to 44.5% on the UCF-QNRF dataset, and CLIP-EBC achieves state-of-the-art performance on the NWPU-Crowd test set, with an MAE of 58.2 and an RMSE of 268.5, representing improvements of 8.6% and 13.3% over the previous best method, STEERER. The code and weights are available at https://github.com/Yiming-M/CLIP-EBC.
comment: This is the author's accepted manuscript. The final version is published in ICME 2025
♻ ☆ UrbanCAD: Towards Highly Controllable and Photorealistic 3D Vehicles for Urban Scene Simulation
Photorealistic 3D vehicle models with high controllability are essential for autonomous driving simulation and data augmentation. While handcrafted CAD models provide flexible controllability, free CAD libraries often lack the high-quality materials necessary for photorealistic rendering. Conversely, reconstructed 3D models offer high-fidelity rendering but lack controllability. In this work, we introduce UrbanCAD, a framework that generates highly controllable and photorealistic 3D vehicle digital twins from a single urban image, leveraging a large collection of free 3D CAD models and handcrafted materials. To achieve this, we propose a novel pipeline that follows a retrieval-optimization manner, adapting to observational data while preserving fine-grained expert-designed priors for both geometry and material. This enables vehicles' realistic 360-degree rendering, background insertion, material transfer, relighting, and component manipulation. Furthermore, given multi-view background perspective and fisheye images, we approximate environment lighting using fisheye images and reconstruct the background with 3DGS, enabling the photorealistic insertion of optimized CAD models into rendered novel view backgrounds. Experimental results demonstrate that UrbanCAD outperforms baselines in terms of photorealism. Additionally, we show that various perception models maintain their accuracy when evaluated on UrbanCAD with in-distribution configurations but degrade when applied to realistic out-of-distribution data generated by our method. This suggests that UrbanCAD is a significant advancement in creating photorealistic, safety-critical driving scenarios for downstream applications.
comment: Project page: https://xdimlab.github.io/UrbanCAD/
♻ ☆ HyperFLINT: Hypernetwork-based Flow Estimation and Temporal Interpolation for Scientific Ensemble Visualization
We present HyperFLINT (Hypernetwork-based FLow estimation and temporal INTerpolation), a novel deep learning-based approach for estimating flow fields, temporally interpolating scalar fields, and facilitating parameter space exploration in spatio-temporal scientific ensemble data. This work addresses the critical need to explicitly incorporate ensemble parameters into the learning process, as traditional methods often neglect these, limiting their ability to adapt to diverse simulation settings and provide meaningful insights into the data dynamics. HyperFLINT introduces a hypernetwork to account for simulation parameters, enabling it to generate accurate interpolations and flow fields for each timestep by dynamically adapting to varying conditions, thereby outperforming existing parameter-agnostic approaches. The architecture features modular neural blocks with convolutional and deconvolutional layers, supported by a hypernetwork that generates weights for the main network, allowing the model to better capture intricate simulation dynamics. A series of experiments demonstrates HyperFLINT's significantly improved performance in flow field estimation and temporal interpolation, as well as its potential in enabling parameter space exploration, offering valuable insights into complex scientific ensembles.
♻ ☆ AuraFusion360: Augmented Unseen Region Alignment for Reference-based 360° Unbounded Scene Inpainting CVPR 2025
Three-dimensional scene inpainting is crucial for applications from virtual reality to architectural visualization, yet existing methods struggle with view consistency and geometric accuracy in 360{\deg} unbounded scenes. We present AuraFusion360, a novel reference-based method that enables high-quality object removal and hole filling in 3D scenes represented by Gaussian Splatting. Our approach introduces (1) depth-aware unseen mask generation for accurate occlusion identification, (2) Adaptive Guided Depth Diffusion, a zero-shot method for accurate initial point placement without requiring additional training, and (3) SDEdit-based detail enhancement for multi-view coherence. We also introduce 360-USID, the first comprehensive dataset for 360{\deg} unbounded scene inpainting with ground truth. Extensive experiments demonstrate that AuraFusion360 significantly outperforms existing methods, achieving superior perceptual quality while maintaining geometric accuracy across dramatic viewpoint changes.
comment: Paper accepted to CVPR 2025. Project page: https://kkennethwu.github.io/aurafusion360/
♻ ☆ FREE-Merging: Fourier Transform for Efficient Model Merging
With the rapid growth of deep learning, there is an increasing availability of open-source models for various tasks. However, single fine-tuned models often fall short of meeting the diverse needs of users. Model merging has thus emerged as an efficient method to integrate the capabilities of existing models into a unified model. Nevertheless, existing model merging methods face challenging trade-offs between performance and deployment costs, primarily due to task interference. For the first time, we reveal that task interference is evident in the frequency domain of model parameters, yet current efforts only focus on spatial domain solutions, which are largely ineffective in addressing frequency domain interference. To mitigate the impact of frequency domain interference, we propose FR-Merging, an innovative method that effectively filters harmful frequency domain interference on the backbone with minimal computational overhead. Since performance loss is inevitable with cost-free methods, we propose a lightweight task-specific expert module that dynamically compensates for information loss during merging. This proposed framework, FREE-Merging (FR-Merging with experts), strikes a balanced trade-off between training cost, inference latency, storage requirements, and performance. We demonstrate the effectiveness of both FR-Merging and FREE-Merging on multiple tasks across CV, NLP, and Multi-Modal domains and show that they can be flexibly adapted to specific needs.
comment: 20 pages, 10 figures
♻ ☆ GCC: Generative Color Constancy via Diffusing a Color Checker CVPR 2025
Color constancy methods often struggle to generalize across different camera sensors due to varying spectral sensitivities. We present GCC, which leverages diffusion models to inpaint color checkers into images for illumination estimation. Our key innovations include (1) a single-step deterministic inference approach that inpaints color checkers reflecting scene illumination, (2) a Laplacian decomposition technique that preserves checker structure while allowing illumination-dependent color adaptation, and (3) a mask-based data augmentation strategy for handling imprecise color checker annotations. By harnessing rich priors from pre-trained diffusion models, GCC demonstrates strong robustness in challenging cross-camera scenarios. These results highlight our method's effective generalization capability across different camera characteristics without requiring sensor-specific training, making it a versatile and practical solution for real-world applications.
comment: Paper accepted to CVPR 2025. Project page: https://chenwei891213.github.io/GCC/
♻ ☆ SpectroMotion: Dynamic 3D Reconstruction of Specular Scenes CVPR 2025
We present SpectroMotion, a novel approach that combines 3D Gaussian Splatting (3DGS) with physically-based rendering (PBR) and deformation fields to reconstruct dynamic specular scenes. Previous methods extending 3DGS to model dynamic scenes have struggled to represent specular surfaces accurately. Our method addresses this limitation by introducing a residual correction technique for accurate surface normal computation during deformation, complemented by a deformable environment map that adapts to time-varying lighting conditions. We implement a coarse-to-fine training strategy significantly enhancing scene geometry and specular color prediction. It is the only existing 3DGS method capable of synthesizing photorealistic real-world dynamic specular scenes, outperforming state-of-the-art methods in rendering complex, dynamic, and specular scenes.
comment: Paper accepted to CVPR 2025. Project page: https://cdfan0627.github.io/spectromotion/
♻ ☆ EmoAttack: Emotion-to-Image Diffusion Models for Emotional Backdoor Generation
Text-to-image diffusion models can generate realistic images based on textual inputs, enabling users to convey their opinions visually through language. Meanwhile, within language, emotion plays a crucial role in expressing personal opinions in our daily lives and the inclusion of maliciously negative content can lead users astray, exacerbating negative emotions. Recognizing the success of diffusion models and the significance of emotion, we investigate a previously overlooked risk associated with text-to-image diffusion models, that is, utilizing emotion in the input texts to introduce negative content and provoke unfavorable emotions in users. Specifically, we identify a new backdoor attack, i.e., emotion-aware backdoor attack (EmoAttack), which introduces malicious negative content triggered by emotional texts during image generation. We formulate such an attack as a diffusion personalization problem to avoid extensive model retraining and propose the EmoBooth. Unlike existing personalization methods, our approach fine-tunes a pre-trained diffusion model by establishing a mapping between a cluster of emotional words and a given reference image containing malicious negative content. To validate the effectiveness of our method, we built a dataset and conducted extensive analysis and discussion about its effectiveness. Given consumers' widespread use of diffusion models, uncovering this threat is critical for society.
♻ ☆ FrugalNeRF: Fast Convergence for Few-shot Novel View Synthesis without Learned Priors CVPR 2025
Neural Radiance Fields (NeRF) face significant challenges in extreme few-shot scenarios, primarily due to overfitting and long training times. Existing methods, such as FreeNeRF and SparseNeRF, use frequency regularization or pre-trained priors but struggle with complex scheduling and bias. We introduce FrugalNeRF, a novel few-shot NeRF framework that leverages weight-sharing voxels across multiple scales to efficiently represent scene details. Our key contribution is a cross-scale geometric adaptation scheme that selects pseudo ground truth depth based on reprojection errors across scales. This guides training without relying on externally learned priors, enabling full utilization of the training data. It can also integrate pre-trained priors, enhancing quality without slowing convergence. Experiments on LLFF, DTU, and RealEstate-10K show that FrugalNeRF outperforms other few-shot NeRF methods while significantly reducing training time, making it a practical solution for efficient and accurate 3D scene reconstruction.
comment: Paper accepted to CVPR 2025. Project page: https://linjohnss.github.io/frugalnerf/
♻ ☆ SPA-VL: A Comprehensive Safety Preference Alignment Dataset for Vision Language Model
The emergence of Vision Language Models (VLMs) has brought unprecedented advances in understanding multimodal information. The combination of textual and visual semantics in VLMs is highly complex and diverse, making the safety alignment of these models challenging. Furthermore, due to the limited study on the safety alignment of VLMs, there is a lack of large-scale, high-quality datasets. To address these limitations, we propose a Safety Preference Alignment dataset for Vision Language Models named SPA-VL. In terms of breadth, SPA-VL covers 6 harmfulness domains, 13 categories, and 53 subcategories, and contains 100,788 samples of the quadruple (question, image, chosen response, rejected response). In terms of depth, the responses are collected from 12 open-source (e.g., QwenVL) and closed-source (e.g., Gemini) VLMs to ensure diversity. The construction of preference data is fully automated, and the experimental results indicate that models trained with alignment techniques on the SPA-VL dataset exhibit substantial improvements in harmlessness and helpfulness while maintaining core capabilities. SPA-VL, as a large-scale, high-quality, and diverse dataset, represents a significant milestone in ensuring that VLMs achieve both harmlessness and helpfulness.
♻ ☆ DeNVeR: Deformable Neural Vessel Representations for Unsupervised Video Vessel Segmentation CVPR 2025
This paper presents Deformable Neural Vessel Representations (DeNVeR), an unsupervised approach for vessel segmentation in X-ray angiography videos without annotated ground truth. DeNVeR utilizes optical flow and layer separation techniques, enhancing segmentation accuracy and adaptability through test-time training. Key contributions include a novel layer separation bootstrapping technique, a parallel vessel motion loss, and the integration of Eulerian motion fields for modeling complex vessel dynamics. A significant component of this research is the introduction of the XACV dataset, the first X-ray angiography coronary video dataset with high-quality, manually labeled segmentation ground truth. Extensive evaluations on both XACV and CADICA datasets demonstrate that DeNVeR outperforms current state-of-the-art methods in vessel segmentation accuracy and generalization capability while maintaining temporal coherency.
comment: Paper accepted to CVPR 2025. Project page: https://kirito878.github.io/DeNVeR/
♻ ☆ FIPER: Generalizable Factorized Features for Robust Low-Level Vision Models
In this work, we propose using a unified representation, termed Factorized Features, for low-level vision tasks, where we test on Single Image Super-Resolution (SISR) and Image Compression. Motivated by the shared principles between these tasks, they require recovering and preserving fine image details, whether by enhancing resolution for SISR or reconstructing compressed data for Image Compression. Unlike previous methods that mainly focus on network architecture, our proposed approach utilizes a basis-coefficient decomposition as well as an explicit formulation of frequencies to capture structural components and multi-scale visual features in images, which addresses the core challenges of both tasks. We replace the representation of prior models from simple feature maps with Factorized Features to validate the potential for broad generalizability. In addition, we further optimize the pipelines by leveraging the mergeable-basis property of our Factorized Features, which consolidates shared structures on multi-frame compression and super-resolution. Extensive experiments show that our unified representation delivers state-of-the-art performance, achieving an average relative improvement of 204.4% in PSNR over the baseline in Super-Resolution (SR) and 9.35% BD-rate reduction in Image Compression compared to the previous SOTA.
comment: Project page: https://jayisaking.github.io/FIPER/
♻ ☆ Empowering LLMs to Understand and Generate Complex Vector Graphics CVPR 2025
The unprecedented advancements in Large Language Models (LLMs) have profoundly impacted natural language processing but have yet to fully embrace the realm of scalable vector graphics (SVG) generation. While LLMs encode partial knowledge of SVG data from web pages during training, recent findings suggest that semantically ambiguous and tokenized representations within LLMs may result in hallucinations in vector primitive predictions. Additionally, LLM training typically lacks modeling and understanding of the rendering sequence of vector paths, which can lead to occlusion between output vector primitives. In this paper, we present LLM4SVG, an initial yet substantial step toward bridging this gap by enabling LLMs to better understand and generate vector graphics. LLM4SVG facilitates a deeper understanding of SVG components through learnable semantic tokens, which precisely encode these tokens and their corresponding properties to generate semantically aligned SVG outputs. Using a series of learnable semantic tokens, a structured dataset for instruction following is developed to support comprehension and generation across two primary tasks. Our method introduces a modular architecture to existing large language models, integrating semantic tags, vector instruction encoders, fine-tuned commands, and powerful LLMs to tightly combine geometric, appearance, and language information. To overcome the scarcity of SVG-text instruction data, we developed an automated data generation pipeline that collected our SVGX-SFT Dataset, consisting of high-quality human-designed SVGs and 580k SVG instruction following data specifically crafted for LLM training, which facilitated the adoption of the supervised fine-tuning strategy popular in LLM development.
comment: Accepted by CVPR 2025. Project Page: https://ximinng.github.io/LLM4SVGProject/
♻ ☆ DiffIR2VR-Zero: Zero-Shot Video Restoration with Diffusion-based Image Restoration Models
We present DiffIR2VR-Zero, a zero-shot framework that enables any pre-trained image restoration diffusion model to perform high-quality video restoration without additional training. While image diffusion models have shown remarkable restoration capabilities, their direct application to video leads to temporal inconsistencies, and existing video restoration methods require extensive retraining for different degradation types. Our approach addresses these challenges through two key innovations: a hierarchical latent warping strategy that maintains consistency across both keyframes and local frames, and a hybrid token merging mechanism that adaptively combines optical flow and feature matching. Through extensive experiments, we demonstrate that our method not only maintains the high-quality restoration of base diffusion models but also achieves superior temporal consistency across diverse datasets and degradation conditions, including challenging scenarios like 8$\times$ super-resolution and severe noise. Importantly, our framework works with any image restoration diffusion model, providing a versatile solution for video enhancement without task-specific training or modifications.
comment: Project page: https://jimmycv07.github.io/DiffIR2VR_web/
♻ ☆ Localized Concept Erasure for Text-to-Image Diffusion Models Using Training-Free Gated Low-Rank Adaptation CVPR 2025
Fine-tuning based concept erasing has demonstrated promising results in preventing generation of harmful contents from text-to-image diffusion models by removing target concepts while preserving remaining concepts. To maintain the generation capability of diffusion models after concept erasure, it is necessary to remove only the image region containing the target concept when it locally appears in an image, leaving other regions intact. However, prior arts often compromise fidelity of the other image regions in order to erase the localized target concept appearing in a specific area, thereby reducing the overall performance of image generation. To address these limitations, we first introduce a framework called localized concept erasure, which allows for the deletion of only the specific area containing the target concept in the image while preserving the other regions. As a solution for the localized concept erasure, we propose a training-free approach, dubbed Gated Low-rank adaptation for Concept Erasure (GLoCE), that injects a lightweight module into the diffusion model. GLoCE consists of low-rank matrices and a simple gate, determined only by several generation steps for concepts without training. By directly applying GLoCE to image embeddings and designing the gate to activate only for target concepts, GLoCE can selectively remove only the region of the target concepts, even when target and remaining concepts coexist within an image. Extensive experiments demonstrated GLoCE not only improves the image fidelity to text prompts after erasing the localized target concepts, but also outperforms prior arts in efficacy, specificity, and robustness by large margin and can be extended to mass concept erasure.
comment: Accepted to CVPR 2025
♻ ☆ DepthSplat: Connecting Gaussian Splatting and Depth CVPR 2025
Gaussian splatting and single-view depth estimation are typically studied in isolation. In this paper, we present DepthSplat to connect Gaussian splatting and depth estimation and study their interactions. More specifically, we first contribute a robust multi-view depth model by leveraging pre-trained monocular depth features, leading to high-quality feed-forward 3D Gaussian splatting reconstructions. We also show that Gaussian splatting can serve as an unsupervised pre-training objective for learning powerful depth models from large-scale multi-view posed datasets. We validate the synergy between Gaussian splatting and depth estimation through extensive ablation and cross-task transfer experiments. Our DepthSplat achieves state-of-the-art performance on ScanNet, RealEstate10K and DL3DV datasets in terms of both depth estimation and novel view synthesis, demonstrating the mutual benefits of connecting both tasks. In addition, DepthSplat enables feed-forward reconstruction from 12 input views (512x960 resolutions) in 0.6 seconds.
comment: CVPR 2025, Project page: https://haofeixu.github.io/depthsplat/, Code: https://github.com/cvg/depthsplat
♻ ☆ Interpreting Object-level Foundation Models via Visual Precision Search CVPR 2025
Advances in multimodal pre-training have propelled object-level foundation models, such as Grounding DINO and Florence-2, in tasks like visual grounding and object detection. However, interpreting these models' decisions has grown increasingly challenging. Existing interpretable attribution methods for object-level task interpretation have notable limitations: (1) gradient-based methods lack precise localization due to visual-textual fusion in foundation models, and (2) perturbation-based methods produce noisy saliency maps, limiting fine-grained interpretability. To address these, we propose a Visual Precision Search method that generates accurate attribution maps with fewer regions. Our method bypasses internal model parameters to overcome attribution issues from multimodal fusion, dividing inputs into sparse sub-regions and using consistency and collaboration scores to accurately identify critical decision-making regions. We also conducted a theoretical analysis of the boundary guarantees and scope of applicability of our method. Experiments on RefCOCO, MS COCO, and LVIS show our approach enhances object-level task interpretability over SOTA for Grounding DINO and Florence-2 across various evaluation metrics, with faithfulness gains of 23.7%, 31.6%, and 20.1% on MS COCO, LVIS, and RefCOCO for Grounding DINO, and 102.9% and 66.9% on MS COCO and RefCOCO for Florence-2. Additionally, our method can interpret failures in visual grounding and object detection tasks, surpassing existing methods across multiple evaluation metrics. The code will be released at https://github.com/RuoyuChen10/VPS.
comment: Accepted to CVPR 2025
♻ ☆ When Large Vision-Language Model Meets Large Remote Sensing Imagery: Coarse-to-Fine Text-Guided Token Pruning
Efficient vision-language understanding of large Remote Sensing Images (RSIs) is meaningful but challenging. Current Large Vision-Language Models (LVLMs) typically employ limited pre-defined grids to process images, leading to information loss when handling gigapixel RSIs. Conversely, using unlimited grids significantly increases computational costs. To preserve image details while reducing computational complexity, we propose a text-guided token pruning method with Dynamic Image Pyramid (DIP) integration. Our method introduces: (i) a Region Focus Module (RFM) that leverages text-aware region localization capability to identify critical vision tokens, and (ii) a coarse-to-fine image tile selection and vision token pruning strategy based on DIP, which is guided by RFM outputs and avoids directly processing the entire large imagery. Additionally, existing benchmarks for evaluating LVLMs' perception ability on large RSI suffer from limited question diversity and constrained image sizes. We construct a new benchmark named LRS-VQA, which contains 7,333 QA pairs across 8 categories, with image length up to 27,328 pixels. Our method outperforms existing high-resolution strategies on four datasets using the same data. Moreover, compared to existing token reduction methods, our approach demonstrates higher efficiency under high-resolution settings. Dataset and code are in https://github.com/VisionXLab/LRS-VQA.
comment: 12 pages, 6 figures, 7 tables
♻ ☆ Interpretable Bilingual Multimodal Large Language Model for Diverse Biomedical Tasks ICLR 2025
Several medical Multimodal Large Languange Models (MLLMs) have been developed to address tasks involving visual images with textual instructions across various medical modalities, achieving impressive results. Most current medical generalist models are region-agnostic, treating the entire image as a holistic representation. However, they struggle to identify which specific regions they are focusing on when generating a sentence. To mimic the behavior of doctors, who typically begin by reviewing the entire image before concentrating on specific regions for a thorough evaluation, we aim to enhance the capability of medical MLLMs in understanding anatomical regions within entire medical scans. To achieve it, we first formulate Region-Centric tasks and construct a large-scale dataset, MedRegInstruct, to incorporate regional information into training. Combining our collected dataset with other medical multimodal corpora for training, we propose a Region-Aware medical MLLM, MedRegA, which is the first bilingual generalist medical AI system to simultaneously handle image-level and region-level medical vision-language tasks across a broad range of modalities. Our MedRegA not only enables three region-centric tasks, but also achieves the best performance for visual question answering, report generation and medical image classification over 8 modalities, showcasing significant versatility. Experiments demonstrate that our model can not only accomplish powerful performance across various medical vision-language tasks in bilingual settings, but also recognize and detect structures in multimodal medical scans, boosting the interpretability and user interactivity of medical MLLMs. Our project page is https://medrega.github.io.
comment: Accepted in ICLR 2025
♻ ☆ Pfungst and Clever Hans: Identifying the unintended cues in a widely used Alzheimer's disease MRI dataset using explainable deep learning
Backgrounds. Deep neural networks have demonstrated high accuracy in classifying Alzheimer's disease (AD). This study aims to enlighten the underlying black-box nature and reveal individual contributions of T1-weighted (T1w) gray-white matter texture, volumetric information and preprocessing on classification performance. Methods. We utilized T1w MRI data from the Alzheimer's Disease Neuroimaging Initiative to distinguish matched AD patients (990 MRIs) from healthy controls (990 MRIs). Preprocessing included skull stripping and binarization at varying thresholds to systematically eliminate texture information. A deep neural network was trained on these configurations, and the model performance was compared using McNemar tests with discrete Bonferroni-Holm correction. Layer-wise Relevance Propagation (LRP) and structural similarity metrics between heatmaps were applied to analyze learned features. Results. Classification performance metrics (accuracy, sensitivity, and specificity) were comparable across all configurations, indicating a negligible influence of T1w gray- and white signal texture. Models trained on binarized images demonstrated similar feature performance and relevance distributions, with volumetric features such as atrophy and skull-stripping features emerging as primary contributors. Conclusions. We revealed a previously undiscovered Clever Hans effect in a widely used AD MRI dataset. Deep neural networks classification predominantly rely on volumetric features, while eliminating gray-white matter T1w texture did not decrease the performance. This study clearly demonstrates an overestimation of the importance of gray-white matter contrasts, at least for widely used structural T1w images, and highlights potential misinterpretation of performance metrics.
♻ ☆ CLIP-Adapter: Better Vision-Language Models with Feature Adapters
Large-scale contrastive vision-language pre-training has shown significant progress in visual representation learning. Unlike traditional visual systems trained by a fixed set of discrete labels, a new paradigm was introduced in \cite{radford2021learning} to directly learn to align images with raw texts in an open-vocabulary setting. On downstream tasks, a carefully chosen text prompt is employed to make zero-shot predictions.~To avoid non-trivial prompt engineering, context optimization \cite{zhou2021coop} has been proposed to learn continuous vectors as task-specific prompts with few-shot training examples.~In this paper, we show that there is an alternative path to achieve better vision-language models other than prompt tuning.~While prompt tuning is for the textual inputs, we propose CLIP-Adapter to conduct fine-tuning with feature adapters on either visual or language branch. Specifically, CLIP-Adapter adopts an additional bottleneck layer to learn new features and performs residual-style feature blending with the original pre-trained features.~As a consequence, CLIP-Adapter is able to outperform context optimization while maintains a simple design. Experiments and extensive ablation studies on various visual classification tasks demonstrate the effectiveness of our approach. Code is released at t https://github.com/gaopengcuhk/CLIP-Adapter.
comment: Accepted by IJCV
♻ ☆ SG-GAN: Fine Stereoscopic-Aware Generation for 3D Brain Point Cloud Up-sampling from a Single Image
In minimally-invasive brain surgeries with indirect and narrow operating environments, 3D brain reconstruction is crucial. However, as requirements of accuracy for some new minimally-invasive surgeries (such as brain-computer interface surgery) are higher and higher, the outputs of conventional 3D reconstruction, such as point cloud (PC), are facing the challenges that sample points are too sparse and the precision is insufficient. On the other hand, there is a scarcity of high-density point cloud datasets, which makes it challenging to train models for direct reconstruction of high-density brain point clouds. In this work, a novel model named stereoscopic-aware graph generative adversarial network (SG-GAN) with two stages is proposed to generate fine high-density PC conditioned on a single image. The Stage-I GAN sketches the primitive shape and basic structure of the organ based on the given image, yielding Stage-I point clouds. The Stage-II GAN takes the results from Stage-I and generates high-density point clouds with detailed features. The Stage-II GAN is capable of correcting defects and restoring the detailed features of the region of interest (ROI) through the up-sampling process. Furthermore, a parameter-free-attention-based free-transforming module is developed to learn the efficient features of input, while upholding a promising performance. Comparing with the existing methods, the SG-GAN model shows superior performance in terms of visual quality, objective measurements, and performance in classification, as demonstrated by comprehensive results measured by several evaluation metrics including PC-to-PC error and Chamfer distance.
comment: Accepted by TETCI
Machine Learning 169
☆ From Interpretation to Correction: A Decentralized Optimization Framework for Exact Convergence in Federated Learning
This work introduces a novel decentralized framework to interpret federated learning (FL) and, consequently, correct the biases introduced by arbitrary client participation and data heterogeneity, which are two typical traits in practical FL. Specifically, we first reformulate the core processes of FedAvg - client participation, local updating, and model aggregation - as stochastic matrix multiplications. This reformulation allows us to interpret FedAvg as a decentralized algorithm. Leveraging the decentralized optimization framework, we are able to provide a concise analysis to quantify the impact of arbitrary client participation and data heterogeneity on FedAvg's convergence point. This insight motivates the development of Federated Optimization with Exact Convergence via Push-pull Strategy (FOCUS), a novel algorithm inspired by the decentralized algorithm that eliminates these biases and achieves exact convergence without requiring the bounded heterogeneity assumption. Furthermore, we theoretically prove that FOCUS exhibits linear convergence (exponential decay) for both strongly convex and non-convex functions satisfying the Polyak-Lojasiewicz condition, regardless of the arbitrary nature of client participation.
☆ Domain Adaptation Framework for Turning Movement Count Estimation with Limited Data
Urban transportation networks are vital for the efficient movement of people and goods, necessitating effective traffic management and planning. An integral part of traffic management is understanding the turning movement counts (TMCs) at intersections, Accurate TMCs at intersections are crucial for traffic signal control, congestion mitigation, and road safety. In general, TMCs are obtained using physical sensors installed at intersections, but this approach can be cost-prohibitive and technically challenging, especially for cities with extensive road networks. Recent advancements in machine learning and data-driven approaches have offered promising alternatives for estimating TMCs. Traffic patterns can vary significantly across different intersections due to factors such as road geometry, traffic signal settings, and local driver behaviors. This domain discrepancy limits the generalizability and accuracy of machine learning models when applied to new or unseen intersections. In response to these limitations, this research proposes a novel framework leveraging domain adaptation (DA) to estimate TMCs at intersections by using traffic controller event-based data, road infrastructure data, and point-of-interest (POI) data. Evaluated on 30 intersections in Tucson, Arizona, the performance of the proposed DA framework was compared with state-of-the-art models and achieved the lowest values in terms of Mean Absolute Error and Root Mean Square Error.
comment: arXiv admin note: substantial text overlap with arXiv:2412.09861
☆ Efficient Model Development through Fine-tuning Transfer
Modern LLMs struggle with efficient updates, as each new pretrained model version requires repeating expensive alignment processes. This challenge also applies to domain- or language-specific models, where fine-tuning on specialized data must be redone for every new base model release. In this paper, we explore the transfer of fine-tuning updates between model versions. Specifically, we derive the diff vector from one source model version, which represents the weight changes from fine-tuning, and apply it to the base model of a different target version. Through empirical evaluations on various open-weight model versions, we show that transferring diff vectors can significantly improve the target base model, often achieving performance comparable to its fine-tuned counterpart. For example, reusing the fine-tuning updates from Llama 3.0 8B leads to an absolute accuracy improvement of 10.7% on GPQA over the base Llama 3.1 8B without additional training, surpassing Llama 3.1 8B Instruct. In a multilingual model development setting, we show that this approach can significantly increase performance on target-language tasks without retraining, achieving an absolute improvement of 4.7% and 15.5% on Global MMLU for Malagasy and Turkish, respectively, compared to Llama 3.1 8B Instruct. Our controlled experiments reveal that fine-tuning transfer is most effective when the source and target models are linearly connected in the parameter space. Additionally, we demonstrate that fine-tuning transfer offers a stronger and more computationally efficient starting point for further fine-tuning. Finally, we propose an iterative recycling-then-finetuning approach for continuous model development, which improves both efficiency and effectiveness. Our findings suggest that fine-tuning transfer is a viable strategy to reduce training costs while maintaining model performance.
comment: 21 pages, 4 figures, 13 tables
Extendable Long-Horizon Planning via Hierarchical Multiscale Diffusion
This paper tackles a novel problem, extendable long-horizon planning-enabling agents to plan trajectories longer than those in training data without compounding errors. To tackle this, we propose the Hierarchical Multiscale Diffuser (HM-Diffuser) and Progressive Trajectory Extension (PTE), an augmentation method that iteratively generates longer trajectories by stitching shorter ones. HM-Diffuser trains on these extended trajectories using a hierarchical structure, efficiently handling tasks across multiple temporal scales. Additionally, we introduce Adaptive Plan Pondering and the Recursive HM-Diffuser, which consolidate hierarchical layers into a single model to process temporal scales recursively. Experimental results demonstrate the effectiveness of our approach, advancing diffusion-based planners for scalable long-horizon planning.
comment: First two authors contributed equally
☆ Fundamental Limits of Perfect Concept Erasure AISTATS 2025
Concept erasure is the task of erasing information about a concept (e.g., gender or race) from a representation set while retaining the maximum possible utility -- information from original representations. Concept erasure is useful in several applications, such as removing sensitive concepts to achieve fairness and interpreting the impact of specific concepts on a model's performance. Previous concept erasure techniques have prioritized robustly erasing concepts over retaining the utility of the resultant representations. However, there seems to be an inherent tradeoff between erasure and retaining utility, making it unclear how to achieve perfect concept erasure while maintaining high utility. In this paper, we offer a fresh perspective toward solving this problem by quantifying the fundamental limits of concept erasure through an information-theoretic lens. Using these results, we investigate constraints on the data distribution and the erasure functions required to achieve the limits of perfect concept erasure. Empirically, we show that the derived erasure functions achieve the optimal theoretical bounds. Additionally, we show that our approach outperforms existing methods on a range of synthetic and real-world datasets using GPT-4 representations.
comment: Accepted at AISTATS 2025
☆ Random feature-based double Vovk-Azoury-Warmuth algorithm for online multi-kernel learning
We introduce a novel multi-kernel learning algorithm, VAW$^2$, for online least squares regression in reproducing kernel Hilbert spaces (RKHS). VAW$^2$ leverages random Fourier feature-based functional approximation and the Vovk-Azoury-Warmuth (VAW) method in a two-level procedure: VAW is used to construct expert strategies from random features generated for each kernel at the first level, and then again to combine their predictions at the second level. A theoretical analysis yields a regret bound of $O(T^{1/2}\ln T)$ in expectation with respect to artificial randomness, when the number of random features scales as $T^{1/2}$. Empirical results on some benchmark datasets demonstrate that VAW$^2$ achieves superior performance compared to the existing online multi-kernel learning algorithms: Raker and OMKL-GF, and to other theoretically grounded method methods involving convex combination of expert predictions at the second level.
☆ Abstracting Geo-specific Terrains to Scale Up Reinforcement Learning
Multi-agent reinforcement learning (MARL) is increasingly ubiquitous in training dynamic and adaptive synthetic characters for interactive simulations on geo-specific terrains. Frameworks such as Unity's ML-Agents help to make such reinforcement learning experiments more accessible to the simulation community. Military training simulations also benefit from advances in MARL, but they have immense computational requirements due to their complex, continuous, stochastic, partially observable, non-stationary, and doctrine-based nature. Furthermore, these simulations require geo-specific terrains, further exacerbating the computational resources problem. In our research, we leverage Unity's waypoints to automatically generate multi-layered representation abstractions of the geo-specific terrains to scale up reinforcement learning while still allowing the transfer of learned policies between different representations. Our early exploratory results on a novel MARL scenario, where each side has differing objectives, indicate that waypoint-based navigation enables faster and more efficient learning while producing trajectories similar to those taken by expert human players in CSGO gaming environments. This research points out the potential of waypoint-based navigation for reducing the computational costs of developing and training MARL models for military training simulations, where geo-specific terrains and differing objectives are crucial.
comment: 10 pages, 6 figures, 2024 Interservice/Industry Training, Simulation, and Education Conference (I/ITSEC)
☆ Peer Disambiguation in Self-Reported Surveys using Graph Attention Networks
Studying peer relationships is crucial in solving complex challenges underserved communities face and designing interventions. The effectiveness of such peer-based interventions relies on accurate network data regarding individual attributes and social influences. However, these datasets are often collected through self-reported surveys, introducing ambiguities in network construction. These ambiguities make it challenging to fully utilize the network data to understand the issues and to design the best interventions. We propose and solve two variations of link ambiguities in such network data -- (i) which among the two candidate links exists, and (ii) if a candidate link exists. We design a Graph Attention Network (GAT) that accounts for personal attributes and network relationships on real-world data with real and simulated ambiguities. We also demonstrate that by resolving these ambiguities, we improve network accuracy, and in turn, improve suicide risk prediction. We also uncover patterns using GNNExplainer to provide additional insights into vital features and relationships. This research demonstrates the potential of Graph Neural Networks (GNN) to advance real-world network data analysis facilitating more effective peer interventions across various fields.
☆ Deep Learning Approaches for Blood Disease Diagnosis Across Hematopoietic Lineages
We present a foundation modeling framework that leverages deep learning to uncover latent genetic signatures across the hematopoietic hierarchy. Our approach trains a fully connected autoencoder on multipotent progenitor cells, reducing over 20,000 gene features to a 256-dimensional latent space that captures predictive information for both progenitor and downstream differentiated cells such as monocytes and lymphocytes. We validate the quality of these embeddings by training feed-forward, transformer, and graph convolutional architectures for blood disease diagnosis tasks. We also explore zero-shot prediction using a progenitor disease state classification model to classify downstream cell conditions. Our models achieve greater than 95% accuracy for multi-class classification, and in the zero-shot setting, we achieve greater than 0.7 F1-score on the binary classification task. Future work should improve embeddings further to increase robustness on lymphocyte classification specifically.
comment: 6 pages, 4 figures
A scalable gene network model of regulatory dynamics in single cells
Single-cell data provide high-dimensional measurements of the transcriptional states of cells, but extracting insights into the regulatory functions of genes, particularly identifying transcriptional mechanisms affected by biological perturbations, remains a challenge. Many perturbations induce compensatory cellular responses, making it difficult to distinguish direct from indirect effects on gene regulation. Modeling how gene regulatory functions shape the temporal dynamics of these responses is key to improving our understanding of biological perturbations. Dynamical models based on differential equations offer a principled way to capture transcriptional dynamics, but their application to single-cell data has been hindered by computational constraints, stochasticity, sparsity, and noise. Existing methods either rely on low-dimensional representations or make strong simplifying assumptions, limiting their ability to model transcriptional dynamics at scale. We introduce a Functional and Learnable model of Cell dynamicS, FLeCS, that incorporates gene network structure into coupled differential equations to model gene regulatory functions. Given (pseudo)time-series single-cell data, FLeCS accurately infers cell dynamics at scale, provides improved functional insights into transcriptional mechanisms perturbed by gene knockouts, both in myeloid differentiation and K562 Perturb-seq experiments, and simulates single-cell trajectories of A549 cells following small-molecule perturbations.
comment: 42 pages, 10 figures
☆ Experience Replay Addresses Loss of Plasticity in Continual Learning
Loss of plasticity is one of the main challenges in continual learning with deep neural networks, where neural networks trained via backpropagation gradually lose their ability to adapt to new tasks and perform significantly worse than their freshly initialized counterparts. The main contribution of this paper is to propose a new hypothesis that experience replay addresses the loss of plasticity in continual learning. Here, experience replay is a form of memory. We provide supporting evidence for this hypothesis. In particular, we demonstrate in multiple different tasks, including regression, classification, and policy evaluation, that by simply adding an experience replay and processing the data in the experience replay with Transformers, the loss of plasticity disappears. Notably, we do not alter any standard components of deep learning. For example, we do not change backpropagation. We do not modify the activation functions. And we do not use any regularization. We conjecture that experience replay and Transformers can address the loss of plasticity because of the in-context learning phenomenon.
comment: 14 pages, 4 figures
☆ Low-resource Machine Translation for Code-switched Kazakh-Russian Language Pair
Machine translation for low resource language pairs is a challenging task. This task could become extremely difficult once a speaker uses code switching. We propose a method to build a machine translation model for code-switched Kazakh-Russian language pair with no labeled data. Our method is basing on generation of synthetic data. Additionally, we present the first codeswitching Kazakh-Russian parallel corpus and the evaluation results, which include a model achieving 16.48 BLEU almost reaching an existing commercial system and beating it by human evaluation.
☆ Unsupervised Learning for Quadratic Assignment
We introduce PLUME search, a data-driven framework that enhances search efficiency in combinatorial optimization through unsupervised learning. Unlike supervised or reinforcement learning, PLUME search learns directly from problem instances using a permutation-based loss with a non-autoregressive approach. We evaluate its performance on the quadratic assignment problem, a fundamental NP-hard problem that encompasses various combinatorial optimization problems. Experimental results demonstrate that PLUME search consistently improves solution quality. Furthermore, we study the generalization behavior and show that the learned model generalizes across different densities and sizes.
comment: preprint
☆ The Coralscapes Dataset: Semantic Scene Understanding in Coral Reefs
Coral reefs are declining worldwide due to climate change and local stressors. To inform effective conservation or restoration, monitoring at the highest possible spatial and temporal resolution is necessary. Conventional coral reef surveying methods are limited in scalability due to their reliance on expert labor time, motivating the use of computer vision tools to automate the identification and abundance estimation of live corals from images. However, the design and evaluation of such tools has been impeded by the lack of large high quality datasets. We release the Coralscapes dataset, the first general-purpose dense semantic segmentation dataset for coral reefs, covering 2075 images, 39 benthic classes, and 174k segmentation masks annotated by experts. Coralscapes has a similar scope and the same structure as the widely used Cityscapes dataset for urban scene segmentation, allowing benchmarking of semantic segmentation models in a new challenging domain which requires expert knowledge to annotate. We benchmark a wide range of semantic segmentation models, and find that transfer learning from Coralscapes to existing smaller datasets consistently leads to state-of-the-art performance. Coralscapes will catalyze research on efficient, scalable, and standardized coral reef surveying methods based on computer vision, and holds the potential to streamline the development of underwater ecological robotics.
☆ ExCoT: Optimizing Reasoning for Text-to-SQL with Execution Feedback
Text-to-SQL demands precise reasoning to convert natural language questions into structured queries. While large language models (LLMs) excel in many reasoning tasks, their ability to leverage Chain-of-Thought (CoT) reasoning for text-to-SQL remains underexplored. We identify critical limitations: zero-shot CoT offers minimal gains, and Direct Preference Optimization (DPO) applied without CoT yields marginal improvements. We propose ExCoT, a novel framework that iteratively optimizes open-source LLMs by combining CoT reasoning with off-policy and on-policy DPO, relying solely on execution accuracy as feedback. This approach eliminates the need for reward models or human-annotated preferences. Our experimental results demonstrate significant performance gains: ExCoT improves execution accuracy on BIRD dev set from 57.37% to 68.51% and on Spider test set from 78.81% to 86.59% for LLaMA-3 70B, with Qwen-2.5-Coder demonstrating similar improvements. Our best model achieves state-of-the-art performance in the single-model setting on both BIRD and Spider datasets, notably achieving 68.53% on the BIRD test set.
☆ Thin-Shell-SfT: Fine-Grained Monocular Non-rigid 3D Surface Tracking with Neural Deformation Fields CVPR 2025
3D reconstruction of highly deformable surfaces (e.g. cloths) from monocular RGB videos is a challenging problem, and no solution provides a consistent and accurate recovery of fine-grained surface details. To account for the ill-posed nature of the setting, existing methods use deformation models with statistical, neural, or physical priors. They also predominantly rely on nonadaptive discrete surface representations (e.g. polygonal meshes), perform frame-by-frame optimisation leading to error propagation, and suffer from poor gradients of the mesh-based differentiable renderers. Consequently, fine surface details such as cloth wrinkles are often not recovered with the desired accuracy. In response to these limitations, we propose ThinShell-SfT, a new method for non-rigid 3D tracking that represents a surface as an implicit and continuous spatiotemporal neural field. We incorporate continuous thin shell physics prior based on the Kirchhoff-Love model for spatial regularisation, which starkly contrasts the discretised alternatives of earlier works. Lastly, we leverage 3D Gaussian splatting to differentiably render the surface into image space and optimise the deformations based on analysis-bysynthesis principles. Our Thin-Shell-SfT outperforms prior works qualitatively and quantitatively thanks to our continuous surface formulation in conjunction with a specially tailored simulation prior and surface-induced 3D Gaussians. See our project page at https://4dqv.mpiinf.mpg.de/ThinShellSfT.
comment: 15 pages, 12 figures and 3 tables; project page: https://4dqv.mpiinf.mpg.de/ThinShellSfT; CVPR 2025
☆ SuperFlow++: Enhanced Spatiotemporal Consistency for Cross-Modal Data Pretraining
LiDAR representation learning has emerged as a promising approach to reducing reliance on costly and labor-intensive human annotations. While existing methods primarily focus on spatial alignment between LiDAR and camera sensors, they often overlook the temporal dynamics critical for capturing motion and scene continuity in driving scenarios. To address this limitation, we propose SuperFlow++, a novel framework that integrates spatiotemporal cues in both pretraining and downstream tasks using consecutive LiDAR-camera pairs. SuperFlow++ introduces four key components: (1) a view consistency alignment module to unify semantic information across camera views, (2) a dense-to-sparse consistency regularization mechanism to enhance feature robustness across varying point cloud densities, (3) a flow-based contrastive learning approach that models temporal relationships for improved scene understanding, and (4) a temporal voting strategy that propagates semantic information across LiDAR scans to improve prediction consistency. Extensive evaluations on 11 heterogeneous LiDAR datasets demonstrate that SuperFlow++ outperforms state-of-the-art methods across diverse tasks and driving conditions. Furthermore, by scaling both 2D and 3D backbones during pretraining, we uncover emergent properties that provide deeper insights into developing scalable 3D foundation models. With strong generalizability and computational efficiency, SuperFlow++ establishes a new benchmark for data-efficient LiDAR-based perception in autonomous driving. The code is publicly available at https://github.com/Xiangxu-0103/SuperFlow
comment: Preprint; 15 pages, 6 figures, 10 tables; Code at https://github.com/Xiangxu-0103/SuperFlow
☆ Tracktention: Leveraging Point Tracking to Attend Videos Faster and Better CVPR 2025
Temporal consistency is critical in video prediction to ensure that outputs are coherent and free of artifacts. Traditional methods, such as temporal attention and 3D convolution, may struggle with significant object motion and may not capture long-range temporal dependencies in dynamic scenes. To address this gap, we propose the Tracktention Layer, a novel architectural component that explicitly integrates motion information using point tracks, i.e., sequences of corresponding points across frames. By incorporating these motion cues, the Tracktention Layer enhances temporal alignment and effectively handles complex object motions, maintaining consistent feature representations over time. Our approach is computationally efficient and can be seamlessly integrated into existing models, such as Vision Transformers, with minimal modification. It can be used to upgrade image-only models to state-of-the-art video ones, sometimes outperforming models natively designed for video prediction. We demonstrate this on video depth prediction and video colorization, where models augmented with the Tracktention Layer exhibit significantly improved temporal consistency compared to baselines.
comment: CVPR 2025. Project website: zlai0.github.io/TrackTention
☆ RCC-PFL: Robust Client Clustering under Noisy Labels in Personalized Federated Learning
We address the problem of cluster identity estimation in a personalized federated learning (PFL) setting in which users aim to learn different personal models. The backbone of effective learning in such a setting is to cluster users into groups whose objectives are similar. A typical approach in the literature is to achieve this by training users' data on different proposed personal models and assign them to groups based on which model achieves the lowest value of the users' loss functions. This process is to be done iteratively until group identities converge. A key challenge in such a setting arises when users have noisy labeled data, which may produce misleading values of their loss functions, and hence lead to ineffective clustering. To overcome this challenge, we propose a label-agnostic data similarity-based clustering algorithm, coined RCC-PFL, with three main advantages: the cluster identity estimation procedure is independent from the training labels; it is a one-shot clustering algorithm performed prior to the training; and it requires fewer communication rounds and less computation compared to iterative-based clustering methods. We validate our proposed algorithm using various models and datasets and show that it outperforms multiple baselines in terms of average accuracy and variance reduction.
comment: to appear in the 2025 IEEE International Conference on Communications
☆ Extensions of regret-minimization algorithm for optimal design
We explore extensions and applications of the regret minimization framework introduced by~\cite{design} for solving optimal experimental design problems. Specifically, we incorporate the entropy regularizer into this framework, leading to a novel sample selection objective and a provable sample complexity bound that guarantees a $(1+\epsilon)$-near optimal solution. We further extend the method to handle regularized optimal design settings. As an application, we use our algorithm to select a small set of representative samples from image classification datasets without relying on label information. To evaluate the quality of the selected samples, we train a logistic regression model and compare performance against several baseline sampling strategies. Experimental results on MNIST, CIFAR-10, and a 50-class subset of ImageNet show that our approach consistently outperforms competing methods in most cases.
☆ Identification of Average Treatment Effects in Nonparametric Panel Models
This paper studies identification of average treatment effects in a panel data setting. It introduces a novel nonparametric factor model and proves identification of average treatment effects. The identification proof is based on the introduction of a consistent estimator. Underlying the proof is a result that there is a consistent estimator for the expected outcome in the absence of the treatment for each unit and time period; this result can be applied more broadly, for example in problems of decompositions of group-level differences in outcomes, such as the much-studied gender wage gap.
☆ Geometric Meta-Learning via Coupled Ricci Flow: Unifying Knowledge Representation and Quantum Entanglement
This paper establishes a unified framework integrating geometric flows with deep learning through three fundamental innovations. First, we propose a thermodynamically coupled Ricci flow that dynamically adapts parameter space geometry to loss landscape topology, formally proved to preserve isometric knowledge embedding (Theorem~\ref{thm:isometric}). Second, we derive explicit phase transition thresholds and critical learning rates (Theorem~\ref{thm:critical}) through curvature blowup analysis, enabling automated singularity resolution via geometric surgery (Lemma~\ref{lem:surgery}). Third, we establish an AdS/CFT-type holographic duality (Theorem~\ref{thm:ads}) between neural networks and conformal field theories, providing entanglement entropy bounds for regularization design. Experiments demonstrate 2.1$\times$ convergence acceleration and 63\% topological simplification while maintaining $\mathcal{O}(N\log N)$ complexity, outperforming Riemannian baselines by 15.2\% in few-shot accuracy. Theoretically, we prove exponential stability (Theorem~\ref{thm:converge}) through a new Lyapunov function combining Perelman entropy with Wasserstein gradient flows, fundamentally advancing geometric deep learning.
comment: 9 pages, submitted to IEEE PAMI
☆ GENIUS: A Generative Framework for Universal Multimodal Search CVPR 2025
Generative retrieval is an emerging approach in information retrieval that generates identifiers (IDs) of target data based on a query, providing an efficient alternative to traditional embedding-based retrieval methods. However, existing models are task-specific and fall short of embedding-based retrieval in performance. This paper proposes GENIUS, a universal generative retrieval framework supporting diverse tasks across multiple modalities and domains. At its core, GENIUS introduces modality-decoupled semantic quantization, transforming multimodal data into discrete IDs encoding both modality and semantics. Moreover, to enhance generalization, we propose a query augmentation that interpolates between a query and its target, allowing GENIUS to adapt to varied query forms. Evaluated on the M-BEIR benchmark, it surpasses prior generative methods by a clear margin. Unlike embedding-based retrieval, GENIUS consistently maintains high retrieval speed across database size, with competitive performance across multiple benchmarks. With additional re-ranking, GENIUS often achieves results close to those of embedding-based methods while preserving efficiency.
comment: Accepted to CVPR 2025
☆ An Overview of Low-Rank Structures in the Training and Adaptation of Large Models
The rise of deep learning has revolutionized data processing and prediction in signal processing and machine learning, yet the substantial computational demands of training and deploying modern large-scale deep models present significant challenges, including high computational costs and energy consumption. Recent research has uncovered a widespread phenomenon in deep networks: the emergence of low-rank structures in weight matrices and learned representations during training. These implicit low-dimensional patterns provide valuable insights for improving the efficiency of training and fine-tuning large-scale models. Practical techniques inspired by this phenomenon, such as low-rank adaptation (LoRA) and training, enable significant reductions in computational cost while preserving model performance. In this paper, we present a comprehensive review of recent advances in exploiting low-rank structures for deep learning and shed light on their mathematical foundations. Mathematically, we present two complementary perspectives on understanding the low-rankness in deep networks: (i) the emergence of low-rank structures throughout the whole optimization dynamics of gradient and (ii) the implicit regularization effects that induce such low-rank structures at convergence. From a practical standpoint, studying the low-rank learning dynamics of gradient descent offers a mathematical foundation for understanding the effectiveness of LoRA in fine-tuning large-scale models and inspires parameter-efficient low-rank training strategies. Furthermore, the implicit low-rank regularization effect helps explain the success of various masked training approaches in deep neural networks, ranging from dropout to masked self-supervised learning.
comment: Authors are listed alphabetically; 27 pages, 10 figures
☆ Capacity-Constrained Online Learning with Delays: Scheduling Frameworks and Regret Trade-offs
We study online learning with oblivious losses and delays under a novel ``capacity constraint'' that limits how many past rounds can be tracked simultaneously for delayed feedback. Under ``clairvoyance'' (i.e., delay durations are revealed upfront each round) and/or ``preemptibility'' (i.e., we have ability to stop tracking previously chosen round feedback), we establish matching upper and lower bounds (up to logarithmic terms) on achievable regret, characterizing the ``optimal capacity'' needed to match the minimax rates of classical delayed online learning, which implicitly assume unlimited capacity. Our algorithms achieve minimax-optimal regret across all capacity levels, with performance gracefully degrading under suboptimal capacity. For $K$ actions and total delay $D$ over $T$ rounds, under clairvoyance and assuming capacity $C = \Omega(\log(T))$, we achieve regret $\widetilde{\Theta}(\sqrt{TK + DK/C + D\log(K)})$ for bandits and $\widetilde{\Theta}(\sqrt{(D+T)\log(K)})$ for full-information feedback. When replacing clairvoyance with preemptibility, we require a known maximum delay bound $d_{\max}$, adding $\smash{\widetilde{O}(d_{\max})}$ to the regret. For fixed delays $d$ (i.e., $D=Td$), the minimax regret is $\Theta\bigl(\sqrt{TK(1+d/C)+Td\log(K)}\bigr)$ and the optimal capacity is $\Theta(\min\{K/\log(K),d\}\bigr)$ in the bandit setting, while in the full-information setting, the minimax regret is $\Theta\bigl(\sqrt{T(d+1)\log(K)}\bigr)$ and the optimal capacity is $\Theta(1)$. For round-dependent and fixed delays, our upper bounds are achieved using novel scheduling policies, based on Pareto-distributed proxy delays and batching techniques. Crucially, our work unifies delayed bandits, label-efficient learning, and online scheduling frameworks, demonstrating that robust online learning under delayed feedback is possible with surprisingly modest tracking capacity.
☆ Ab-initio simulation of excited-state potential energy surfaces with transferable deep quantum Monte Carlo
The accurate quantum chemical calculation of excited states is a challenging task, often requiring computationally demanding methods. When entire ground and excited potential energy surfaces (PESs) are desired, e.g., to predict the interaction of light excitation and structural changes, one is often forced to use cheaper computational methods at the cost of reduced accuracy. Here we introduce a novel method for the geometrically transferable optimization of neural network wave functions that leverages weight sharing and dynamical ordering of electronic states. Our method enables the efficient prediction of ground and excited-state PESs and their intersections at the highest accuracy, demonstrating up to two orders of magnitude cost reduction compared to single-point calculations. We validate our approach on three challenging excited-state PESs, including ethylene, the carbon dimer, and the methylenimmonium cation, indicating that transferable deep-learning QMC can pave the way towards highly accurate simulation of excited-state dynamics.
comment: 21 pages, 4 figures
☆ IgCraft: A versatile sequence generation framework for antibody discovery and engineering
Designing antibody sequences to better resemble those observed in natural human repertoires is a key challenge in biologics development. We introduce IgCraft: a multi-purpose model for paired human antibody sequence generation, built on Bayesian Flow Networks. IgCraft presents one of the first unified generative modeling frameworks capable of addressing multiple antibody sequence design tasks with a single model, including unconditional sampling, sequence inpainting, inverse folding, and CDR motif scaffolding. Our approach achieves competitive results across the full spectrum of these tasks while constraining generation to the space of human antibody sequences, exhibiting particular strengths in CDR motif scaffolding (grafting) where we achieve state-of-the-art performance in terms of humanness and preservation of structural properties. By integrating previously separate tasks into a single scalable generative model, IgCraft provides a versatile platform for sampling human antibody sequences under a variety of contexts relevant to antibody discovery and engineering. Model code and weights are publicly available at github.com/mgreenig/IgCraft.
☆ A Systematic Review of EEG-based Machine Intelligence Algorithms for Depression Diagnosis, and Monitoring
Depression disorder is a serious health condition that has affected the lives of millions of people around the world. Diagnosis of depression is a challenging practice that relies heavily on subjective studies and, in most cases, suffers from late findings. Electroencephalography (EEG) biomarkers have been suggested and investigated in recent years as a potential transformative objective practice. In this article, for the first time, a detailed systematic review of EEG-based depression diagnosis approaches is conducted using advanced machine learning techniques and statistical analyses. For this, 938 potentially relevant articles (since 1985) were initially detected and filtered into 139 relevant articles based on the review scheme 'preferred reporting items for systematic reviews and meta-analyses (PRISMA).' This article compares and discusses the selected articles and categorizes them according to the type of machine learning techniques and statistical analyses. Algorithms, preprocessing techniques, extracted features, and data acquisition systems are discussed and summarized. This review paper explains the existing challenges of the current algorithms and sheds light on the future direction of the field. This systematic review outlines the issues and challenges in machine intelligence for the diagnosis of EEG depression that can be addressed in future studies and possibly in future wearable technologies.
☆ Domain-incremental White Blood Cell Classification with Privacy-aware Continual Learning
White blood cell (WBC) classification plays a vital role in hematology for diagnosing various medical conditions. However, it faces significant challenges due to domain shifts caused by variations in sample sources (e.g., blood or bone marrow) and differing imaging conditions across hospitals. Traditional deep learning models often suffer from catastrophic forgetting in such dynamic environments, while foundation models, though generally robust, experience performance degradation when the distribution of inference data differs from that of the training data. To address these challenges, we propose a generative replay-based Continual Learning (CL) strategy designed to prevent forgetting in foundation models for WBC classification. Our method employs lightweight generators to mimic past data with a synthetic latent representation to enable privacy-preserving replay. To showcase the effectiveness, we carry out extensive experiments with a total of four datasets with different task ordering and four backbone models including ResNet50, RetCCL, CTransPath, and UNI. Experimental results demonstrate that conventional fine-tuning methods degrade performance on previously learned tasks and struggle with domain shifts. In contrast, our continual learning strategy effectively mitigates catastrophic forgetting, preserving model performance across varying domains. This work presents a practical solution for maintaining reliable WBC classification in real-world clinical settings, where data distributions frequently evolve.
☆ LogQuant: Log-Distributed 2-Bit Quantization of KV Cache with Superior Accuracy Preservation ICLR 2025
We introduce LogQuant, a groundbreaking 2-bit quantization technique for KV Cache in large language model (LLM) inference, delivering substantial memory savings while preserving superior performance. Previous methods either assume that later tokens are more important or attempt to predict important tokens based on earlier attention patterns. Both approaches, however, can result in performance bottlenecks or frequent mispredictions. LogQuant takes a different approach. By applying a log-based filtering mechanism, it selectively compresses the KV Cache across the entire context, achieving better performance with the same or even reduced memory footprint compared to existing methods. In benchmark tests, it enhances throughput by 25% and boosts batch size by 60% without increasing memory consumption. For challenging tasks such as Math and Code Completion, LogQuant improves accuracy by 40% to 200% at the same compression ratio, outperforming comparable techniques.LogQuant integrates effortlessly with popular inference frameworks like Python's transformers library. Implementation can be available in https://github.com/Concyclics/LogQuantKV.
comment: Accepted by ICLR 2025 Workshop on Sparsity in LLMs (SLLM)
☆ PAVE: Patching and Adapting Video Large Language Models CVPR2025
Pre-trained video large language models (Video LLMs) exhibit remarkable reasoning capabilities, yet adapting these models to new tasks involving additional modalities or data types (e.g., audio or 3D information) remains challenging. In this paper, we present PAVE, a flexible framework for adapting pre-trained Video LLMs to downstream tasks with side-channel signals, such as audio, 3D cues, or multi-view videos. PAVE introduces lightweight adapters, referred to as "patches," which add a small number of parameters and operations to a base model without modifying its architecture or pre-trained weights. In doing so, PAVE can effectively adapt the pre-trained base model to support diverse downstream tasks, including audio-visual question answering, 3D reasoning, multi-view video recognition, and high frame rate video understanding. Across these tasks, PAVE significantly enhances the performance of the base model, surpassing state-of-the-art task-specific models while incurring a minor cost of ~0.1% additional FLOPs and parameters. Further, PAVE supports multi-task learning and generalizes well across different Video LLMs. Our code is available at https://github.com/dragonlzm/PAVE.
comment: CVPR2025 Camera Ready
☆ PyGraph: Robust Compiler Support for CUDA Graphs in PyTorch
CUDA Graphs -- a recent hardware feature introduced for NVIDIA GPUs -- aim to reduce CPU launch overhead by capturing and launching a series of GPU tasks (kernels) as a DAG. However, deploying CUDA Graphs faces several challenges today due to the static structure of a graph. It also incurs performance overhead due to data copy. In fact, we show a counter-intuitive result -- deploying CUDA Graphs hurts performance in many cases. We introduce PyGraph, a novel approach to automatically harness the power of CUDA Graphs within PyTorch2. Driven by three key observations, PyGraph embodies three novel optimizations: it enables wider deployment of CUDA Graphs, reduces GPU kernel parameter copy overheads, and selectively deploys CUDA Graphs based on a cost-benefit analysis. PyGraph seamlessly integrates with PyTorch2's compilation toolchain, enabling efficient use of CUDA Graphs without manual modifications to the code. We evaluate PyGraph across various machine learning benchmarks, demonstrating substantial performance improvements over PyTorch2.
☆ LPOSS: Label Propagation Over Patches and Pixels for Open-vocabulary Semantic Segmentation
We propose a training-free method for open-vocabulary semantic segmentation using Vision-and-Language Models (VLMs). Our approach enhances the initial per-patch predictions of VLMs through label propagation, which jointly optimizes predictions by incorporating patch-to-patch relationships. Since VLMs are primarily optimized for cross-modal alignment and not for intra-modal similarity, we use a Vision Model (VM) that is observed to better capture these relationships. We address resolution limitations inherent to patch-based encoders by applying label propagation at the pixel level as a refinement step, significantly improving segmentation accuracy near class boundaries. Our method, called LPOSS+, performs inference over the entire image, avoiding window-based processing and thereby capturing contextual interactions across the full image. LPOSS+ achieves state-of-the-art performance among training-free methods, across a diverse set of datasets. Code: https://github.com/vladan-stojnic/LPOSS
☆ BiPrompt-SAM: Enhancing Image Segmentation via Explicit Selection between Point and Text Prompts
Segmentation is a fundamental task in computer vision, with prompt-driven methods gaining prominence due to their flexibility. The recent Segment Anything Model (SAM) has demonstrated powerful point-prompt segmentation capabilities, while text-based segmentation models offer rich semantic understanding. However, existing approaches rarely explore how to effectively combine these complementary modalities for optimal segmentation performance. This paper presents BiPrompt-SAM, a novel dual-modal prompt segmentation framework that fuses the advantages of point and text prompts through an explicit selection mechanism. Specifically, we leverage SAM's inherent ability to generate multiple mask candidates, combined with a semantic guidance mask from text prompts, and explicitly select the most suitable candidate based on similarity metrics. This approach can be viewed as a simplified Mixture of Experts (MoE) system, where the point and text modules act as distinct "experts," and the similarity scoring serves as a rudimentary "gating network." We conducted extensive evaluations on both the Endovis17 medical dataset and RefCOCO series natural image datasets. On Endovis17, BiPrompt-SAM achieved 89.55\% mDice and 81.46\% mIoU, comparable to state-of-the-art specialized medical segmentation models. On the RefCOCO series datasets, our method attained 87.1\%, 86.5\%, and 85.8\% IoU, significantly outperforming existing approaches. Experiments demonstrate that our explicit dual-selection method effectively combines the spatial precision of point prompts with the semantic richness of text prompts, particularly excelling in scenarios involving semantically complex objects, multiple similar objects, and partial occlusions. BiPrompt-SAM not only provides a simple yet effective implementation but also offers a new perspective on multi-modal prompt fusion.
☆ Automated Video-EEG Analysis in Epilepsy Studies: Advances and Challenges
Epilepsy is typically diagnosed through electroencephalography (EEG) and long-term video-EEG (vEEG) monitoring. The manual analysis of vEEG recordings is time-consuming, necessitating automated tools for seizure detection. Recent advancements in machine learning have shown promise in real-time seizure detection and prediction using EEG and video data. However, diversity of seizure symptoms, markup ambiguities, and limited availability of multimodal datasets hinder progress. This paper reviews the latest developments in automated video-EEG analysis and discusses the integration of multimodal data. We also propose a novel pipeline for treatment effect estimation from vEEG data using concept-based learning, offering a pathway for future research in this domain.
☆ Interpretable Deep Regression Models with Interval-Censored Failure Time Data
Deep neural networks (DNNs) have become powerful tools for modeling complex data structures through sequentially integrating simple functions in each hidden layer. In survival analysis, recent advances of DNNs primarily focus on enhancing model capabilities, especially in exploring nonlinear covariate effects under right censoring. However, deep learning methods for interval-censored data, where the unobservable failure time is only known to lie in an interval, remain underexplored and limited to specific data type or model. This work proposes a general regression framework for interval-censored data with a broad class of partially linear transformation models, where key covariate effects are modeled parametrically while nonlinear effects of nuisance multi-modal covariates are approximated via DNNs, balancing interpretability and flexibility. We employ sieve maximum likelihood estimation by leveraging monotone splines to approximate the cumulative baseline hazard function. To ensure reliable and tractable estimation, we develop an EM algorithm incorporating stochastic gradient descent. We establish the asymptotic properties of parameter estimators and show that the DNN estimator achieves minimax-optimal convergence. Extensive simulations demonstrate superior estimation and prediction accuracy over state-of-the-art methods. Applying our method to the Alzheimer's Disease Neuroimaging Initiative dataset yields novel insights and improved predictive performance compared to traditional approaches.
☆ How to RETIRE Tabular Data in Favor of Discrete Digital Signal Representation
The successes achieved by deep neural networks in computer vision tasks have led in recent years to the emergence of a new research area dubbed Multi-Dimensional Encoding (MDE). Methods belonging to this family aim to transform tabular data into a homogeneous form of discrete digital signals (images) to apply convolutional networks to initially unsuitable problems. Despite the successive emerging works, the pool of multi-dimensional encoding methods is still low, and the scope of research on existing modality encoding techniques is quite limited. To contribute to this area of research, we propose the Radar-based Encoding from Tabular to Image REpresentation (RETIRE), which allows tabular data to be represented as radar graphs, capturing the feature characteristics of each problem instance. RETIRE was compared with a pool of state-of-the-art MDE algorithms as well as with XGBoost in terms of classification accuracy and computational complexity. In addition, an analysis was carried out regarding transferability and explainability to provide more insight into both RETIRE and existing MDE techniques. The results obtained, supported by statistical analysis, confirm the superiority of RETIRE over other established MDE methods.
comment: 16 pages, 6 figures, 2 tables
☆ MindfulLIME: A Stable Solution for Explanations of Machine Learning Models with Enhanced Localization Precision -- A Medical Image Case Study
Ensuring transparency in machine learning decisions is critically important, especially in sensitive sectors such as healthcare, finance, and justice. Despite this, some popular explainable algorithms, such as Local Interpretable Model-agnostic Explanations (LIME), often produce unstable explanations due to the random generation of perturbed samples. Random perturbation introduces small changes or noise to modified instances of the original data, leading to inconsistent explanations. Even slight variations in the generated samples significantly affect the explanations provided by such models, undermining trust and hindering the adoption of interpretable models. To address this challenge, we propose MindfulLIME, a novel algorithm that intelligently generates purposive samples using a graph-based pruning algorithm and uncertainty sampling. MindfulLIME substantially improves the consistency of visual explanations compared to random sampling approaches. Our experimental evaluation, conducted on a widely recognized chest X-ray dataset, confirms MindfulLIME's stability with a 100% success rate in delivering reliable explanations under identical conditions. Additionally, MindfulLIME improves the localization precision of visual explanations by reducing the distance between the generated explanations and the actual local annotations compared to LIME. We also performed comprehensive experiments considering various segmentation algorithms and sample numbers, focusing on stability, quality, and efficiency. The results demonstrate the outstanding performance of MindfulLIME across different segmentation settings, generating fewer high-quality samples within a reasonable processing time. By addressing the stability limitations of LIME in image data, MindfulLIME enhances the trustworthiness and interpretability of machine learning models in specific medical imaging applications, a critical domain.
☆ On What Depends the Robustness of Multi-source Models to Missing Data in Earth Observation?
In recent years, the development of robust multi-source models has emerged in the Earth Observation (EO) field. These are models that leverage data from diverse sources to improve predictive accuracy when there is missing data. Despite these advancements, the factors influencing the varying effectiveness of such models remain poorly understood. In this study, we evaluate the predictive performance of six state-of-the-art multi-source models in predicting scenarios where either a single data source is missing or only a single source is available. Our analysis reveals that the efficacy of these models is intricately tied to the nature of the task, the complementarity among data sources, and the model design. Surprisingly, we observe instances where the removal of certain data sources leads to improved predictive performance, challenging the assumption that incorporating all available data is always beneficial. These findings prompt critical reflections on model complexity and the necessity of all collected data sources, potentially shaping the way for more streamlined approaches in EO applications.
comment: Accepted at IEEE International Geoscience and Remote Sensing Symposium 2025
☆ Invertible Koopman neural operator for data-driven modeling of partial differential equations
Koopman operator theory is a popular candidate for data-driven modeling because it provides a global linearization representation for nonlinear dynamical systems. However, existing Koopman operator-based methods suffer from shortcomings in constructing the well-behaved observable function and its inverse and are inefficient enough when dealing with partial differential equations (PDEs). To address these issues, this paper proposes the Invertible Koopman Neural Operator (IKNO), a novel data-driven modeling approach inspired by the Koopman operator theory and neural operator. IKNO leverages an Invertible Neural Network to parameterize observable function and its inverse simultaneously under the same learnable parameters, explicitly guaranteeing the reconstruction relation, thus eliminating the dependency on the reconstruction loss, which is an essential improvement over the original Koopman Neural Operator (KNO). The structured linear matrix inspired by the Koopman operator theory is parameterized to learn the evolution of observables' low-frequency modes in the frequency space rather than directly in the observable space, sustaining IKNO is resolution-invariant like other neural operators. Moreover, with preprocessing such as interpolation and dimension expansion, IKNO can be extended to operator learning tasks defined on non-Cartesian domains. We fully support the above claims based on rich numerical and real-world examples and demonstrate the effectiveness of IKNO and superiority over other neural operators.
comment: 25 pages, 10 figures
☆ Data-efficient rapid prediction of urban airflow and temperature fields for complex building geometries
Accurately predicting urban microclimate, including wind speed and temperature, based solely on building geometry requires capturing complex interactions between buildings and airflow, particularly long-range wake effects influenced by directional geometry. Traditional methods relying on computational fluid dynamics (CFD) are prohibitively expensive for large-scale simulations, while data-driven approaches struggle with limited training data and the need to model both local and far-field dependencies. In response, we propose a novel framework that leverages a multi-directional distance feature (MDDF) combined with localized training to achieve effective wind field predictions with minimal CFD data. By reducing the problem's dimensionality, localized training effectively increases the number of training samples, while MDDF encodes the surrounding geometric information to accurately model wake dynamics and flow redirection. Trained on only 24 CFD simulations, our localized Fourier neural operator (Local-FNO) model generates full 3D wind velocity and temperature predictions in under one minute, yielding a 500-fold speedup over conventional CFD methods. With mean absolute errors of 0.3 m/s for wind speed and 0.3 $^{\circ}$C for temperature on unseen urban configurations, our method demonstrates strong generalization capabilities and significant potential for practical urban applications.
☆ BiblioPage: A Dataset of Scanned Title Pages for Bibliographic Metadata Extraction ICDAR2025
Manual digitization of bibliographic metadata is time consuming and labor intensive, especially for historical and real-world archives with highly variable formatting across documents. Despite advances in machine learning, the absence of dedicated datasets for metadata extraction hinders automation. To address this gap, we introduce BiblioPage, a dataset of scanned title pages annotated with structured bibliographic metadata. The dataset consists of approximately 2,000 monograph title pages collected from 14 Czech libraries, spanning a wide range of publication periods, typographic styles, and layout structures. Each title page is annotated with 16 bibliographic attributes, including title, contributors, and publication metadata, along with precise positional information in the form of bounding boxes. To extract structured information from this dataset, we valuated object detection models such as YOLO and DETR combined with transformer-based OCR, achieving a maximum mAP of 52 and an F1 score of 59. Additionally, we assess the performance of various visual large language models, including LlamA 3.2-Vision and GPT-4o, with the best model reaching an F1 score of 67. BiblioPage serves as a real-world benchmark for bibliographic metadata extraction, contributing to document understanding, document question answering, and document information extraction. Dataset and evaluation scripts are availible at: https://github.com/DCGM/biblio-dataset
comment: Submitted to ICDAR2025 conference
☆ Towards Reliable Time Series Forecasting under Future Uncertainty: Ambiguity and Novelty Rejection Mechanisms
In real-world time series forecasting, uncertainty and lack of reliable evaluation pose significant challenges. Notably, forecasting errors often arise from underfitting in-distribution data and failing to handle out-of-distribution inputs. To enhance model reliability, we introduce a dual rejection mechanism combining ambiguity and novelty rejection. Ambiguity rejection, using prediction error variance, allows the model to abstain under low confidence, assessed through historical error variance analysis without future ground truth. Novelty rejection, employing Variational Autoencoders and Mahalanobis distance, detects deviations from training data. This dual approach improves forecasting reliability in dynamic environments by reducing errors and adapting to data changes, advancing reliability in complex scenarios.
☆ RGB-Th-Bench: A Dense benchmark for Visual-Thermal Understanding of Vision Language Models
We introduce RGB-Th-Bench, the first benchmark designed to evaluate the ability of Vision-Language Models (VLMs) to comprehend RGB-Thermal image pairs. While VLMs have demonstrated remarkable progress in visual reasoning and multimodal understanding, their evaluation has been predominantly limited to RGB-based benchmarks, leaving a critical gap in assessing their capabilities in infrared vision tasks. Existing visible-infrared datasets are either task-specific or lack high-quality annotations necessary for rigorous model evaluation. To address these limitations, RGB-Th-Bench provides a comprehensive evaluation framework covering 14 distinct skill dimensions, with a total of 1,600+ expert-annotated Yes/No questions. The benchmark employs two accuracy metrics: a standard question-level accuracy and a stricter skill-level accuracy, which evaluates model robustness across multiple questions within each skill dimension. This design ensures a thorough assessment of model performance, including resilience to adversarial and hallucinated responses. We conduct extensive evaluations on 19 state-of-the-art VLMs, revealing significant performance gaps in RGB-Thermal understanding. Our results show that even the strongest models struggle with thermal image comprehension, with performance heavily constrained by their RGB-based capabilities. Additionally, the lack of large-scale application-specific and expert-annotated thermal-caption-pair datasets in pre-training is an important reason of the observed performance gap. RGB-Th-Bench highlights the urgent need for further advancements in multimodal learning to bridge the gap between visible and thermal image understanding. The dataset is available through this link, and the evaluation code will also be made publicly available.
☆ Enhancing Graphical Lasso: A Robust Scheme for Non-Stationary Mean Data
This work addresses the problem of graph learning from data following a Gaussian Graphical Model (GGM) with a time-varying mean. Graphical Lasso (GL), the standard method for estimating sparse precision matrices, assumes that the observed data follows a zero-mean Gaussian distribution. However, this assumption is often violated in real-world scenarios where the mean evolves over time due to external influences, trends, or regime shifts. When the mean is not properly accounted for, applying GL directly can lead to estimating a biased precision matrix, hence hindering the graph learning task. To overcome this limitation, we propose Graphical Lasso with Adaptive Targeted Adaptive Importance Sampling (GL-ATAIS), an iterative method that jointly estimates the time-varying mean and the precision matrix. Our approach integrates Bayesian inference with frequentist estimation, leveraging importance sampling to obtain an estimate of the mean while using a regularized maximum likelihood estimator to infer the precision matrix. By iteratively refining both estimates, GL-ATAIS mitigates the bias introduced by time-varying means, leading to more accurate graph recovery. Our numerical evaluation demonstrates the impact of properly accounting for time-dependent means and highlights the advantages of GL-ATAIS over standard GL in recovering the true graph structure.
☆ An Efficient Data Reuse with Tile-Based Adaptive Stationary for Transformer Accelerators ISCA
Transformer-based models have become the \textit{de facto} backbone across many fields, such as computer vision and natural language processing. However, as these models scale in size, external memory access (EMA) for weight and activations becomes a critical bottleneck due to its significantly higher energy consumption compared to internal computations. While most prior work has focused on optimizing the self-attention mechanism, little attention has been given to optimizing data transfer during linear projections, where EMA costs are equally important. In this paper, we propose the Tile-based Adaptive Stationary (TAS) scheme that selects the input or weight stationary in a tile granularity, based on the input sequence length. Our experimental results demonstrate that TAS can significantly reduce EMA by more than 97\% compared to traditional stationary schemes, while being compatible with various attention optimization techniques and hardware accelerators.
comment: to be published in IEEE International Symposium on Circuits and Systems (IEEE ISCAS 2025)
☆ Kernel Learning Assisted Synthesis Condition Exploration for Ternary Spinel
Machine learning and high-throughput experimentation have greatly accelerated the discovery of mixed metal oxide catalysts by leveraging their compositional flexibility. However, the lack of established synthesis routes for solid-state materials remains a significant challenge in inorganic chemistry. An interpretable machine learning model is therefore essential, as it provides insights into the key factors governing phase formation. Here, we focus on the formation of single-phase Fe$_2$(ZnCo)O$_4$, synthesized via a high-throughput co-precipitation method. We combined a kernel classification model with a novel application of global SHAP analysis to pinpoint the experimental features most critical to single phase synthesizability by interpreting the contributions of each feature. Global SHAP analysis reveals that precursor and precipitating agent contributions to single-phase spinel formation align closely with established crystal growth theories. These results not only underscore the importance of interpretable machine learning in refining synthesis protocols but also establish a framework for data-informed experimental design in inorganic synthesis.
☆ Red Teaming with Artificial Intelligence-Driven Cyberattacks: A Scoping Review
The progress of artificial intelligence (AI) has made sophisticated methods available for cyberattacks and red team activities. These AI attacks can automate the process of penetrating a target or collecting sensitive data. The new methods can also accelerate the execution of the attacks. This review article examines the use of AI technologies in cybersecurity attacks. It also tries to describe typical targets for such attacks. We employed a scoping review methodology to analyze articles and identify AI methods, targets, and models that red teams can utilize to simulate cybercrime. From the 470 records screened, 11 were included in the review. Various cyberattack methods were identified, targeting sensitive data, systems, social media profiles, passwords, and URLs. The application of AI in cybercrime to develop versatile attack models presents an increasing threat. Furthermore, AI-based techniques in red team use can provide new ways to address these issues.
comment: An earlier version first published in Good Practices and New Perspectives in Information Systems and Technologies (pp. 129-138), 2024 by Springer Nature
☆ Optimization through In-Context Learning and Iterative LLM Prompting for Nuclear Engineering Design Problems
The optimization of nuclear engineering designs, such as nuclear fuel assembly configurations, involves managing competing objectives like reactivity control and power distribution. This study explores the use of Optimization by Prompting, an iterative approach utilizing large language models (LLMs), to address these challenges. The method is straightforward to implement, requiring no hyperparameter tuning or complex mathematical formulations. Optimization problems can be described in plain English, with only an evaluator and a parsing script needed for execution. The in-context learning capabilities of LLMs enable them to understand problem nuances, therefore, they have the potential to surpass traditional metaheuristic optimization methods. This study demonstrates the application of LLMs as optimizers to Boiling Water Reactor (BWR) fuel lattice design, showing the capability of commercial LLMs to achieve superior optimization results compared to traditional methods.
comment: Codes and data are available upon request
☆ Learning to chain-of-thought with Jensen's evidence lower bound
We propose a way to optimize chain-of-thought with reinforcement learning, but without external reward function. Our algorithm relies on viewing chain-of-thought as latent variable as part of a probabilistic inference problem. Contrary to the full evidence lower bound, we propose to apply a much simpler Jensen's lower bound, which derives tractable objectives with simple algorithmic components (e.g., without the need for parametric approximate posterior), making it more conducive to modern large-scale training. The lower bound approach naturally interpolates other methods such as supervised fine-tuning and online reinforcement learning, whose practical trade-offs we will illustrate. Finally, we show that on mathematical reasoning problems, optimizing with Jensen's lower bound is as effective as policy gradient with external reward. Taken together, our results showcase as a proof of concept to this new algorithmic paradigm's potential to more generic applications.
☆ RL-finetuning LLMs from on- and off-policy data with a single algorithm
We introduce a novel reinforcement learning algorithm (AGRO, for Any-Generation Reward Optimization) for fine-tuning large-language models. AGRO leverages the concept of generation consistency, which states that the optimal policy satisfies the notion of consistency across any possible generation of the model. We derive algorithms that find optimal solutions via the sample-based policy gradient and provide theoretical guarantees on their convergence. Our experiments demonstrate the effectiveness of AGRO in both on-policy and off-policy settings, showing improved performance on the mathematical reasoning dataset over baseline algorithms.
☆ Lean Formalization of Generalization Error Bound by Rademacher Complexity
We formalize the generalization error bound using Rademacher complexity in the Lean 4 theorem prover. Generalization error quantifies the gap between a learning machine's performance on given training data versus unseen test data, and Rademacher complexity serves as an estimate of this error based on the complexity of learning machines, or hypothesis class. Unlike traditional methods such as PAC learning and VC dimension, Rademacher complexity is applicable across diverse machine learning scenarios including deep learning and kernel methods. We formalize key concepts and theorems, including the empirical and population Rademacher complexities, and establish generalization error bounds through formal proofs of McDiarmid's inequality, Hoeffding's lemma, and symmetrization arguments.
☆ Optimizing Language Models for Inference Time Objectives using Reinforcement Learning
In this work, we investigate the merits of explicitly optimizing for inference time algorithmic performance during model training. We show how optimizing for inference time performance can improve overall model efficacy. We consider generic inference time objectives with $k$ samples, with a focus on pass@$k$ and majority voting as two main applications. With language model training on reasoning datasets, we showcase the performance trade-off enabled by training with such objectives. When training on code generation tasks, we show that the approach significantly improves pass@$k$ objectives compared to the baseline method.
☆ Boosting the Transferability of Audio Adversarial Examples with Acoustic Representation Optimization ICME 2025
With the widespread application of automatic speech recognition (ASR) systems, their vulnerability to adversarial attacks has been extensively studied. However, most existing adversarial examples are generated on specific individual models, resulting in a lack of transferability. In real-world scenarios, attackers often cannot access detailed information about the target model, making query-based attacks unfeasible. To address this challenge, we propose a technique called Acoustic Representation Optimization that aligns adversarial perturbations with low-level acoustic characteristics derived from speech representation models. Rather than relying on model-specific, higher-layer abstractions, our approach leverages fundamental acoustic representations that remain consistent across diverse ASR architectures. By enforcing an acoustic representation loss to guide perturbations toward these robust, lower-level representations, we enhance the cross-model transferability of adversarial examples without degrading audio quality. Our method is plug-and-play and can be integrated with any existing attack methods. We evaluate our approach on three modern ASR models, and the experimental results demonstrate that our method significantly improves the transferability of adversarial examples generated by previous methods while preserving the audio quality.
comment: Accepted to ICME 2025
☆ Post-Hoc Calibrated Anomaly Detection
Deep unsupervised anomaly detection has seen improvements in a supervised binary classification paradigm in which auxiliary external data is included in the training set as anomalous data in a process referred to as outlier exposure, which opens the possibility of exploring the efficacy of post-hoc calibration for anomaly detection and localization. Post-hoc Platt scaling and Beta calibration are found to improve results with gradient-based input perturbation, as well as post-hoc training with a strictly proper loss of a base model initially trained on an unsupervised loss. Post-hoc calibration is also found at times to be more effective using random synthesized spectral data as labeled anomalous data in the calibration set, suggesting that outlier exposure is superior only for initial training.
☆ SINR: Sparsity Driven Compressed Implicit Neural Representations
Implicit Neural Representations (INRs) are increasingly recognized as a versatile data modality for representing discretized signals, offering benefits such as infinite query resolution and reduced storage requirements. Existing signal compression approaches for INRs typically employ one of two strategies: 1. direct quantization with entropy coding of the trained INR; 2. deriving a latent code on top of the INR through a learnable transformation. Thus, their performance is heavily dependent on the quantization and entropy coding schemes employed. In this paper, we introduce SINR, an innovative compression algorithm that leverages the patterns in the vector spaces formed by weights of INRs. We compress these vector spaces using a high-dimensional sparse code within a dictionary. Further analysis reveals that the atoms of the dictionary used to generate the sparse code do not need to be learned or transmitted to successfully recover the INR weights. We demonstrate that the proposed approach can be integrated with any existing INR-based signal compression technique. Our results indicate that SINR achieves substantial reductions in storage requirements for INRs across various configurations, outperforming conventional INR-based compression baselines. Furthermore, SINR maintains high-quality decoding across diverse data modalities, including images, occupancy fields, and Neural Radiance Fields.
☆ FedMM-X: A Trustworthy and Interpretable Framework for Federated Multi-Modal Learning in Dynamic Environments
As artificial intelligence systems increasingly operate in Real-world environments, the integration of multi-modal data sources such as vision, language, and audio presents both unprecedented opportunities and critical challenges for achieving trustworthy intelligence. In this paper, we propose a novel framework that unifies federated learning with explainable multi-modal reasoning to ensure trustworthiness in decentralized, dynamic settings. Our approach, called FedMM-X (Federated Multi-Modal Explainable Intelligence), leverages cross-modal consistency checks, client-level interpretability mechanisms, and dynamic trust calibration to address challenges posed by data heterogeneity, modality imbalance, and out-of-distribution generalization. Through rigorous evaluation across federated multi-modal benchmarks involving vision-language tasks, we demonstrate improved performance in both accuracy and interpretability while reducing vulnerabilities to adversarial and spurious correlations. Further, we introduce a novel trust score aggregation method to quantify global model reliability under dynamic client participation. Our findings pave the way toward developing robust, interpretable, and socially responsible AI systems in Real-world environments.
☆ Causal Bayesian Optimization with Unknown Graphs
Causal Bayesian Optimization (CBO) is a methodology designed to optimize an outcome variable by leveraging known causal relationships through targeted interventions. Traditional CBO methods require a fully and accurately specified causal graph, which is a limitation in many real-world scenarios where such graphs are unknown. To address this, we propose a new method for the CBO framework that operates without prior knowledge of the causal graph. Consistent with causal bandit theory, we demonstrate through theoretical analysis and that focusing on the direct causal parents of the target variable is sufficient for optimization, and provide empirical validation in the context of CBO. Furthermore we introduce a new method that learns a Bayesian posterior over the direct parents of the target variable. This allows us to optimize the outcome variable while simultaneously learning the causal structure. Our contributions include a derivation of the closed-form posterior distribution for the linear case. In the nonlinear case where the posterior is not tractable, we present a Gaussian Process (GP) approximation that still enables CBO by inferring the parents of the outcome variable. The proposed method performs competitively with existing benchmarks and scales well to larger graphs, making it a practical tool for real-world applications where causal information is incomplete.
☆ Noise Resilient Over-The-Air Federated Learning In Heterogeneous Wireless Networks
In 6G wireless networks, Artificial Intelligence (AI)-driven applications demand the adoption of Federated Learning (FL) to enable efficient and privacy-preserving model training across distributed devices. Over-The-Air Federated Learning (OTA-FL) exploits the superposition property of multiple access channels, allowing edge users in 6G networks to efficiently share spectral resources and perform low-latency global model aggregation. However, these advantages come with challenges, as traditional OTA-FL techniques suffer due to the joint effects of Additive White Gaussian Noise (AWGN) at the server, fading, and both data and system heterogeneity at the participating edge devices. In this work, we propose the novel Noise Resilient Over-the-Air Federated Learning (NoROTA-FL) framework to jointly tackle these challenges in federated wireless networks. In NoROTA-FL, the local optimization problems find controlled inexact solutions, which manifests as an additional proximal constraint at the clients. This approach provides robustness against straggler-induced partial work, heterogeneity, noise, and fading. From a theoretical perspective, we leverage the zeroth- and first-order inexactness and establish convergence guarantees for non-convex optimization problems in the presence of heterogeneous data and varying system capabilities. Experimentally, we validate NoROTA-FL on real-world datasets, including FEMNIST, CIFAR10, and CIFAR100, demonstrating its robustness in noisy and heterogeneous environments. Compared to state-of-the-art baselines such as COTAF and FedProx, NoROTA-FL achieves significantly more stable convergence and higher accuracy, particularly in the presence of stragglers.
☆ VectorFit : Adaptive Singular & Bias Vector Fine-Tuning of Pre-trained Foundation Models
Popular PEFT methods achieve parameter efficiency by assuming that incremental weight updates are inherently low-rank, which often leads to a performance gap compared to full fine-tuning. While recent methods have attempted to address this limitation, they typically lack sufficient parameter and memory efficiency. We propose VectorFit, an effective and easily deployable approach that adaptively trains the singular vectors and biases of pre-trained weight matrices. We demonstrate that the utilization of structural and transformational characteristics of pre-trained weights enables high-rank updates comparable to those of full fine-tuning. As a result, VectorFit achieves superior performance with 9X less trainable parameters compared to state-of-the-art PEFT methods. Through extensive experiments over 17 datasets spanning diverse language and vision tasks such as natural language understanding and generation, question answering, image classification, and image generation, we exhibit that VectorFit consistently outperforms baselines, even in extremely low-budget scenarios.
☆ A Spatiotemporal Radar-Based Precipitation Model for Water Level Prediction and Flood Forecasting
Study Region: Goslar and G\"ottingen, Lower Saxony, Germany. Study Focus: In July 2017, the cities of Goslar and G\"ottingen experienced severe flood events characterized by short warning time of only 20 minutes, resulting in extensive regional flooding and significant damage. This highlights the critical need for a more reliable and timely flood forecasting system. This paper presents a comprehensive study on the impact of radar-based precipitation data on forecasting river water levels in Goslar. Additionally, the study examines how precipitation influences water level forecasts in G\"ottingen. The analysis integrates radar-derived spatiotemporal precipitation patterns with hydrological sensor data obtained from ground stations to evaluate the effectiveness of this approach in improving flood prediction capabilities. New Hydrological Insights for the Region: A key innovation in this paper is the use of residual-based modeling to address the non-linearity between precipitation images and water levels, leading to a Spatiotemporal Radar-based Precipitation Model with residuals (STRPMr). Unlike traditional hydrological models, our approach does not rely on upstream data, making it independent of additional hydrological inputs. This independence enhances its adaptability and allows for broader applicability in other regions with RADOLAN precipitation. The deep learning architecture integrates (2+1)D convolutional neural networks for spatial and temporal feature extraction with LSTM for timeseries forecasting. The results demonstrate the potential of the STRPMr for capturing extreme events and more accurate flood forecasting.
comment: 28 pages, 11 figures, 6 tables
☆ DataPlatter: Boosting Robotic Manipulation Generalization with Minimal Costly Data
The growing adoption of Vision-Language-Action (VLA) models in embodied AI intensifies the demand for diverse manipulation demonstrations. However, high costs associated with data collection often result in insufficient data coverage across all scenarios, which limits the performance of the models. It is observed that the spatial reasoning phase (SRP) in large workspace dominates the failure cases. Fortunately, this data can be collected with low cost, underscoring the potential of leveraging inexpensive data to improve model performance. In this paper, we introduce the DataPlatter method, a framework that decouples training trajectories into distinct task stages and leverages abundant easily collectible SRP data to enhance VLA model's generalization. Through analysis we demonstrate that sub-task-specific training with additional SRP data with proper proportion can act as a performance catalyst for robot manipulation, maximizing the utilization of costly physical interaction phase (PIP) data. Experiments show that through introducing large proportion of cost-effective SRP trajectories into a limited set of PIP data, we can achieve a maximum improvement of 41\% on success rate in zero-shot scenes, while with the ability to transfer manipulation skill to novel targets.
☆ Improved Alignment of Modalities in Large Vision Language Models
Recent advancements in vision-language models have achieved remarkable results in making language models understand vision inputs. However, a unified approach to align these models across diverse tasks such as image captioning and visual question answering remains a challenge. Existing methods either require very big language models or very big datasets which is not efficient in utilizing existing models. This paper addresses this gap and devises a training strategy of auto-regressive vision-language models, to unify vision-language tasks like image-captioning and visual question answering. We propose four training stages for aligning the vision model with the language model, in other words, the language model is given an ability to process visual inputs. We also devise different attention masks for training transformer-based language models that improve the quality of visual features. Further, we introduce some findings, 1) the attention mask should not be applied on visual inputs, 2) the Language model converges faster on AI- generated data, 3) More work should be done in the alignment stage during the pre-training of the model, 4) the model can easily adapt to any downstream tasks like visual question answering on healthcare datasets like PathVQA. After training the model for one epoch for all the stages, it outperforms large models like VILA-13 billion models on common benchmarks like CIDEr scores on COCO and Flickr30k datasets and achieves very close scores to GIT-2 on the same dataset despite being a much smaller model trained on a much smaller dataset. All of the training is done using best practices available like multi- GPU parallel training, lower-precision training with 16-bit float numbers, faster attention (SDPA), and gradient accumulation, and completed the training within 12 hours.
☆ A stochastic gradient descent algorithm with random search directions
Stochastic coordinate descent algorithms are efficient methods in which each iterate is obtained by fixing most coordinates at their values from the current iteration, and approximately minimizing the objective with respect to the remaining coordinates. However, this approach is usually restricted to canonical basis vectors of $\mathbb{R}^d$. In this paper, we develop a new class of stochastic gradient descent algorithms with random search directions which uses the directional derivative of the gradient estimate following more general random vectors. We establish the almost sure convergence of these algorithms with decreasing step. We further investigate their central limit theorem and pay particular attention to analyze the impact of the search distributions on the asymptotic covariance matrix. We also provide the non-asymptotic $\mathbb{L}^p$ rates of convergence.
☆ SMT-EX: An Explainable Surrogate Modeling Toolbox for Mixed-Variables Design Exploration
Surrogate models are of high interest for many engineering applications, serving as cheap-to-evaluate time-efficient approximations of black-box functions to help engineers and practitioners make decisions and understand complex systems. As such, the need for explainability methods is rising and many studies have been performed to facilitate knowledge discovery from surrogate models. To respond to these enquiries, this paper introduces SMT-EX, an enhancement of the open-source Python Surrogate Modeling Toolbox (SMT) that integrates explainability techniques into a state-of-the-art surrogate modelling framework. More precisely, SMT-EX includes three key explainability methods: Shapley Additive Explanations, Partial Dependence Plot, and Individual Conditional Expectations. A peculiar explainability dependency of SMT has been developed for such purpose that can be easily activated once the surrogate model is built, offering a user-friendly and efficient tool for swift insight extraction. The effectiveness of SMT-EX is showcased through two test cases. The first case is a 10-variable wing weight problem with purely continuous variables and the second one is a 3-variable mixed-categorical cantilever beam bending problem. Relying on SMT-EX analyses for these problems, we demonstrate its versatility in addressing a diverse range of problem characteristics. SMT-Explainability is freely available on Github: https://github.com/SMTorg/smt-explainability .
☆ Bayesian Optimization of a Lightweight and Accurate Neural Network for Aerodynamic Performance Prediction
Ensuring high accuracy and efficiency of predictive models is paramount in the aerospace industry, particularly in the context of multidisciplinary design and optimization processes. These processes often require numerous evaluations of complex objective functions, which can be computationally expensive and time-consuming. To build efficient and accurate predictive models, we propose a new approach that leverages Bayesian Optimization (BO) to optimize the hyper-parameters of a lightweight and accurate Neural Network (NN) for aerodynamic performance prediction. To clearly describe the interplay between design variables, hierarchical and categorical kernels are used in the BO formulation. We demonstrate the efficiency of our approach through two comprehensive case studies, where the optimized NN significantly outperforms baseline models and other publicly available NNs in terms of accuracy and parameter efficiency. For the drag coefficient prediction task, the Mean Absolute Percentage Error (MAPE) of our optimized model drops from 0.1433\% to 0.0163\%, which is nearly an order of magnitude improvement over the baseline model. Additionally, our model achieves a MAPE of 0.82\% on a benchmark aircraft self-noise prediction problem, significantly outperforming existing models (where their MAPE values are around 2 to 3\%) while requiring less computational resources. The results highlight the potential of our framework to enhance the scalability and performance of NNs in large-scale MDO problems, offering a promising solution for the aerospace industry.
☆ Extracting Interpretable Logic Rules from Graph Neural Networks
Graph neural networks (GNNs) operate over both input feature spaces and combinatorial graph structures, making it challenging to understand the rationale behind their predictions. As GNNs gain widespread popularity and demonstrate success across various domains, such as drug discovery, studying their interpretability has become a critical task. To address this, many explainability methods have been proposed, with recent efforts shifting from instance-specific explanations to global concept-based explainability. However, these approaches face several limitations, such as relying on predefined concepts and explaining only a limited set of patterns. To address this, we propose a novel framework, LOGICXGNN, for extracting interpretable logic rules from GNNs. LOGICXGNN is model-agnostic, efficient, and data-driven, eliminating the need for predefined concepts. More importantly, it can serve as a rule-based classifier and even outperform the original neural models. Its interpretability facilitates knowledge discovery, as demonstrated by its ability to extract detailed and accurate chemistry knowledge that is often overlooked by existing methods. Another key advantage of LOGICXGNN is its ability to generate new graph instances in a controlled and transparent manner, offering significant potential for applications such as drug design. We empirically demonstrate these merits through experiments on real-world datasets such as MUTAG and BBBP.
comment: 12 pages, 4 figures
☆ A Probabilistic Neuro-symbolic Layer for Algebraic Constraint Satisfaction
In safety-critical applications, guaranteeing the satisfaction of constraints over continuous environments is crucial, e.g., an autonomous agent should never crash into obstacles or go off-road. Neural models struggle in the presence of these constraints, especially when they involve intricate algebraic relationships. To address this, we introduce a differentiable probabilistic layer that guarantees the satisfaction of non-convex algebraic constraints over continuous variables. This probabilistic algebraic layer (PAL) can be seamlessly plugged into any neural architecture and trained via maximum likelihood without requiring approximations. PAL defines a distribution over conjunctions and disjunctions of linear inequalities, parameterized by polynomials. This formulation enables efficient and exact renormalization via symbolic integration, which can be amortized across different data points and easily parallelized on a GPU. We showcase PAL and our integration scheme on a number of benchmarks for algebraic constraint integration and on real-world trajectory data.
☆ Data-centric Federated Graph Learning with Large Language Models
In federated graph learning (FGL), a complete graph is divided into multiple subgraphs stored in each client due to privacy concerns, and all clients jointly train a global graph model by only transmitting model parameters. A pain point of FGL is the heterogeneity problem, where nodes or structures present non-IID properties among clients (e.g., different node label distributions), dramatically undermining the convergence and performance of FGL. To address this, existing efforts focus on design strategies at the model level, i.e., they design models to extract common knowledge to mitigate heterogeneity. However, these model-level strategies fail to fundamentally address the heterogeneity problem as the model needs to be designed from scratch when transferring to other tasks. Motivated by large language models (LLMs) having achieved remarkable success, we aim to utilize LLMs to fully understand and augment local text-attributed graphs, to address data heterogeneity at the data level. In this paper, we propose a general framework LLM4FGL that innovatively decomposes the task of LLM for FGL into two sub-tasks theoretically. Specifically, for each client, it first utilizes the LLM to generate missing neighbors and then infers connections between generated nodes and raw nodes. To improve the quality of generated nodes, we design a novel federated generation-and-reflection mechanism for LLMs, without the need to modify the parameters of the LLM but relying solely on the collective feedback from all clients. After neighbor generation, all the clients utilize a pre-trained edge predictor to infer the missing edges. Furthermore, our framework can seamlessly integrate as a plug-in with existing FGL methods. Experiments on three real-world datasets demonstrate the superiority of our method compared to advanced baselines.
comment: ongoing work
☆ VecTrans: LLM Transformation Framework for Better Auto-vectorization on High-performance CPU
Large language models (LLMs) have demonstrated great capabilities in code generation, yet their effective application in compiler optimizations remains an open challenge due to issues such as hallucinations and a lack of domain-specific reasoning. Vectorization, a crucial optimization for enhancing code performance, often fails because of the compiler's inability to recognize complex code patterns, which commonly require extensive empirical expertise. LLMs, with their ability to capture intricate patterns, thus providing a promising solution to this challenge. This paper presents VecTrans, a novel framework that leverages LLMs to enhance compiler-based code vectorization. VecTrans first employs compiler analysis to identify potentially vectorizable code regions. It then utilizes an LLM to refactor these regions into patterns that are more amenable to the compiler's auto-vectorization. To ensure semantic correctness, VecTrans further integrates a hybrid validation mechanism at the intermediate representation (IR) level. With the above efforts, VecTrans combines the adaptability of LLMs with the precision of compiler vectorization, thereby effectively opening up the vectorization opportunities. Experimental results show that among all 50 TSVC functions unvectorizable by Clang, GCC, and BiShengCompiler, VecTrans successfully vectorizes 23 cases (46%) and achieves an average speedup of 2.02x, greatly surpassing state-of-the-art performance.
☆ FuXi-RTM: A Physics-Guided Prediction Framework with Radiative Transfer Modeling
Similar to conventional video generation, current deep learning-based weather prediction frameworks often lack explicit physical constraints, leading to unphysical outputs that limit their reliability for operational forecasting. Among various physical processes requiring proper representation, radiation plays a fundamental role as it drives Earth's weather and climate systems. However, accurate simulation of radiative transfer processes remains challenging for traditional numerical weather prediction (NWP) models due to their inherent complexity and high computational costs. Here, we propose FuXi-RTM, a hybrid physics-guided deep learning framework designed to enhance weather forecast accuracy while enforcing physical consistency. FuXi-RTM integrates a primary forecasting model (FuXi) with a fixed deep learning-based radiative transfer model (DLRTM) surrogate that efficiently replaces conventional radiation parameterization schemes. This represents the first deep learning-based weather forecasting framework to explicitly incorporate physical process modeling. Evaluated over a comprehensive 5-year dataset, FuXi-RTM outperforms its unconstrained counterpart in 88.51% of 3320 variable and lead time combinations, with improvements in radiative flux predictions. By incorporating additional physical processes, FuXi-RTM paves the way for next-generation weather forecasting systems that are both accurate and physically consistent.
☆ Quantifying the Ease of Reproducing Training Data in Unconditional Diffusion Models
Diffusion models, which have been advancing rapidly in recent years, may generate samples that closely resemble the training data. This phenomenon, known as memorization, may lead to copyright issues. In this study, we propose a method to quantify the ease of reproducing training data in unconditional diffusion models. The average of a sample population following the Langevin equation in the reverse diffusion process moves according to a first-order ordinary differential equation (ODE). This ODE establishes a 1-to-1 correspondence between images and their noisy counterparts in the latent space. Since the ODE is reversible and the initial noisy images are sampled randomly, the volume of an image's projected area represents the probability of generating those images. We examined the ODE, which projects images to latent space, and succeeded in quantifying the ease of reproducing training data by measuring the volume growth rate in this process. Given the relatively low computational complexity of this method, it allows us to enhance the quality of training data by detecting and modifying the easily memorized training samples.
☆ A novel forecasting framework combining virtual samples and enhanced Transformer models for tourism demand forecasting
Accurate tourism demand forecasting is hindered by limited historical data and complex spatiotemporal dependencies among tourist origins. A novel forecasting framework integrating virtual sample generation and a novel Transformer predictor addresses constraints arising from restricted data availability. A spatiotemporal GAN produces realistic virtual samples by dynamically modeling spatial correlations through a graph convolutional network, and an enhanced Transformer captures local patterns with causal convolutions and long-term dependencies with self-attention,eliminating autoregressive decoding. A joint training strategy refines virtual sample generation based on predictor feedback to maintain robust performance under data-scarce conditions. Experimental evaluations on real-world daily and monthly tourism demand datasets indicate a reduction in average MASE by 18.37% compared to conventional Transformer-based models, demonstrating improved forecasting accuracy. The integration of adaptive spatiotemporal sample augmentation with a specialized Transformer can effectively address limited-data forecasting scenarios in tourism management.
☆ Multi-Agent Deep Reinforcement Learning for Safe Autonomous Driving with RICS-Assisted MEC
Environment sensing and fusion via onboard sensors are envisioned to be widely applied in future autonomous driving networks. This paper considers a vehicular system with multiple self-driving vehicles that is assisted by multi-access edge computing (MEC), where image data collected by the sensors is offloaded from cellular vehicles to the MEC server using vehicle-to-infrastructure (V2I) links. Sensory data can also be shared among surrounding vehicles via vehicle-to-vehicle (V2V) communication links. To improve spectrum utilization, the V2V links may reuse the same frequency spectrum with V2I links, which may cause severe interference. To tackle this issue, we leverage reconfigurable intelligent computational surfaces (RICSs) to jointly enable V2I reflective links and mitigate interference appearing at the V2V links. Considering the limitations of traditional algorithms in addressing this problem, such as the assumption for quasi-static channel state information, which restricts their ability to adapt to dynamic environmental changes and leads to poor performance under frequently varying channel conditions, in this paper, we formulate the problem at hand as a Markov game. Our novel formulation is applied to time-varying channels subject to multi-user interference and introduces a collaborative learning mechanism among users. The considered optimization problem is solved via a driving safety-enabled multi-agent deep reinforcement learning (DS-MADRL) approach that capitalizes on the RICS presence. Our extensive numerical investigations showcase that the proposed reinforcement learning approach achieves faster convergence and significant enhancements in both data rate and driving safety, as compared to various state-of-the-art benchmarks.
☆ Continual Learning With Quasi-Newton Methods
Catastrophic forgetting remains a major challenge when neural networks learn tasks sequentially. Elastic Weight Consolidation (EWC) attempts to address this problem by introducing a Bayesian-inspired regularization loss to preserve knowledge of previously learned tasks. However, EWC relies on a Laplace approximation where the Hessian is simplified to the diagonal of the Fisher information matrix, assuming uncorrelated model parameters. This overly simplistic assumption often leads to poor Hessian estimates, limiting its effectiveness. To overcome this limitation, we introduce Continual Learning with Sampled Quasi-Newton (CSQN), which leverages Quasi-Newton methods to compute more accurate Hessian approximations. CSQN captures parameter interactions beyond the diagonal without requiring architecture-specific modifications, making it applicable across diverse tasks and architectures. Experimental results across four benchmarks demonstrate that CSQN consistently outperforms EWC and other state-of-the-art baselines, including rehearsal-based methods. CSQN reduces EWC's forgetting by 50 percent and improves its performance by 8 percent on average. Notably, CSQN achieves superior results on three out of four benchmarks, including the most challenging scenarios, highlighting its potential as a robust solution for continual learning.
comment: Published in IEEE Access
☆ Quantifying Symptom Causality in Clinical Decision Making: An Exploration Using CausaLM
Current machine learning approaches to medical diagnosis often rely on correlational patterns between symptoms and diseases, risking misdiagnoses when symptoms are ambiguous or common across multiple conditions. In this work, we move beyond correlation to investigate the causal influence of key symptoms-specifically "chest pain" on diagnostic predictions. Leveraging the CausaLM framework, we generate counterfactual text representations in which target concepts are effectively "forgotten" enabling a principled estimation of the causal effect of that concept on a model's predicted disease distribution. By employing Textual Representation-based Average Treatment Effect (TReATE), we quantify how the presence or absence of a symptom shapes the model's diagnostic outcomes, and contrast these findings against correlation-based baselines such as CONEXP. Our results offer deeper insight into the decision-making behavior of clinical NLP models and have the potential to inform more trustworthy, interpretable, and causally-grounded decision support tools in medical practice.
☆ Causal invariant geographic network representations with feature and structural distribution shifts
The existing methods learn geographic network representations through deep graph neural networks (GNNs) based on the i.i.d. assumption. However, the spatial heterogeneity and temporal dynamics of geographic data make the out-of-distribution (OOD) generalisation problem particularly salient. The latter are particularly sensitive to distribution shifts (feature and structural shifts) between testing and training data and are the main causes of the OOD generalisation problem. Spurious correlations are present between invariant and background representations due to selection biases and environmental effects, resulting in the model extremes being more likely to learn background representations. The existing approaches focus on background representation changes that are determined by shifts in the feature distributions of nodes in the training and test data while ignoring changes in the proportional distributions of heterogeneous and homogeneous neighbour nodes, which we refer to as structural distribution shifts. We propose a feature-structure mixed invariant representation learning (FSM-IRL) model that accounts for both feature distribution shifts and structural distribution shifts. To address structural distribution shifts, we introduce a sampling method based on causal attention, encouraging the model to identify nodes possessing strong causal relationships with labels or nodes that are more similar to the target node. Inspired by the Hilbert-Schmidt independence criterion, we implement a reweighting strategy to maximise the orthogonality of the node representations, thereby mitigating the spurious correlations among the node representations and suppressing the learning of background representations. Our experiments demonstrate that FSM-IRL exhibits strong learning capabilities on both geographic and social network datasets in OOD scenarios.
comment: 15 pages, 3 figures, 8 tables
☆ Towards Build Optimization Using Digital Twins
Despite the indisputable benefits of Continuous Integration (CI) pipelines (or builds), CI still presents significant challenges regarding long durations, failures, and flakiness. Prior studies addressed CI challenges in isolation, yet these issues are interrelated and require a holistic approach for effective optimization. To bridge this gap, this paper proposes a novel idea of developing Digital Twins (DTs) of build processes to enable global and continuous improvement. To support such an idea, we introduce the CI Build process Digital Twin (CBDT) framework as a minimum viable product. This framework offers digital shadowing functionalities, including real-time build data acquisition and continuous monitoring of build process performance metrics. Furthermore, we discuss guidelines and challenges in the practical implementation of CBDTs, including (1) modeling different aspects of the build process using Machine Learning, (2) exploring what-if scenarios based on historical patterns, and (3) implementing prescriptive services such as automated failure and performance repair to continuously improve build processes.
comment: Accepted at the 21st International Conference on Predictive Models and Data Analytics in Software Engineering - PROMISE 2025
☆ Social Network User Profiling for Anomaly Detection Based on Graph Neural Networks
This study proposes a risk pricing anomaly detection method for social network user portraits based on graph neural networks (GNNs), aiming to improve the ability to identify abnormal users in social network environments. In view of the limitations of traditional methods in social network data modeling, this paper combines graph autoencoders (GAEs) and graph attention networks (GATs) to achieve accurate detection of abnormal users through dynamic aggregation of neighbor features and reconstruction error evaluation. The Facebook Page-Page Network dataset is used in the experiment and compared with VAE, GNN, Transformer and GAE. The results show that the proposed method achieves the best performance in AUC, F1-score, Precision and Recall, verifying its effectiveness. In addition, this paper explores the computational efficiency of the model in large-scale data and looks forward to combining self-supervised learning, federated learning, and other technologies in the future to improve the robustness and privacy protection of risk assessment. The research results can provide efficient anomaly detection solutions for financial risk control, social security management, and other fields.
☆ Interpretable Generative Models through Post-hoc Concept Bottlenecks CVPR 2025
Concept bottleneck models (CBM) aim to produce inherently interpretable models that rely on human-understandable concepts for their predictions. However, existing approaches to design interpretable generative models based on CBMs are not yet efficient and scalable, as they require expensive generative model training from scratch as well as real images with labor-intensive concept supervision. To address these challenges, we present two novel and low-cost methods to build interpretable generative models through post-hoc techniques and we name our approaches: concept-bottleneck autoencoder (CB-AE) and concept controller (CC). Our proposed approaches enable efficient and scalable training without the need of real data and require only minimal to no concept supervision. Additionally, our methods generalize across modern generative model families including generative adversarial networks and diffusion models. We demonstrate the superior interpretability and steerability of our methods on numerous standard datasets like CelebA, CelebA-HQ, and CUB with large improvements (average ~25%) over the prior work, while being 4-15x faster to train. Finally, a large-scale user study is performed to validate the interpretability and steerability of our methods.
comment: CVPR 2025. Project Page: https://lilywenglab.github.io/posthoc-generative-cbm/
☆ Flow to Learn: Flow Matching on Neural Network Parameters ICLR
Foundational language models show a remarkable ability to learn new concepts during inference via context data. However, similar work for images lag behind. To address this challenge, we introduce FLoWN, a flow matching model that learns to generate neural network parameters for different tasks. Our approach models the flow on latent space, while conditioning the process on context data. Experiments verify that FLoWN attains various desiderata for a meta-learning model. In addition, it matches or exceeds baselines on in-distribution tasks, provides better initializations for classifier training, and is performant on out-of-distribution few-shot tasks while having a fine-tuning mechanism to improve performance.
comment: Accepted at the ICLR Workshop on Neural Network Weights as a New Data Modality 2025
☆ Data-driven Mesoscale Weather Forecasting Combining Swin-Unet and Diffusion Models
Data-driven weather prediction models exhibit promising performance and advance continuously. In particular, diffusion models represent fine-scale details without spatial smoothing, which is crucial for mesoscale predictions, such as heavy rainfall forecasting. However, the applications of diffusion models to mesoscale prediction remain limited. To address this gap, this study proposes an architecture that combines a diffusion model with Swin-Unet as a deterministic model, achieving mesoscale predictions while maintaining flexibility. The proposed architecture trains the two models independently, allowing the diffusion model to remain unchanged when the deterministic model is updated. Comparisons using the Fractions Skill Score and power spectral analysis demonstrate that incorporating the diffusion model leads to improved accuracy compared to predictions without it. These findings underscore the potential of the proposed architecture to enhance mesoscale predictions, particularly for strong rainfall events, while maintaining flexibility.
☆ QUAD: Quantization and Parameter-Efficient Tuning of LLM with Activation Decomposition
Large Language Models (LLMs) excel in diverse applications but suffer inefficiency due to massive scale. While quantization reduces computational costs, existing methods degrade accuracy in medium-sized LLMs (e.g., Llama-3-8B) due to activation outliers. To address this, we propose QUAD (Quantization with Activation Decomposition), a framework leveraging Singular Value Decomposition (SVD) to suppress activation outliers for effective 4-bit quantization. QUAD estimates activation singular vectors offline using calibration data to construct an orthogonal transformation matrix P, shifting outliers to additional dimensions in full precision while quantizing rest components to 4-bit. Additionally, QUAD enables parameter-efficient fine-tuning via adaptable full-precision outlier weights, narrowing the accuracy gap between quantized and full-precision models. Experiments demonstrate that QUAD achieves 94% ~ 96% accuracy under W4A4 quantization and 98% accuracy with W4A4/A8 and parameter-efficient fine-tuning for Llama-3 and Qwen-2.5 models. Our code is available at \href{https://github.com/hyx1999/Quad}{repository}.
comment: 18 pages, 8 figures, 8 tables
☆ Optimal Parameter Adaptation for Safety-Critical Control via Safe Barrier Bayesian Optimization
Safety is of paramount importance in control systems to avoid costly risks and catastrophic damages. The control barrier function (CBF) method, a promising solution for safety-critical control, poses a new challenge of enhancing control performance due to its direct modification of original control design and the introduction of uncalibrated parameters. In this work, we shed light on the crucial role of configurable parameters in the CBF method for performance enhancement with a systematical categorization. Based on that, we propose a novel framework combining the CBF method with Bayesian optimization (BO) to optimize the safe control performance. Considering feasibility/safety-critical constraints, we develop a safe version of BO using the barrier-based interior method to efficiently search for promising feasible configurable parameters. Furthermore, we provide theoretical criteria of our framework regarding safety and optimality. An essential advantage of our framework lies in that it can work in model-agnostic environments, leaving sufficient flexibility in designing objective and constraint functions. Finally, simulation experiments on swing-up control and high-fidelity adaptive cruise control are conducted to demonstrate the effectiveness of our framework.
comment: Preprent manuscript, review only
☆ Stop Walking in Circles! Bailing Out Early in Projected Gradient Descent CVPR
Projected Gradient Descent (PGD) under the $L_\infty$ ball has become one of the defacto methods used in adversarial robustness evaluation for computer vision (CV) due to its reliability and efficacy, making a strong and easy-to-implement iterative baseline. However, PGD is computationally demanding to apply, especially when using thousands of iterations is the current best-practice recommendation to generate an adversarial example for a single image. In this work, we introduce a simple novel method for early termination of PGD based on cycle detection by exploiting the geometry of how PGD is implemented in practice and show that it can produce large speedup factors while providing the \emph{exact} same estimate of model robustness as standard PGD. This method substantially speeds up PGD without sacrificing any attack strength, enabling evaluations of robustness that were previously computationally intractable.
comment: To appear in the 2025 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
♻ ☆ Aether: Geometric-Aware Unified World Modeling
The integration of geometric reconstruction and generative modeling remains a critical challenge in developing AI systems capable of human-like spatial reasoning. This paper proposes Aether, a unified framework that enables geometry-aware reasoning in world models by jointly optimizing three core capabilities: (1) 4D dynamic reconstruction, (2) action-conditioned video prediction, and (3) goal-conditioned visual planning. Through task-interleaved feature learning, Aether achieves synergistic knowledge sharing across reconstruction, prediction, and planning objectives. Building upon video generation models, our framework demonstrates unprecedented synthetic-to-real generalization despite never observing real-world data during training. Furthermore, our approach achieves zero-shot generalization in both action following and reconstruction tasks, thanks to its intrinsic geometric modeling. Remarkably, even without real-world data, its reconstruction performance is comparable with or even better than that of domain-specific models. Additionally, Aether employs camera trajectories as geometry-informed action spaces, enabling effective action-conditioned prediction and visual planning. We hope our work inspires the community to explore new frontiers in physically-reasonable world modeling and its applications.
comment: Project Page: https://aether-world.github.io/
♻ ☆ TARDIS: Mitigating Temporal Misalignment via Representation Steering
Language models often struggle with temporal misalignment, performance degradation caused by shifts in the temporal distribution of data. Continuously updating models to avoid degradation is expensive. Can models be adapted without updating model weights? We present TARDIS, an unsupervised representation editing method that addresses this challenge. TARDIS extracts steering vectors from unlabeled data and adjusts the model's representations to better align with the target time period's distribution. Our experiments reveal that TARDIS enhances downstream task performance without the need for fine-tuning, can mitigate temporal misalignment even when exact target time period data is unavailable, and remains efficient even when the temporal information of the target data points is unknown at inference time.
♻ ☆ Geometric Preference Elicitation for Minimax Regret Optimization in Uncertainty Matroids
This paper presents an efficient preference elicitation framework for uncertain matroid optimization, where precise weight information is unavailable, but insights into possible weight values are accessible. The core innovation of our approach lies in its ability to systematically elicit user preferences, aligning the optimization process more closely with decision-makers' objectives. By incrementally querying preferences between pairs of elements, we iteratively refine the parametric uncertainty regions, leveraging the structural properties of matroids. Our method aims to achieve the exact optimum by reducing regret with a few elicitation rounds. Additionally, our approach avoids the computation of Minimax Regret and the use of Linear programming solvers at every iteration, unlike previous methods. Experimental results on four standard matroids demonstrate that our method reaches optimality more quickly and with fewer preference queries than existing techniques.
♻ ☆ AutoBayes: A Compositional Framework for Generalized Variational Inference
We introduce a new compositional framework for generalized variational inference, clarifying the different parts of a model, how they interact, and how they compose. We explain that both exact Bayesian inference and the loss functions typical of variational inference (such as variational free energy and its generalizations) satisfy chain rules akin to that of reverse-mode automatic differentiation, and we advocate for exploiting this to build and optimize models accordingly. To this end, we construct a series of compositional tools: for building models; for constructing their inversions; for attaching local loss functions; and for exposing parameters. Finally, we explain how the resulting parameterized statistical games may be optimized locally, too. We illustrate our framework with a number of classic examples, pointing to new areas of extensibility that are revealed.
comment: 15 pages. v2: fixed a typo
♻ ☆ A Universal Model Combining Differential Equations and Neural Networks for Ball Trajectory Prediction
This paper presents a data driven universal ball trajectory prediction method integrated with physics equations. Existing methods are designed for specific ball types and struggle to generalize. This challenge arises from three key factors. First, learning-based models require large datasets but suffer from accuracy drops in unseen scenarios. Second, physics-based models rely on complex formulas and detailed inputs, yet accurately obtaining ball states, such as spin, is often impractical. Third, integrating physical principles with neural networks to achieve high accuracy, fast inference, and strong generalization remains difficult. To address these issues, we propose an innovative approach that incorporates physics-based equations and neural networks. We first derive three generalized physical formulas. Then, using a neural network and observed trajectory points, we infer certain parameters while fitting the remaining ones. These formulas enable precise trajectory prediction with minimal training data: only a few dozen samples. Extensive experiments demonstrate our method superiority in generalization, real-time performance, and accuracy.
comment: This submission was made without my advisor's consent, and I mistakenly uploaded an incorrect version of the paper. Additionally, some content in the paper should not be made publicly available at this time, as per my advisor's wishes. I apologize for any inconvenience this may have caused
♻ ☆ Learning Causal Transition Matrix for Instance-dependent Label Noise
Noisy labels are both inevitable and problematic in machine learning methods, as they negatively impact models' generalization ability by causing overfitting. In the context of learning with noise, the transition matrix plays a crucial role in the design of statistically consistent algorithms. However, the transition matrix is often considered unidentifiable. One strand of methods typically addresses this problem by assuming that the transition matrix is instance-independent; that is, the probability of mislabeling a particular instance is not influenced by its characteristics or attributes. This assumption is clearly invalid in complex real-world scenarios. To better understand the transition relationship and relax this assumption, we propose to study the data generation process of noisy labels from a causal perspective. We discover that an unobservable latent variable can affect either the instance itself, the label annotation procedure, or both, which complicates the identification of the transition matrix. To address various scenarios, we have unified these observations within a new causal graph. In this graph, the input instance is divided into a noise-resistant component and a noise-sensitive component based on whether they are affected by the latent variable. These two components contribute to identifying the ``causal transition matrix'', which approximates the true transition matrix with theoretical guarantee. In line with this, we have designed a novel training framework that explicitly models this causal relationship and, as a result, achieves a more accurate model for inferring the clean label.
♻ ☆ Using Anomaly Detection to Detect Poisoning Attacks in Federated Learning Applications
Adversarial attacks such as poisoning attacks have attracted the attention of many machine learning researchers. Traditionally, poisoning attacks attempt to inject adversarial training data in order to manipulate the trained model. In federated learning (FL), data poisoning attacks can be generalized to model poisoning attacks, which cannot be detected by simpler methods due to the lack of access to local training data by the detector. State-of-the-art poisoning attack detection methods for FL have various weaknesses, e.g., the number of attackers has to be known or not high enough, working with i.i.d. data only, and high computational complexity. To overcome above weaknesses, we propose a novel framework for detecting poisoning attacks in FL, which employs a reference model based on a public dataset and an auditor model to detect malicious updates. We implemented a detector based on the proposed framework and using a one-class support vector machine (OC-SVM), which reaches the lowest possible computational complexity O(K) where K is the number of clients. We evaluated our detector's performance against state-of-the-art (SOTA) poisoning attacks for two typical applications of FL: electrocardiograph (ECG) classification and human activity recognition (HAR). Our experimental results validated the performance of our detector over other SOTA detection methods.
comment: We will updated this article soon
♻ ☆ Lessons and Insights from a Unifying Study of Parameter-Efficient Fine-Tuning (PEFT) in Visual Recognition CVPR 2025
Parameter-efficient fine-tuning (PEFT) has attracted significant attention due to the growth of pre-trained model sizes and the need to fine-tune (FT) them for superior downstream performance. Despite a surge in new PEFT methods, a systematic study to understand their performance and suitable application scenarios is lacking, leaving questions like "when to apply PEFT" and "which method to use" largely unanswered, especially in visual recognition. In this paper, we conduct a unifying empirical study of representative PEFT methods with Vision Transformers. We systematically tune their hyperparameters to fairly compare their accuracy on downstream tasks. Our study offers a practical user guide and unveils several new insights. First, if tuned carefully, different PEFT methods achieve similar accuracy in the low-shot benchmark VTAB-1K. This includes simple approaches like FT the bias terms that were reported inferior. Second, despite similar accuracy, we find that PEFT methods make different mistakes and high-confidence predictions, likely due to their different inductive biases. Such an inconsistency (or complementarity) opens up the opportunity for ensemble methods, and we make preliminary attempts at this. Third, going beyond the commonly used low-shot tasks, we find that PEFT is also useful in many-shot regimes, achieving comparable or better accuracy than full FT while using significantly fewer parameters. Lastly, we investigate PEFT's ability to preserve a pre-trained model's robustness to distribution shifts (e.g., CLIP). Perhaps not surprisingly, PEFT approaches outperform full FT alone. However, with weight-space ensembles, full FT can better balance target distribution and distribution shift performance, suggesting a future research direction for robust PEFT.
comment: CVPR 2025. The code is available at https://github.com/OSU-MLB/ViT_PEFT_Vision
♻ ☆ A Mechanistic Explanatory Strategy for XAI
Despite significant advancements in XAI, scholars continue to note a persistent lack of robust conceptual foundations and integration with broader discourse on scientific explanation. In response, emerging XAI research increasingly draws on explanatory strategies from various scientific disciplines and the philosophy of science to address these gaps. This paper outlines a mechanistic strategy for explaining the functional organization of deep learning systems, situating recent developments in AI explainability within a broader philosophical context. According to the mechanistic approach, explaining opaque AI systems involves identifying the mechanisms underlying decision-making processes. For deep neural networks, this means discerning functionally relevant components - such as neurons, layers, circuits, or activation patterns - and understanding their roles through decomposition, localization, and recomposition. Proof-of-principle case studies from image recognition and language modeling align this theoretical framework with recent research from OpenAI and Anthropic. The findings suggest that pursuing mechanistic explanations can uncover elements that traditional explainability techniques may overlook, ultimately contributing to more thoroughly explainable AI.
comment: Forthcoming in M\"uller, V. C., Dung, L., L\"ohr, G., & Rumana, A. (Eds.). Philosophy of Artificial Intelligence: The State of the Art, Synthese Library, Springer Nature. Please cite the published version
♻ ☆ LoTUS: Large-Scale Machine Unlearning with a Taste of Uncertainty CVPR 2025
We present LoTUS, a novel Machine Unlearning (MU) method that eliminates the influence of training samples from pre-trained models, avoiding retraining from scratch. LoTUS smooths the prediction probabilities of the model up to an information-theoretic bound, mitigating its over-confidence stemming from data memorization. We evaluate LoTUS on Transformer and ResNet18 models against eight baselines across five public datasets. Beyond established MU benchmarks, we evaluate unlearning on ImageNet1k, a large-scale dataset, where retraining is impractical, simulating real-world conditions. Moreover, we introduce the novel Retrain-Free Jensen-Shannon Divergence (RF-JSD) metric to enable evaluation under real-world conditions. The experimental results show that LoTUS outperforms state-of-the-art methods in terms of both efficiency and effectiveness. Code: https://github.com/cspartalis/LoTUS.
comment: Accepted as a main conference paper at CVPR 2025 (https://cvpr.thecvf.com/virtual/2025/poster/33292)
♻ ☆ Evaluating Negative Sampling Approaches for Neural Topic Models
Negative sampling has emerged as an effective technique that enables deep learning models to learn better representations by introducing the paradigm of learn-to-compare. The goal of this approach is to add robustness to deep learning models to learn better representation by comparing the positive samples against the negative ones. Despite its numerous demonstrations in various areas of computer vision and natural language processing, a comprehensive study of the effect of negative sampling in an unsupervised domain like topic modeling has not been well explored. In this paper, we present a comprehensive analysis of the impact of different negative sampling strategies on neural topic models. We compare the performance of several popular neural topic models by incorporating a negative sampling technique in the decoder of variational autoencoder-based neural topic models. Experiments on four publicly available datasets demonstrate that integrating negative sampling into topic models results in significant enhancements across multiple aspects, including improved topic coherence, richer topic diversity, and more accurate document classification. Manual evaluations also indicate that the inclusion of negative sampling into neural topic models enhances the quality of the generated topics. These findings highlight the potential of negative sampling as a valuable tool for advancing the effectiveness of neural topic models.
comment: Code is available at: https://github.com/AdhyaSuman/Eval_NegTM
♻ ☆ Inductive Moment Matching
Diffusion models and Flow Matching generate high-quality samples but are slow at inference, and distilling them into few-step models often leads to instability and extensive tuning. To resolve these trade-offs, we propose Inductive Moment Matching (IMM), a new class of generative models for one- or few-step sampling with a single-stage training procedure. Unlike distillation, IMM does not require pre-training initialization and optimization of two networks; and unlike Consistency Models, IMM guarantees distribution-level convergence and remains stable under various hyperparameters and standard model architectures. IMM surpasses diffusion models on ImageNet-256x256 with 1.99 FID using only 8 inference steps and achieves state-of-the-art 2-step FID of 1.98 on CIFAR-10 for a model trained from scratch.
♻ ☆ Dataset Distillation for Quantum Neural Networks
Training Quantum Neural Networks (QNNs) on large amount of classical data can be both time consuming as well as expensive. Higher amount of training data would require higher number of gradient descent steps to reach convergence. This, in turn would imply that the QNN will require higher number of quantum executions, thereby driving up its overall execution cost. In this work, we propose performing the dataset distillation process for QNNs, where we use a novel quantum variant of classical LeNet model containing residual connection and trainable Hermitian observable in the Parametric Quantum Circuit (PQC) of the QNN. This approach yields highly informative yet small number of training data at similar performance as the original data. We perform distillation for MNIST and Cifar-10 datasets, and on comparison with classical models observe that both the datasets yield reasonably similar post-inferencing accuracy on quantum LeNet (91.9% MNIST, 50.3% Cifar-10) compared to classical LeNet (94% MNIST, 54% Cifar-10). We also introduce a non-trainable Hermitian for ensuring stability in the distillation process and note marginal reduction of up to 1.8% (1.3%) for MNIST (Cifar-10) dataset.
comment: 5 pages, 4 figures, 2 tables
♻ ☆ Deep learning framework for action prediction reveals multi-timescale locomotor control
Modeling movement in real-world tasks is a fundamental goal for motor control, biomechanics, and rehabilitation engineering. However, widely used data-driven models of essential tasks like locomotion make simplifying assumptions such as linear and fixed timescale mappings between past inputs and future actions, which do not generalize to real-world contexts. Here, we develop a deep learning-based framework for action prediction with architecture-dependent trial embeddings, outperforming traditional models across contexts (walking and running, treadmill and overground, varying terrains) and input modalities (multiple body states, gaze). We find that neural network architectures with flexible input history-dependence like GRU and Transformer perform best overall. By quantifying the model's predictions relative to an autoregressive baseline, we identify context- and modality-dependent timescales. These analyses reveal that there is greater reliance on fast-timescale predictions in complex terrain, gaze predicts future foot placement before body states, and the full-body state predictions precede those by center-of-mass-relevant states. This deep learning framework for action prediction provides quantifiable insights into the control of real-world locomotion and can be extended to other actions, contexts, and populations.
♻ ☆ Persistent Homology for Structural Characterization in Disordered Systems
We propose a unified framework based on persistent homology (PH) to characterize both local and global structures in disordered systems. It can simultaneously generate local and global descriptors using the same algorithm and data structure, and has shown to be highly effective and interpretable in predicting particle rearrangements and classifying global phases. We also demonstrated that using a single variable enables a linear SVM to achieve nearly perfect three-phase classification. Inspired by this discovery, we define a non-parametric metric, the Separation Index (SI), which not only achieves this classification without sacrificing significant performance but also establishes a connection between particle environments and the global phase structure. Our methods provide an effective framework for understanding and analyzing the properties of disordered materials, with broad potential applications in materials science and even wider studies of complex systems.
comment: 21 pages, 19 figures
♻ ☆ Semi-Decision-Focused Learning with Deep Ensembles: A Practical Framework for Robust Portfolio Optimization ICLR 2025
I propose Semi-Decision-Focused Learning, a practical adaptation of Decision-Focused Learning for portfolio optimization. Rather than directly optimizing complex financial metrics, I employ simple target portfolios (Max-Sortino or One-Hot) and train models with a convex, cross-entropy loss. I further incorporate Deep Ensemble methods to reduce variance and stabilize performance. Experiments on two universes (one upward-trending and another range-bound) show consistent outperformance over baseline portfolios, demonstrating the effectiveness and robustness of my approach. Code is available at https://github.com/sDFLwDE/sDFLwDE
comment: ICLR 2025 Advances in Financial AI Workshop
♻ ☆ Denoising Diffusion Variational Inference: Diffusion Models as Expressive Variational Posteriors AAAI 2025
We propose denoising diffusion variational inference (DDVI), a black-box variational inference algorithm for latent variable models which relies on diffusion models as flexible approximate posteriors. Specifically, our method introduces an expressive class of diffusion-based variational posteriors that perform iterative refinement in latent space; we train these posteriors with a novel regularized evidence lower bound (ELBO) on the marginal likelihood inspired by the wake-sleep algorithm. Our method is easy to implement (it fits a regularized extension of the ELBO), is compatible with black-box variational inference, and outperforms alternative classes of approximate posteriors based on normalizing flows or adversarial networks. We find that DDVI improves inference and learning in deep latent variable models across common benchmarks as well as on a motivating task in biology -- inferring latent ancestry from human genomes -- where it outperforms strong baselines on the Thousand Genomes dataset.
comment: published at AAAI 2025; code available at https://github.com/topwasu/DDVI
♻ ☆ Mitigating Data Redundancy to Revitalize Transformer-based Long-Term Time Series Forecasting System
Long-term time-series forecasting (LTSF) is fundamental to various real-world applications, where Transformer-based models have become the dominant framework due to their ability to capture long-range dependencies. However, these models often experience overfitting due to data redundancy in rolling forecasting settings, limiting their generalization ability particularly evident in longer sequences with highly similar adjacent data. In this work, we introduce CLMFormer, a novel framework that mitigates redundancy through curriculum learning and a memory-driven decoder. Specifically, we progressively introduce Bernoulli noise to the training samples, which effectively breaks the high similarity between adjacent data points. This curriculum-driven noise introduction aids the memory-driven decoder by supplying more diverse and representative training data, enhancing the decoder's ability to model seasonal tendencies and dependencies in the time-series data. To further enhance forecasting accuracy, we introduce a memory-driven decoder. This component enables the model to capture seasonal tendencies and dependencies in the time-series data and leverages temporal relationships to facilitate the forecasting process. Extensive experiments on six real-world LTSF benchmarks show that CLMFormer consistently improves Transformer-based models by up to 30%, demonstrating its effectiveness in long-horizon forecasting.
comment: ACM TIST
♻ ☆ Map-Based Path Loss Prediction in Multiple Cities Using Convolutional Neural Networks
Radio deployments and spectrum planning benefit from path loss predictions. Obstructions along a communications link are often considered implicitly or through derived metrics such as representative clutter height or total obstruction depth. In this paper, we propose a path-specific path loss prediction method that uses convolutional neural networks to automatically perform feature extraction from 2-D obstruction height maps. Our methods result in low prediction error in a variety of environments without requiring derived metrics.
comment: 5 pages, 3 figures, 3 tables
♻ ☆ Dataset-learning duality and emergent criticality
In artificial neural networks, the activation dynamics of non-trainable variables is strongly coupled to the learning dynamics of trainable variables. During the activation pass, the boundary neurons (e.g., input neurons) are mapped to the bulk neurons (e.g., hidden neurons), and during the learning pass, both bulk and boundary neurons are mapped to changes in trainable variables (e.g., weights and biases). For example, in feed-forward neural networks, forward propagation is the activation pass and backward propagation is the learning pass. We show that a composition of the two maps establishes a duality map between a subspace of non-trainable boundary variables (e.g., dataset) and a tangent subspace of trainable variables (i.e., learning). In general, the dataset-learning duality is a complex non-linear map between high-dimensional spaces. We use duality to study the emergence of criticality, or the power-law distribution of fluctuations of the trainable variables, using a toy model at learning equilibrium. In particular, we show that criticality can emerge in the learning system even from the dataset in a non-critical state, and that the power-law distribution can be modified by changing either the activation function or the loss function.
comment: 22 pages, 5 figures, 1 table. Improved analysis; main results unchanged
♻ ☆ Training Domain Draft Models for Speculative Decoding: Best Practices and Insights SC
Speculative decoding is an effective method for accelerating inference of large language models (LLMs) by employing a small draft model to predict the output of a target model. However, when adapting speculative decoding to domain-specific target models, the acceptance rate of the generic draft model drops significantly due to domain shift. In this work, we systematically investigate knowledge distillation techniques for training domain draft models to improve their speculation accuracy. We compare white-box and black-box distillation approaches and explore their effectiveness in various data accessibility scenarios, including historical user queries, curated domain data, and synthetically generated alignment data. Our experiments across Function Calling, Biology, and Chinese domains show that offline distillation consistently outperforms online distillation by 11% to 25%, white-box distillation surpasses black-box distillation by 2% to 10%, and data scaling trends hold across domains. Additionally, we find that synthetic data can effectively align draft models and achieve 80% to 93% of the performance of training on historical user queries. These findings provide practical guidelines for training domain-specific draft models to improve speculative decoding efficiency.
comment: Published as a workshop paper at SCOPE - ICLR 2025
♻ ☆ FedECA: A Federated External Control Arm Method for Causal Inference with Time-To-Event Data in Distributed Settings
External control arms (ECA) can inform the early clinical development of experimental drugs and provide efficacy evidence for regulatory approval. However, the main challenge in implementing ECA lies in accessing real-world or historical clinical trials data. Indeed, regulations protecting patients' rights by strictly controlling data processing make pooling data from multiple sources in a central server often difficult. To address these limitations, we develop a new method, 'FedECA' that leverages federated learning (FL) to enable inverse probability of treatment weighting (IPTW) for time-to-event outcomes on separate cohorts without needing to pool data. To showcase the potential of FedECA, we apply it in different settings of increasing complexity culminating with a real-world use-case in which FedECA provides evidence for a differential effect between two drugs that would have otherwise gone unnoticed. By sharing our code, we hope FedECA will foster the creation of federated research networks and thus accelerate drug development.
comment: code available at: https://github.com/owkin/fedeca, bug in SMD computation present in v1 and v2 fixed, many experiments on real data added + fix in YODA experiments using imputed data instead of raw data (v3->v4) + affiliations fix + more precise wording for acknowledgments, real-world experiment results fixed by excluding data with bias + text polished (v5->v6)
♻ ☆ Ambient Noise Full Waveform Inversion with Neural Operators
Numerical simulations of seismic wave propagation are crucial for investigating velocity structures and improving seismic hazard assessment. However, standard methods such as finite difference or finite element are computationally expensive. Recent studies have shown that a new class of machine learning models, called neural operators, can solve the elastodynamic wave equation orders of magnitude faster than conventional methods. Full waveform inversion is a prime beneficiary of the accelerated simulations. Neural operators, as end-to-end differentiable operators, combined with automatic differentiation, provide an alternative approach to the adjoint-state method. Since neural operators do not involve the Born approximation, when used for full waveform inversion they have the potential to include additional phases and alleviate cycle-skipping problems present in traditional adjoint-state formulations. In this study, we demonstrate the first application of neural operators for full waveform inversion on a real seismic dataset, which consists of several nodal transects collected across the San Gabriel, Chino, and San Bernardino basins in the Los Angeles metropolitan area.
comment: Added references
♻ ☆ Similarity-Dissimilarity Loss for Multi-label Supervised Contrastive Learning
Supervised contrastive learning has achieved remarkable success by leveraging label information; however, determining positive samples in multi-label scenarios remains a critical challenge. In multi-label supervised contrastive learning (MSCL), relations among multi-label samples are not yet fully defined, leading to ambiguity in identifying positive samples and formulating contrastive loss functions to construct the representation space. To address these challenges, we: (i) first define five distinct multi-label relations in MSCL to systematically identify positive samples, (ii) introduce a novel Similarity-Dissimilarity Loss that dynamically re-weights samples through computing the similarity and dissimilarity factors between positive samples and given anchors based on multi-label relations, and (iii) further provide theoretical grounded proof for our method through rigorous mathematical analysis that supports the formulation and effectiveness of the proposed loss function. We conduct the experiments across both image and text modalities, and extend the evaluation to medical domain. The results demonstrate that our method consistently outperforms baselines in a comprehensive evaluation, confirming its effectiveness and robustness. Code is available at: https://github.com/guangminghuang/similarity-dissimilarity-loss.
♻ ☆ Fidelity-Imposed Displacement Editing for the Learn2Reg 2024 SHG-BF Challenge
Co-examination of second-harmonic generation (SHG) and bright-field (BF) microscopy enables the differentiation of tissue components and collagen fibers, aiding the analysis of human breast and pancreatic cancer tissues. However, large discrepancies between SHG and BF images pose challenges for current learning-based registration models in aligning SHG to BF. In this paper, we propose a novel multi-modal registration framework that employs fidelity-imposed displacement editing to address these challenges. The framework integrates batch-wise contrastive learning, feature-based pre-alignment, and instance-level optimization. Experimental results from the Learn2Reg COMULISglobe SHG-BF Challenge validate the effectiveness of our method, securing the 1st place on the online leaderboard.
comment: Due to the instruction and conflict of co-author
♻ ☆ Pretraining Generative Flow Networks with Inexpensive Rewards for Molecular Graph Generation
Generative Flow Networks (GFlowNets) have recently emerged as a suitable framework for generating diverse and high-quality molecular structures by learning from rewards treated as unnormalized distributions. Previous works in this framework often restrict exploration by using predefined molecular fragments as building blocks, limiting the chemical space that can be accessed. In this work, we introduce Atomic GFlowNets (A-GFNs), a foundational generative model leveraging individual atoms as building blocks to explore drug-like chemical space more comprehensively. We propose an unsupervised pre-training approach using drug-like molecule datasets, which teaches A-GFNs about inexpensive yet informative molecular descriptors such as drug-likeliness, topological polar surface area, and synthetic accessibility scores. These properties serve as proxy rewards, guiding A-GFNs towards regions of chemical space that exhibit desirable pharmacological properties. We further implement a goal-conditioned finetuning process, which adapts A-GFNs to optimize for specific target properties. In this work, we pretrain A-GFN on a subset of ZINC dataset, and by employing robust evaluation metrics we show the effectiveness of our approach when compared to other relevant baseline methods for a wide range of drug design tasks.
comment: arXiv admin note: text overlap with arXiv:2409.09702
♻ ☆ Elastic Federated Learning over Open Radio Access Network (O-RAN) for Concurrent Execution of Multiple Distributed Learning Tasks
Federated learning (FL) is a popular distributed machine learning (ML) technique in Internet of Things (IoT) networks, where resource-constrained devices collaboratively train ML models while preserving data privacy. However, implementation of FL over 5G-and-beyond wireless networks faces key challenges caused by (i) dynamics of the wireless network conditions and (ii) the coexistence of multiple FL-services in the system. In this paper, we unveil two key phenomena that arise from these challenges: over/under-provisioning of resources and perspective-driven load balancing, both of which significantly impact FL performance in IoT environments. We take the first steps towards addressing these phenomena by proposing a novel distributed ML architecture called elastic FL (EFL). EFL unleashes the full potential of Open RAN (O-RAN) systems and introduces an elastic resource provisioning methodology to execute FL-services. It further constitutes a multi-time-scale FL management system that introduces three dedicated network control functionalities tailored for FL-services, including (i) non-real-time (non-RT) system descriptor, which trains ML-based applications to predict both system and FL-related dynamics and parameters; (ii) near-RT FL controller, which handles O-RAN slicing and mobility management for the seamless execution of FL-services; (iii) FL MAC scheduler, which conducts real-time resource allocation to the end clients of various FL-services. We finally prototype EFL to demonstrate its potential in improving the performance of FL-services.
comment: 9 pages, 4 figures
♻ ☆ Autoregressive Action Sequence Learning for Robotic Manipulation
Designing a universal policy architecture that performs well across diverse robots and task configurations remains a key challenge. In this work, we address this by representing robot actions as sequential data and generating actions through autoregressive sequence modeling. Existing autoregressive architectures generate end-effector waypoints sequentially as word tokens in language modeling, which are limited to low-frequency control tasks. Unlike language, robot actions are heterogeneous and often include continuous values -- such as joint positions, 2D pixel coordinates, and end-effector poses -- which are not easily suited for language-based modeling. Based on this insight, we introduce a straightforward enhancement: we extend causal transformers' single-token prediction to support predicting a variable number of tokens in a single step through our Chunking Causal Transformer (CCT). This enhancement enables robust performance across diverse tasks of various control frequencies, greater efficiency by having fewer autoregression steps, and lead to a hybrid action sequence design by mixing different types of actions and using a different chunk size for each action type. Based on CCT, we propose the Autoregressive Policy (ARP) architecture, which solves manipulation tasks by generating hybrid action sequences. We evaluate ARP across diverse robotic manipulation environments, including Push-T, ALOHA, and RLBench, and show that ARP, as a universal architecture, matches or outperforms the environment-specific state-of-the-art in all tested benchmarks, while being more efficient in computation and parameter sizes. Videos of our real robot demonstrations, all source code and the pretrained models of ARP can be found at http://github.com/mlzxy/arp.
comment: (RA-L 2025) Add a new figure to explain why chunking autoregression works. Put back the previous in-depth discussion for arxiv release
♻ ☆ LLAVIDAL: A Large LAnguage VIsion Model for Daily Activities of Living CVPR 2025
Current Large Language Vision Models (LLVMs) trained on web videos perform well in general video understanding but struggle with fine-grained details, complex human-object interactions (HOI), and view-invariant representation learning essential for Activities of Daily Living (ADL). This limitation stems from a lack of specialized ADL video instruction-tuning datasets and insufficient modality integration to capture discriminative action representations. To address this, we propose a semi-automated framework for curating ADL datasets, creating ADL-X, a multiview, multimodal RGBS instruction-tuning dataset. Additionally, we introduce LLAVIDAL, an LLVM integrating videos, 3D skeletons, and HOIs to model ADL's complex spatiotemporal relationships. For training LLAVIDAL a simple joint alignment of all modalities yields suboptimal results; thus, we propose a Multimodal Progressive (MMPro) training strategy, incorporating modalities in stages following a curriculum. We also establish ADL MCQ and video description benchmarks to assess LLVM performance in ADL tasks. Trained on ADL-X, LLAVIDAL achieves state-of-the-art performance across ADL benchmarks. Code and data will be made publicly available at: https://adl-x.github.io/.
comment: CVPR 2025 Camera Ready
♻ ☆ Linear Diffusion Networks
We present Linear Diffusion Networks (LDNs), a novel architecture that reinterprets sequential data processing as a unified diffusion process. Our model integrates adaptive diffusion modules with localized nonlinear updates and a diffusion-inspired attention mechanism. This design enables efficient global information propagation while preserving fine-grained temporal details. LDN overcomes the limitations of conventional recurrent and transformer models by allowing full parallelization across time steps and supporting robust multi-scale temporal representations. Experiments on benchmark sequence modeling tasks demonstrate that LDN delivers competitive performance across ImageNet and LRA tasks.
♻ ☆ Unlocking Guidance for Discrete State-Space Diffusion and Flow Models
Generative models on discrete state-spaces have a wide range of potential applications, particularly in the domain of natural sciences. In continuous state-spaces, controllable and flexible generation of samples with desired properties has been realized using guidance on diffusion and flow models. However, these guidance approaches are not readily amenable to discrete state-space models. Consequently, we introduce a general and principled method for applying guidance on such models. Our method depends on leveraging continuous-time Markov processes on discrete state-spaces, which unlocks computational tractability for sampling from a desired guided distribution. We demonstrate the utility of our approach, Discrete Guidance, on a range of applications including guided generation of small-molecules, DNA sequences and protein sequences.
♻ ☆ Functional Acceleration for Policy Mirror Descent
We apply functional acceleration to the Policy Mirror Descent (PMD) general family of algorithms, which cover a wide range of novel and fundamental methods in Reinforcement Learning (RL). Leveraging duality, we propose a momentum-based PMD update. By taking the functional route, our approach is independent of the policy parametrization and applicable to large-scale optimization, covering previous applications of momentum at the level of policy parameters as a special case. We theoretically analyze several properties of this approach and complement with a numerical ablation study, which serves to illustrate the policy optimization dynamics on the value polytope, relative to different algorithmic design choices in this space. We further characterize numerically several features of the problem setting relevant for functional acceleration, and lastly, we investigate the impact of approximation on their learning mechanics.
♻ ☆ Mambular: A Sequential Model for Tabular Deep Learning
The analysis of tabular data has traditionally been dominated by gradient-boosted decision trees (GBDTs), known for their proficiency with mixed categorical and numerical features. However, recent deep learning innovations are challenging this dominance. This paper investigates the use of autoregressive state-space models for tabular data and compares their performance against established benchmark models. Additionally, we explore various adaptations of these models, including different pooling strategies, feature interaction mechanisms, and bi-directional processing techniques to understand their effectiveness for tabular data. Our findings indicate that interpreting features as a sequence and processing them and their interactions through structured state-space layers can lead to significant performance improvement. This research underscores the versatility of autoregressive models in tabular data analysis, positioning them as a promising alternative that could substantially enhance deep learning capabilities in this traditionally challenging area. The source code is available at https://github.com/basf/mamba-tabular.
♻ ☆ DeepIFSAC: Deep Imputation of Missing Values Using Feature and Sample Attention within Contrastive Framework
Missing values of varying patterns and rates in real-world tabular data pose a significant challenge in developing reliable data-driven models. The most commonly used statistical and machine learning methods for missing value imputation may be ineffective when the missing rate is high and not random. This paper explores row and column attention in tabular data as between-feature and between-sample attention in a novel framework to reconstruct missing values. The proposed method uses CutMix data augmentation within a contrastive learning framework to improve the uncertainty of missing value estimation. The performance and generalizability of trained imputation models are evaluated in set-aside test data folds with missing values. The proposed framework is compared with 11 state-of-the-art statistical, machine learning, and deep imputation methods using 12 diverse tabular data sets. The average performance rank of our proposed method demonstrates its superiority over the state-of-the-art methods for missing rates between 10% and 90% and three missing value types, especially when the missing values are not random. The quality of the imputed data using our proposed method is compared in a downstream patient classification task using real-world electronic health records. This paper highlights the heterogeneity of tabular data sets to recommend imputation methods based on missing value types and data characteristics.
♻ ☆ Decomposing The Dark Matter of Sparse Autoencoders
Sparse autoencoders (SAEs) are a promising technique for decomposing language model activations into interpretable linear features. However, current SAEs fall short of completely explaining model performance, resulting in "dark matter": unexplained variance in activations. This work investigates dark matter as an object of study in its own right. Surprisingly, we find that much of SAE dark matter -- about half of the error vector itself and >90% of its norm -- can be linearly predicted from the initial activation vector. Additionally, we find that the scaling behavior of SAE error norms at a per token level is remarkably predictable: larger SAEs mostly struggle to reconstruct the same contexts as smaller SAEs. We build on the linear representation hypothesis to propose models of activations that might lead to these observations. These insights imply that the part of the SAE error vector that cannot be linearly predicted ("nonlinear" error) might be fundamentally different from the linearly predictable component. To validate this hypothesis, we empirically analyze nonlinear SAE error and show that 1) it contains fewer not yet learned features, 2) SAEs trained on it are quantitatively worse, and 3) it is responsible for a proportional amount of the downstream increase in cross entropy loss when SAE activations are inserted into the model. Finally, we examine two methods to reduce nonlinear SAE error: inference time gradient pursuit, which leads to a very slight decrease in nonlinear error, and linear transformations from earlier layer SAE outputs, which leads to a larger reduction.
comment: Published in TMLR. Code at https://github.com/JoshEngels/SAE-Dark-Matter
♻ ☆ Phylo2Vec: a vector representation for binary trees
Binary phylogenetic trees inferred from biological data are central to understanding the shared history among evolutionary units. However, inferring the placement of latent nodes in a tree is computationally expensive. State-of-the-art methods rely on carefully designed heuristics for tree search, using different data structures for easy manipulation (e.g., classes in object-oriented programming languages) and readable representation of trees (e.g., Newick-format strings). Here, we present Phylo2Vec, a parsimonious encoding for phylogenetic trees that serves as a unified approach for both manipulating and representing phylogenetic trees. Phylo2Vec maps any binary tree with $n$ leaves to a unique integer vector of length $n-1$. The advantages of Phylo2Vec are fourfold: i) fast tree sampling, (ii) compressed tree representation compared to a Newick string, iii) quick and unambiguous verification if two binary trees are identical topologically, and iv) systematic ability to traverse tree space in very large or small jumps. As a proof of concept, we use Phylo2Vec for maximum likelihood inference on five real-world datasets and show that a simple hill-climbing-based optimisation scheme can efficiently traverse the vastness of tree space from a random to an optimal tree.
comment: 38 pages, 9 figures, 1 table, 2 supplementary figures
♻ ☆ LLM4DV: Using Large Language Models for Hardware Test Stimuli Generation
Hardware design verification (DV) is a process that checks the functional equivalence of a hardware design against its specifications, improving hardware reliability and robustness. A key task in the DV process is the test stimuli generation, which creates a set of conditions or inputs for testing. These test conditions are often complex and specific to the given hardware design, requiring substantial human engineering effort to optimize. We seek a solution of automated and efficient testing for arbitrary hardware designs that takes advantage of large language models (LLMs). LLMs have already shown promising results for improving hardware design automation, but remain under-explored for hardware DV. In this paper, we propose an open-source benchmarking framework named LLM4DV that efficiently orchestrates LLMs for automated hardware test stimuli generation. Our analysis evaluates six different LLMs involving six prompting improvements over eight hardware designs and provides insight for future work on LLMs development for efficient automated DV.
♻ ☆ Simplifying Deep Temporal Difference Learning
Q-learning played a foundational role in the field reinforcement learning (RL). However, TD algorithms with off-policy data, such as Q-learning, or nonlinear function approximation like deep neural networks require several additional tricks to stabilise training, primarily a large replay buffer and target networks. Unfortunately, the delayed updating of frozen network parameters in the target network harms the sample efficiency and, similarly, the large replay buffer introduces memory and implementation overheads. In this paper, we investigate whether it is possible to accelerate and simplify off-policy TD training while maintaining its stability. Our key theoretical result demonstrates for the first time that regularisation techniques such as LayerNorm can yield provably convergent TD algorithms without the need for a target network or replay buffer, even with off-policy data. Empirically, we find that online, parallelised sampling enabled by vectorised environments stabilises training without the need for a large replay buffer. Motivated by these findings, we propose PQN, our simplified deep online Q-Learning algorithm. Surprisingly, this simple algorithm is competitive with more complex methods like: Rainbow in Atari, PPO-RNN in Craftax, QMix in Smax, and can be up to 50x faster than traditional DQN without sacrificing sample efficiency. In an era where PPO has become the go-to RL algorithm, PQN reestablishes off-policy Q-learning as a viable alternative.
♻ ☆ HyperFLINT: Hypernetwork-based Flow Estimation and Temporal Interpolation for Scientific Ensemble Visualization
We present HyperFLINT (Hypernetwork-based FLow estimation and temporal INTerpolation), a novel deep learning-based approach for estimating flow fields, temporally interpolating scalar fields, and facilitating parameter space exploration in spatio-temporal scientific ensemble data. This work addresses the critical need to explicitly incorporate ensemble parameters into the learning process, as traditional methods often neglect these, limiting their ability to adapt to diverse simulation settings and provide meaningful insights into the data dynamics. HyperFLINT introduces a hypernetwork to account for simulation parameters, enabling it to generate accurate interpolations and flow fields for each timestep by dynamically adapting to varying conditions, thereby outperforming existing parameter-agnostic approaches. The architecture features modular neural blocks with convolutional and deconvolutional layers, supported by a hypernetwork that generates weights for the main network, allowing the model to better capture intricate simulation dynamics. A series of experiments demonstrates HyperFLINT's significantly improved performance in flow field estimation and temporal interpolation, as well as its potential in enabling parameter space exploration, offering valuable insights into complex scientific ensembles.
♻ ☆ Locally Private Nonparametric Contextual Multi-armed Bandits
Motivated by privacy concerns in sequential decision-making on sensitive data, we address the challenge of nonparametric contextual multi-armed bandits (MAB) under local differential privacy (LDP). We develop a uniform-confidence-bound-type estimator, showing its minimax optimality supported by a matching minimax lower bound. We further consider the case where auxiliary datasets are available, subject also to (possibly heterogeneous) LDP constraints. Under the widely-used covariate shift framework, we propose a jump-start scheme to effectively utilize the auxiliary data, the minimax optimality of which is further established by a matching lower bound. Comprehensive experiments on both synthetic and real-world datasets validate our theoretical results and underscore the effectiveness of the proposed methods.
♻ ☆ MetaSel: A Test Selection Approach for Fine-tuned DNN Models
Deep Neural Networks (DNNs) face challenges during deployment due to data distribution shifts. Fine-tuning adapts pre-trained models to new contexts requiring smaller labeled sets. However, testing fine-tuned models under constrained labeling budgets remains a critical challenge. This paper introduces MetaSel, a new approach, tailored for fine-tuned DNN models, to select tests from unlabeled inputs. MetaSel assumes that fine-tuned and pre-trained models share related data distributions and exhibit similar behaviors for many inputs. However, their behaviors diverge within the input subspace where fine-tuning alters decision boundaries, making those inputs more prone to misclassification. Unlike general approaches that rely solely on the DNN model and its input set, MetaSel leverages information from both the fine-tuned and pre-trained models and their behavioral differences to estimate misclassification probability for unlabeled test inputs, enabling more effective test selection. Our extensive empirical evaluation, comparing MetaSel against 10 state-of-the-art approaches and involving 68 fine-tuned models across weak, medium, and strong distribution shifts, demonstrates that MetaSel consistently delivers significant improvements in Test Relative Coverage (TRC) over existing baselines, particularly under highly constrained labeling budgets. MetaSel shows average TRC improvements of 28.46% to 56.18% over the most frequent second-best baselines while maintaining a high TRC median and low variability. Our results confirm MetaSel's practicality, robustness, and cost-effectiveness for test selection in the context of fine-tuned models.
♻ ☆ FW-Merging: Scaling Model Merging with Frank-Wolfe Optimization
Model merging has emerged as a promising approach for multi-task learning (MTL), offering a data-efficient alternative to conventional fine-tuning. However, with the rapid development of the open-source AI ecosystem and the increasing availability of fine-tuned foundation models, existing model merging methods face two key limitations: (i) They are primarily designed for in-house fine-tuned models, making them less adaptable to diverse model sources with partially unknown model and task information, (ii) They struggle to scale effectively when merging numerous model checkpoints. To address these challenges, we formulate model merging as a constrained optimization problem and introduce a novel approach: Frank-Wolfe Merging (FW-Merging). Inspired by Frank-Wolfe optimization, our approach iteratively selects the most relevant model in the pool to minimize a linear approximation of the objective function and then executes a local merging similar to the Frank-Wolfe update. The objective function is designed to capture the desired behavior of the target-merged model, while the fine-tuned candidate models define the constraint set. More importantly, FW-Merging serves as an orthogonal technique for existing merging methods, seamlessly integrating with them to further enhance accuracy performance. Our experiments show that FW-Merging scales across diverse model sources, remaining stable with 16 irrelevant models and improving by 15.3% with 16 relevant models on 20 CV tasks, while maintaining constant memory overhead, unlike the linear overhead of data-informed merging methods. Compared with the state-of-the-art approaches, FW-Merging surpasses the data-free merging method by 32.8% and outperforms the data-informed Adamerging by 8.39% when merging 20 ViT models. Our code is open-sourced at github.com/hmarkc/FW-Merging.
♻ ☆ Internet of Things-Based Smart Precision Farming in Soilless Agriculture:Opportunities and Challenges for Global Food Security
The rapid growth of the global population and the continuous decline in cultivable land pose significant threats to food security. This challenge worsens as climate change further reduces the availability of farmland. Soilless agriculture, such as hydroponics, aeroponics, and aquaponics, offers a sustainable solution by enabling efficient crop cultivation in controlled environments. The integration of the Internet of Things (IoT) with smart precision farming improves resource efficiency, automates environmental control, and ensures stable and high-yield crop production. IoT-enabled smart farming systems utilize real-time monitoring, data-driven decision-making, and automation to optimize water and nutrient usage while minimizing human intervention. This paper explores the opportunities and challenges of IoT-based soilless farming, highlighting its role in sustainable agriculture, urban farming, and global food security. These advanced farming methods ensure greater productivity, resource conservation, and year-round cultivation. However, they also face challenges such as high initial investment, technological dependency, and energy consumption. Through a comprehensive study, bibliometric analysis, and comparative analysis, this research highlights current trends and research gaps. It also outlines future directions for researchers, policymakers, and industry stakeholders to drive innovation and scalability in IoT-driven soilless agriculture. By emphasizing the benefits of vertical farming and Controlled Environment Agriculture (CEA)-enabled soilless techniques, this paper supports informed decision-making to address food security challenges and promote sustainable agricultural innovations.
♻ ☆ DeltaZip: Efficient Serving of Multiple Full-Model-Tuned LLMs EuroSys 2025
Fine-tuning large language models (LLMs) greatly improves model quality for downstream tasks. However, serving many fine-tuned LLMs concurrently is challenging due to the sporadic, bursty, and varying request patterns of different LLMs. To bridge this gap, we present DeltaZip, an LLM serving system that efficiently serves multiple full-parameter fine-tuned models concurrently by aggressively compressing model deltas by up to 10x while maintaining high model quality. The key insight behind this design is that fine-tuning results in small-magnitude changes to the pre-trained model. By co-designing the serving system with the compression algorithm, DeltaZip achieves 2x to 12x improvement in throughput compared to the state-of-the-art systems.
comment: EuroSys 2025'
♻ ☆ Pfungst and Clever Hans: Identifying the unintended cues in a widely used Alzheimer's disease MRI dataset using explainable deep learning
Backgrounds. Deep neural networks have demonstrated high accuracy in classifying Alzheimer's disease (AD). This study aims to enlighten the underlying black-box nature and reveal individual contributions of T1-weighted (T1w) gray-white matter texture, volumetric information and preprocessing on classification performance. Methods. We utilized T1w MRI data from the Alzheimer's Disease Neuroimaging Initiative to distinguish matched AD patients (990 MRIs) from healthy controls (990 MRIs). Preprocessing included skull stripping and binarization at varying thresholds to systematically eliminate texture information. A deep neural network was trained on these configurations, and the model performance was compared using McNemar tests with discrete Bonferroni-Holm correction. Layer-wise Relevance Propagation (LRP) and structural similarity metrics between heatmaps were applied to analyze learned features. Results. Classification performance metrics (accuracy, sensitivity, and specificity) were comparable across all configurations, indicating a negligible influence of T1w gray- and white signal texture. Models trained on binarized images demonstrated similar feature performance and relevance distributions, with volumetric features such as atrophy and skull-stripping features emerging as primary contributors. Conclusions. We revealed a previously undiscovered Clever Hans effect in a widely used AD MRI dataset. Deep neural networks classification predominantly rely on volumetric features, while eliminating gray-white matter T1w texture did not decrease the performance. This study clearly demonstrates an overestimation of the importance of gray-white matter contrasts, at least for widely used structural T1w images, and highlights potential misinterpretation of performance metrics.
♻ ☆ Federated Causal Inference: Multi-Study ATE Estimation beyond Meta-Analysis
We study Federated Causal Inference, an approach to estimate treatment effects from decentralized data across centers. We compare three classes of Average Treatment Effect (ATE) estimators derived from the Plug-in G-Formula, ranging from simple meta-analysis to one-shot and multi-shot federated learning, the latter leveraging the full data to learn the outcome model (albeit requiring more communication). Focusing on Randomized Controlled Trials (RCTs), we derive the asymptotic variance of these estimators for linear models. Our results provide practical guidance on selecting the appropriate estimator for various scenarios, including heterogeneity in sample sizes, covariate distributions, treatment assignment schemes, and center effects. We validate these findings with a simulation study.
♻ ☆ Communities in the Kuramoto Model: Dynamics and Detection via Path Signatures
The behavior of multivariate dynamical processes is often governed by underlying structural connections that relate the components of the system. For example, brain activity which is often measured via time series is determined by an underlying structural graph, where nodes represent neurons or brain regions and edges represent cortical connectivity. Existing methods for inferring structural connections from observed dynamics, such as correlation-based or spectral techniques, may fail to fully capture complex relationships in high-dimensional time series in an interpretable way. Here, we propose the use of path signatures a mathematical framework that encodes geometric and temporal properties of continuous paths to address this problem. Path signatures provide a reparametrization-invariant characterization of dynamical data and, in particular, can be used to compute the lead matrix which reveals lead-lag phenomena. We showcase our approach on time series from coupled oscillators in the Kuramoto model defined on a stochastic block model graph, termed the Kuramoto stochastic block model (KSBM). Using mean-field theory and Gaussian approximations, we analytically derive reduced models of KSBM dynamics in different temporal regimes and theoretically characterize the lead matrix in these settings. Leveraging these insights, we propose a novel signature-based community detection algorithm, achieving exact recovery of structural communities from observed time series in multiple KSBM instances. Our results demonstrate that path signatures provide a novel perspective on analyzing complex neural data and other high-dimensional systems, explicitly exploiting temporal functional relationships to infer underlying structure.
comment: 46 pages, 13 figures
♻ ☆ A Quantum Neural Network Transfer-Learning Model for Forecasting Problems with Continuous and Discrete Variables
This study introduces simple yet effective continuous- and discrete-variable quantum neural network (QNN) models as a transfer-learning approach for forecasting tasks. The CV-QNN features a single quantum layer with two qubits to establish entanglement and utilizes a minimal set of quantum gates, including displacement, rotation, beam splitter, squeezing, and a non-Gaussian cubic-phase gate, with a maximum of eight trainable parameters. A key advantage of this model is its ability to be trained on a single dataset, after which the learned parameters can be transferred to other forecasting problems with little to no fine-tuning. Initially trained on the Kurdistan load demand dataset, the model's frozen parameters are successfully applied to various forecasting tasks, including energy consumption, traffic flow, weather conditions, and cryptocurrency price prediction, demonstrating strong performance. Furthermore, the study introduces a discrete-variable quantum model with an equivalent 2- and 4-wire configuration and presents a performance assessment, showing good but relatively lower effectiveness compared to the continuous-variable model.
♻ ☆ One-vs.-One Mitigation of Intersectional Bias: A General Method to Extend Fairness-Aware Binary Classification
With the widespread adoption of machine learning in the real world, the impact of the discriminatory bias has attracted attention. In recent years, various methods to mitigate the bias have been proposed. However, most of them have not considered intersectional bias, which brings unfair situations where people belonging to specific subgroups of a protected group are treated worse when multiple sensitive attributes are taken into consideration. To mitigate this bias, in this paper, we propose a method called One-vs.-One Mitigation by applying a process of comparison between each pair of subgroups related to sensitive attributes to the fairness-aware machine learning for binary classification. We compare our method and the conventional fairness-aware binary classification methods in comprehensive settings using three approaches (pre-processing, in-processing, and post-processing), six metrics (the ratio and difference of demographic parity, equalized odds, and equal opportunity), and two real-world datasets (Adult and COMPAS). As a result, our method mitigates the intersectional bias much better than conventional methods in all the settings. With the result, we open up the potential of fairness-aware binary classification for solving more realistic problems occurring when there are multiple sensitive attributes.
♻ ☆ MetaToken: Detecting Hallucination in Image Descriptions by Meta Classification
Large Vision Language Models (LVLMs) have shown remarkable capabilities in multimodal tasks like visual question answering or image captioning. However, inconsistencies between the visual information and the generated text, a phenomenon referred to as hallucinations, remain an unsolved problem with regard to the trustworthiness of LVLMs. To address this problem, recent works proposed to incorporate computationally costly Large (Vision) Language Models in order to detect hallucinations on a sentence- or subsentence-level. In this work, we introduce MetaToken, a lightweight binary classifier to detect hallucinations on the token-level at negligible cost. Based on a statistical analysis, we reveal key factors of hallucinations in LVLMs. MetaToken can be applied to any open-source LVLM without any knowledge about ground truth data providing a calibrated detection of hallucinations. We evaluate our method on four state-of-the-art LVLMs demonstrating the effectiveness of our approach.
♻ ☆ Imitation Learning with Limited Actions via Diffusion Planners and Deep Koopman Controllers ICRA
Recent advances in diffusion-based robot policies have demonstrated significant potential in imitating multi-modal behaviors. However, these approaches typically require large quantities of demonstration data paired with corresponding robot action labels, creating a substantial data collection burden. In this work, we propose a plan-then-control framework aimed at improving the action-data efficiency of inverse dynamics controllers by leveraging observational demonstration data. Specifically, we adopt a Deep Koopman Operator framework to model the dynamical system and utilize observation-only trajectories to learn a latent action representation. This latent representation can then be effectively mapped to real high-dimensional continuous actions using a linear action decoder, requiring minimal action-labeled data. Through experiments on simulated robot manipulation tasks and a real robot experiment with multi-modal expert demonstrations, we demonstrate that our approach significantly enhances action-data efficiency and achieves high task success rates with limited action data.
comment: Accepted to IEEE International Conference on Robotics and Automation (ICRA) 2025
♻ ☆ Exploring Robustness of Image Recognition Models on Hardware Accelerators
As the usage of Artificial Intelligence (AI) on resource-intensive and safety-critical tasks increases, a variety of Machine Learning (ML) compilers have been developed, enabling compatibility of Deep Neural Networks (DNNs) with a variety of hardware acceleration devices. However, given that DNNs are widely utilized for challenging and demanding tasks, the behavior of these compilers must be verified. To this direction, we propose MutateNN, a tool that utilizes elements of both differential and mutation testing in order to examine the robustness of image recognition models when deployed on hardware accelerators with different capabilities, in the presence of faults in their target device code - introduced either by developers, or problems in their compilation process. We focus on the image recognition domain by applying mutation testing to 7 well-established DNN models, introducing 21 mutations of 6 different categories. We deployed our mutants on 4 different hardware acceleration devices of varying capabilities and observed that DNN models presented discrepancies of up to 90.3% in mutants related to conditional operators across devices. We also observed that mutations related to layer modification, arithmetic types and input affected severely the overall model performance (up to 99.8%) or led to model crashes, in a consistent manner across devices.
comment: 7 pages, 6 figures
♻ ☆ Early Classification of Time Series: Taxonomy and Benchmark
In many situations, the measurements of a studied phenomenon are provided sequentially, and the prediction of its class needs to be made as early as possible so as not to incur too high a time penalty, but not too early and risk paying the cost of misclassification. This problem has been particularly studied in the case of time series, and is known as Early Classification of Time Series (ECTS). Although it has been the subject of a growing body of literature, there is still a lack of a systematic, shared evaluation protocol to compare the relative merits of the various existing methods. This document begins by situating these methods within a principle-based taxonomy. It defines dimensions for organizing their evaluation, and then reports the results of a very extensive set of experiments along these dimensions involving nine state-of-the art ECTS algorithms. In addition, these and other experiments can be carried out using an open-source library in which most of the existing ECTS algorithms have been implemented (see https://github.com/ML-EDM/ml_edm).
♻ ☆ RL-RC-DoT: A Block-level RL agent for Task-Aware Video Compression
Video encoders optimize compression for human perception by minimizing reconstruction error under bit-rate constraints. In many modern applications such as autonomous driving, an overwhelming majority of videos serve as input for AI systems performing tasks like object recognition or segmentation, rather than being watched by humans. It is therefore useful to optimize the encoder for a downstream task instead of for perceptual image quality. However, a major challenge is how to combine such downstream optimization with existing standard video encoders, which are highly efficient and popular. Here, we address this challenge by controlling the Quantization Parameters (QPs) at the macro-block level to optimize the downstream task. This granular control allows us to prioritize encoding for task-relevant regions within each frame. We formulate this optimization problem as a Reinforcement Learning (RL) task, where the agent learns to balance long-term implications of choosing QPs on both task performance and bit-rate constraints. Notably, our policy does not require the downstream task as an input during inference, making it suitable for streaming applications and edge devices such as vehicles. We demonstrate significant improvements in two tasks, car detection, and ROI (saliency) encoding. Our approach improves task performance for a given bit rate compared to traditional task agnostic encoding methods, paving the way for more efficient task-aware video compression.
♻ ☆ Solvation Free Energies from Neural Thermodynamic Integration
We present a method for computing free-energy differences using thermodynamic integration with a neural network potential that interpolates between two target Hamiltonians. The interpolation is defined at the sample distribution level, and the neural network potential is optimized to match the corresponding equilibrium potential at every intermediate time-step. Once the interpolating potentials and samples are well-aligned, the free-energy difference can be estimated using (neural) thermodynamic integration. To target molecular systems, we simultaneously couple Lennard-Jones and electrostatic interactions and model the rigid-body rotation of molecules. We report accurate results for several benchmark systems: a Lennard-Jones particle in a Lennard-Jones fluid, as well as the insertion of both water and methane solutes in a water solvent at atomistic resolution using a simple three-body neural-network potential.
♻ ☆ Towards LLM Unlearning Resilient to Relearning Attacks: A Sharpness-Aware Minimization Perspective and Beyond
The LLM unlearning technique has recently been introduced to comply with data regulations and address the safety and ethical concerns of LLMs by removing the undesired data-model influence. However, state-of-the-art unlearning methods face a critical vulnerability: they are susceptible to ``relearning'' the removed information from a small number of forget data points, known as relearning attacks. In this paper, we systematically investigate how to make unlearned models robust against such attacks. For the first time, we establish a connection between robust unlearning and sharpness-aware minimization (SAM) through a unified robust optimization framework, in an analogy to adversarial training designed to defend against adversarial attacks. Our analysis for SAM reveals that smoothness optimization plays a pivotal role in mitigating relearning attacks. Thus, we further explore diverse smoothing strategies to enhance unlearning robustness. Extensive experiments on benchmark datasets, including WMDP and MUSE, demonstrate that SAM and other smoothness optimization approaches consistently improve the resistance of LLM unlearning to relearning attacks. Notably, smoothness-enhanced unlearning also helps defend against (input-level) jailbreaking attacks, broadening our proposal's impact in robustifying LLM unlearning. Codes are available at https://github.com/OPTML-Group/Unlearn-Smooth.
♻ ☆ Towards Understanding the Influence of Training Samples on Explanations IJCAI 2024
Explainable AI (XAI) is widely used to analyze AI systems' decision-making, such as providing counterfactual explanations for recourse. When unexpected explanations occur, users may want to understand the training data properties shaping them. Under the umbrella of data valuation, first approaches have been proposed that estimate the influence of data samples on a given model. This process not only helps determine the data's value, but also offers insights into how individual, potentially noisy, or misleading examples affect a model, which is crucial for interpretable AI. In this work, we apply the concept of data valuation to the significant area of model evaluations, focusing on how individual training samples impact a model's internal reasoning rather than the predictive performance only. Hence, we introduce the novel problem of identifying training samples shaping a given explanation or related quantity, and investigate the particular case of the cost of computational recourse. We propose an algorithm to identify such influential samples and conduct extensive empirical evaluations in two case studies.
comment: Extended version of the paper accepted at the "Workshop on Explainable Artificial Intelligence (XAI)" at IJCAI 2024
♻ ☆ Neuromorphic Principles for Efficient Large Language Models on Intel Loihi 2 ICLR
Large language models (LLMs) deliver impressive performance but require large amounts of energy. In this work, we present a MatMul-free LLM architecture adapted for Intel's neuromorphic processor, Loihi 2. Our approach leverages Loihi 2's support for low-precision, event-driven computation and stateful processing. Our hardware-aware quantized model on GPU demonstrates that a 370M parameter MatMul-free model can be quantized with no accuracy loss. Based on preliminary results, we report up to 3x higher throughput with 2x less energy, compared to transformer-based LLMs on an edge GPU, with significantly better scaling. Further hardware optimizations will increase throughput and decrease energy consumption. These results show the potential of neuromorphic hardware for efficient inference and pave the way for efficient reasoning models capable of generating complex, long-form text rapidly and cost-effectively.
comment: Accepted to International Conference on Learning Representations (ICLR) Workshop on Scalable Optimization for Efficient and Adaptive Foundation Models (SCOPE)
♻ ☆ Probabilistic Shielding for Safe Reinforcement Learning AAAI 2025
In real-life scenarios, a Reinforcement Learning (RL) agent aiming to maximise their reward, must often also behave in a safe manner, including at training time. Thus, much attention in recent years has been given to Safe RL, where an agent aims to learn an optimal policy among all policies that satisfy a given safety constraint. However, strict safety guarantees are often provided through approaches based on linear programming, and thus have limited scaling. In this paper we present a new, scalable method, which enjoys strict formal guarantees for Safe RL, in the case where the safety dynamics of the Markov Decision Process (MDP) are known, and safety is defined as an undiscounted probabilistic avoidance property. Our approach is based on state-augmentation of the MDP, and on the design of a shield that restricts the actions available to the agent. We show that our approach provides a strict formal safety guarantee that the agent stays safe at training and test time. Furthermore, we demonstrate that our approach is viable in practice through experimental evaluation.
comment: 13 pages, 3 figures, Conference: AAAI 2025
♻ ☆ Practical multi-fidelity machine learning: fusion of deterministic and Bayesian models
Multi-fidelity machine learning methods address the accuracy-efficiency trade-off by integrating scarce, resource-intensive high-fidelity data with abundant but less accurate low-fidelity data. We propose a practical multi-fidelity strategy for problems spanning low- and high-dimensional domains, integrating a non-probabilistic regression model for the low-fidelity with a Bayesian model for the high-fidelity. The models are trained in a staggered scheme, where the low-fidelity model is transfer-learned to the high-fidelity data and a Bayesian model is trained to learn the residual between the data and the transfer-learned model. This three-model strategy -- deterministic low-fidelity, transfer-learning, and Bayesian residual -- leads to a prediction that includes uncertainty quantification for noisy and noiseless multi-fidelity data. The strategy is general and unifies the topic, highlighting the expressivity trade-off between the transfer-learning and Bayesian models (a complex transfer-learning model leads to a simpler Bayesian model, and vice versa). We propose modeling choices for two scenarios, and argue in favor of using a linear transfer-learning model that fuses 1) kernel ridge regression for low-fidelity with Gaussian processes for high-fidelity; or 2) deep neural network for low-fidelity with a Bayesian neural network for high-fidelity. We demonstrate the effectiveness and efficiency of the proposed strategies and contrast them with the state-of-the-art based on various numerical examples and two engineering problems. The results indicate that the proposed approach achieves comparable performance in both mean and uncertainty estimation while significantly reducing training time for machine learning modeling in data-scarce scenarios. Moreover, in data-rich settings, it outperforms other multi-fidelity architectures by effectively mitigating overfitting.
comment: 39 Pages, 26 Figures
♻ ☆ Unsupervised Blind Joint Dereverberation and Room Acoustics Estimation with Diffusion Models
This paper presents an unsupervised method for single-channel blind dereverberation and room impulse response (RIR) estimation, called BUDDy. The algorithm is rooted in Bayesian posterior sampling: it combines a likelihood model enforcing fidelity to the reverberant measurement, and an anechoic speech prior implemented by an unconditional diffusion model. We design a parametric filter representing the RIR, with exponential decay for each frequency subband. Room acoustics estimation and speech dereverberation are jointly carried out, as the filter parameters are iteratively estimated and the speech utterance refined along the reverse diffusion trajectory. In a blind scenario where the RIR is unknown, BUDDy successfully performs speech dereverberation in various acoustic scenarios, significantly outperforming other blind unsupervised baselines. Unlike supervised methods, which often struggle to generalize, BUDDy seamlessly adapts to different acoustic conditions. This paper extends our previous work by offering new experimental results and insights into the algorithm's versatility. We demonstrate the robustness of our proposed method to new acoustic and speaker conditions, as well as its adaptability to high-resolution singing voice dereverberation, using both instrumental metrics and subjective listening evaluation. We study BUDDy's performance for RIR estimation and observe it surpasses a state-of-the-art supervised DNN-based estimator on mismatched acoustic conditions. Finally, we investigate the sensitivity of informed dereverberation methods to RIR estimation errors, thereby motivating the joint acoustic estimation and dereverberation design. Audio examples and code can be found online.
comment: Submitted to IEEE/ACM Transactions on Audio, Speech and Language Processing
♻ ☆ Investigating Data Hierarchies in Multifidelity Machine Learning for Excitation Energies
Recent progress in machine learning (ML) has made high-accuracy quantum chemistry (QC) calculations more accessible. Of particular interest are multifidelity machine learning (MFML) methods where training data from differing accuracies or fidelities are used. These methods usually employ a fixed scaling factor, $\gamma$, to relate the number of training samples across different fidelities, which reflects the cost and assumed sparsity of the data. This study investigates the impact of modifying $\gamma$ on model efficiency and accuracy for the prediction of vertical excitation energies using the QeMFi benchmark dataset. Further, this work introduces QC compute time informed scaling factors, denoted as $\theta$, that vary based on QC compute times at different fidelities. A novel error metric, error contours of MFML, is proposed to provide a comprehensive view of model error contributions from each fidelity. The results indicate that high model accuracy can be achieved with just 2 training samples at the target fidelity when a larger number of samples from lower fidelities are used. This is further illustrated through a novel concept, the $\Gamma$-curve, which compares model error against the time-cost of generating training samples, demonstrating that multifidelity models can achieve high accuracy while minimizing training data costs.
comment: Modified errors to be relative MAE. Transferability tests of training on QeMFi and testing on QUESTDB have now been added
♻ ☆ To FP8 and Back Again: Quantifying Reduced Precision Effects on LLM Training Stability
The massive computational costs associated with large language model (LLM) pretraining have spurred great interest in reduced-precision floating-point representations to accelerate the process. As a result, the BrainFloat16 (BF16) precision has become the de facto standard for LLM training, with hardware support included in recent generations of accelerators. This trend has gone even further in the latest processors, where FP8 has recently been introduced. However, prior experience with FP16, which was found to be less stable than BF16, raises concerns as to whether FP8, with even fewer bits than FP16, can be a cost-effective option for LLM training. We argue that reduced-precision training schemes must have similar training stability and hyperparameter sensitivities to their higher-precision counterparts in order to be cost-effective. However, we find that currently available methods for FP8 training are not robust enough to allow their use as economical replacements. This prompts us to investigate the stability of reduced-precision LLM training in terms of robustness across random seeds, learning rates, and datasets. To this end, we propose new evaluation techniques and a new metric for quantifying loss landscape sharpness in autoregressive language models. By simulating incremental bit reductions in floating-point representations, we analyze the relationship between representational power and training stability with the intent of aiding future research into the field.
♻ ☆ Benchmarking Data Efficiency in $Δ$-ML and Multifidelity Models for Quantum Chemistry
The development of machine learning (ML) methods has made quantum chemistry (QC) calculations more accessible by reducing the compute cost incurred in conventional QC methods. This has since been translated into the overhead cost of generating training data. Increased work in reducing the cost of generating training data resulted in the development of $\Delta$-ML and multifidelity machine learning methods which use data at more than one QC level of accuracy, or fidelity. This work compares the data costs associated with $\Delta$-ML, multifidelity machine learning (MFML), and optimized MFML (o-MFML) in contrast with a newly introduced Multifidelity$\Delta$-Machine Learning (MF$\Delta$ML) method for the prediction of ground state energies, vertical excitation energies, and the magnitude of electronic contribution of molecular dipole moments from the multifidelity benchmark dataset QeMFi. This assessment is made on the basis of training data generation cost associated with each model and is compared with the single fidelity kernel ridge regression (KRR) case. The results indicate that the use of multifidelity methods surpasses the standard $\Delta$-ML approaches in cases of a large number of predictions. For applications which require only a few evaluations to be made using ML models, while the $\Delta$-ML method might be favored, the MF$\Delta$ML method is shown to be more efficient.
comment: Supplementary sections S1-S4 and FIG.~S1-S4, Table S1.Work modified to include benchmarks for 3 more QC properties: first and second excitation energies, magnitude of electronic dipole moment
♻ ☆ The HalluRAG Dataset: Detecting Closed-Domain Hallucinations in RAG Applications Using an LLM's Internal States
Detecting hallucinations in large language models (LLMs) is critical for enhancing their reliability and trustworthiness. Most research focuses on hallucinations as deviations from information seen during training. However, the opaque nature of an LLM's parametric knowledge complicates the understanding of why generated texts appear ungrounded: The LLM might not have picked up the necessary knowledge from large and often inaccessible datasets, or the information might have been changed or contradicted during further training. Our focus is on hallucinations involving information not used in training, which we determine by using recency to ensure the information emerged after a cut-off date. This study investigates these hallucinations by detecting them at sentence level using different internal states of various LLMs. We present HalluRAG, a dataset designed to train classifiers on these hallucinations. Depending on the model and quantization, MLPs trained on HalluRAG detect hallucinations with test accuracies ranging up to 75 %, with Mistral-7B-Instruct-v0.1 achieving the highest test accuracies. Our results show that IAVs detect hallucinations as effectively as CEVs and reveal that answerable and unanswerable prompts are encoded differently as separate classifiers for these categories improved accuracy. However, HalluRAG showed some limited generalizability, advocating for more diversity in datasets on hallucinations.
comment: 19 pages, 3 figures
♻ ☆ Scale generalisation properties of extended scale-covariant and scale-invariant Gaussian derivative networks on image datasets with spatial scaling variations
This paper presents an in-depth analysis of the scale generalisation properties of the scale-covariant and scale-invariant Gaussian derivative networks, complemented with both conceptual and algorithmic extensions. For this purpose, Gaussian derivative networks (GaussDerNets) are evaluated on new rescaled versions of the Fashion-MNIST and the CIFAR-10 datasets, with spatial scaling variations over a factor of 4 in the testing data, that are not present in the training data. Additionally, evaluations on the previously existing STIR datasets show that the GaussDerNets achieve better scale generalisation than previously reported for these datasets for other types of deep networks. We first experimentally demonstrate that the GaussDerNets have quite good scale generalisation properties on the new datasets, and that average pooling of feature responses over scales may sometimes also lead to better results than the previously used approach of max pooling over scales. Then, we demonstrate that using a spatial max pooling mechanism after the final layer enables localisation of non-centred objects in image domain, with maintained scale generalisation properties. We also show that regularisation during training, by applying dropout across the scale channels, referred to as scale-channel dropout, improves both the performance and the scale generalisation. In additional ablation studies, we demonstrate that discretisations of GaussDerNets, based on the discrete analogue of the Gaussian kernel in combination with central difference operators, perform best or among the best, compared to a set of other discrete approximations of the Gaussian derivative kernels. Finally, by visualising the activation maps and the learned receptive fields, we demonstrate that the GaussDerNets have very good explainability properties.
comment: 52 pages, 24 figures, 18 tables
♻ ☆ Inverting Transformer-based Vision Models
Understanding the mechanisms underlying deep neural networks in computer vision remains a fundamental challenge. While many previous approaches have focused on visualizing intermediate representations within deep neural networks, particularly convolutional neural networks, these techniques have yet to be thoroughly explored in transformer-based vision models. In this study, we apply a modular approach of training inverse models to reconstruct input images from intermediate layers within a Detection Transformer and a Vision Transformer, showing that this approach is efficient and feasible. Through qualitative and quantitative evaluations of reconstructed images, we generate insights into the underlying mechanisms of these architectures, highlighting their similarities and differences in terms of contextual shape and preservation of image details, inter-layer correlation, and robustness to color perturbations. Our analysis illustrates how these properties emerge within the models, contributing to a deeper understanding of transformer-based vision models. The code for reproducing our experiments is available at github.com/wiskott-lab/inverse-tvm.
♻ ☆ Rank Reduction Autoencoders
The choice of an appropriate bottleneck dimension and the application of effective regularization are both essential for Autoencoders to learn meaningful representations from unlabeled data. In this paper, we introduce a new class of deterministic autoencoders, Rank Reduction Autoencoders (RRAEs), which regularize their latent spaces by employing a truncated singular value decomposition (SVD) during training. In RRAEs, the bottleneck is defined by the rank of the latent matrix, thereby alleviating the dependence of the encoder/decoder architecture on the bottleneck size. This approach enabled us to propose an adaptive algorithm (aRRAEs) that efficiently determines the optimal bottleneck size during training. We empirically demonstrate that both RRAEs and aRRAEs are stable, scalable, and reliable, as they do not introduce any additional training hyperparameters. We evaluate our proposed architecture on a synthetic data set, as well as on MNIST, Fashion MNIST, and CelebA. Our results show that RRAEs offer several advantages over Vanilla AEs with both large and small latent spaces, and outperform other regularizing AE architectures.
♻ ☆ Nonparametric estimation of Hawkes processes with RKHSs AISTATS 2025
This paper addresses nonparametric estimation of nonlinear multivariate Hawkes processes, where the interaction functions are assumed to lie in a reproducing kernel Hilbert space (RKHS). Motivated by applications in neuroscience, the model allows complex interaction functions, in order to express exciting and inhibiting effects, but also a combination of both (which is particularly interesting to model the refractory period of neurons), and considers in return that conditional intensities are rectified by the ReLU function. The latter feature incurs several methodological challenges, for which workarounds are proposed in this paper. In particular, it is shown that a representer theorem can be obtained for approximated versions of the log-likelihood and the least-squares criteria. Based on it, we propose an estimation method, that relies on two common approximations (of the ReLU function and of the integral operator). We provide a bound that controls the impact of these approximations. Numerical results on synthetic data confirm this fact as well as the good asymptotic behavior of the proposed estimator. It also shows that our method achieves a better performance compared to related nonparametric estimation techniques and suits neuronal applications.
comment: AISTATS 2025
♻ ☆ A Note on Estimation Error Bound and Grouping Effect of Transfer Elastic Net
The Transfer Elastic Net is an estimation method for linear regression models that combines $\ell_1$ and $\ell_2$ norm penalties to facilitate knowledge transfer. In this study, we derive a non-asymptotic $\ell_2$ norm estimation error bound for the estimator and discuss scenarios where the Transfer Elastic Net effectively works. Furthermore, we examine situations where it exhibits the grouping effect, which states that the estimates corresponding to highly correlated predictors have a small difference.
♻ ☆ KL-geodesics flow matching with a novel sampling scheme
Non-autoregressive language models generate all tokens simultaneously, offering potential speed advantages over traditional autoregressive models, but they face challenges in modeling the complex dependencies inherent in text data. In this work, we investigate a conditional flow matching approach for text generation. We represent tokens as one-hot vectors in a \(V\)-dimensional simplex and utilize geodesics under the Kullback-Leibler (KL) divergence, which correspond to linear interpolation in logit space. We provide a theoretical justification that maximizing the conditional likelihood \(P_{\theta}(x_1 \mid x_t, t)\) yields the exact flow matching velocity under logit interpolation. To address the suboptimal performance of basic inference, we propose a novel empirical sampling scheme that iteratively samples from the conditional distribution and introduces additional noise, significantly improving results despite lacking full theoretical underpinnings. Furthermore, we propose a hybrid inference method that combines the basic approach with the sampling scheme. This method demonstrates superior performance on both conditional and unconditional text generation experiments compared to previous SOTA method for discrete flow matching.
♻ ☆ Understanding and Reducing the Class-Dependent Effects of Data Augmentation with A Two-Player Game Approach
Data augmentation is widely applied and has shown its benefits in different machine learning tasks. However, as recently observed, it may have an unfair effect in multi-class classification. While data augmentation generally improves the overall performance (and therefore is beneficial for many classes), it can actually be detrimental for other classes, which can be problematic in some application domains. In this paper, to counteract this phenomenon, we propose CLAM, a CLAss-dependent Multiplicative-weights method. To derive it, we first formulate the training of a classifier as a non-linear optimization problem that aims at simultaneously maximizing the individual class performances and balancing them. By rewriting this optimization problem as an adversarial two-player game, we propose a novel multiplicative weight algorithm, for which we prove the convergence. Interestingly, our formulation also reveals that the class-dependent effects of data augmentation is not due to data augmentation only, but is in fact a general phenomenon. Our empirical results over five datasets demonstrate that the performance of learned classifiers is indeed more fairly distributed over classes, with only limited impact on the average accuracy.
♻ ☆ Expert Race: A Flexible Routing Strategy for Scaling Diffusion Transformer with Mixture of Experts
Diffusion models have emerged as mainstream framework in visual generation. Building upon this success, the integration of Mixture of Experts (MoE) methods has shown promise in enhancing model scalability and performance. In this paper, we introduce Race-DiT, a novel MoE model for diffusion transformers with a flexible routing strategy, Expert Race. By allowing tokens and experts to compete together and select the top candidates, the model learns to dynamically assign experts to critical tokens. Additionally, we propose per-layer regularization to address challenges in shallow layer learning, and router similarity loss to prevent mode collapse, ensuring better expert utilization. Extensive experiments on ImageNet validate the effectiveness of our approach, showcasing significant performance gains while promising scaling properties.
♻ ☆ A Closer Look at Time Steps is Worthy of Triple Speed-Up for Diffusion Model Training
Training diffusion models is always a computation-intensive task. In this paper, we introduce a novel speed-up method for diffusion model training, called, which is based on a closer look at time steps. Our key findings are: i) Time steps can be empirically divided into acceleration, deceleration, and convergence areas based on the process increment. ii) These time steps are imbalanced, with many concentrated in the convergence area. iii) The concentrated steps provide limited benefits for diffusion training. To address this, we design an asymmetric sampling strategy that reduces the frequency of steps from the convergence area while increasing the sampling probability for steps from other areas. Additionally, we propose a weighting strategy to emphasize the importance of time steps with rapid-change process increments. As a plug-and-play and architecture-agnostic approach, SpeeD consistently achieves 3-times acceleration across various diffusion architectures, datasets, and tasks. Notably, due to its simple design, our approach significantly reduces the cost of diffusion model training with minimal overhead. Our research enables more researchers to train diffusion models at a lower cost.
♻ ☆ Conditional Shift-Robust Conformal Prediction for Graph Neural Network
Graph Neural Networks (GNNs) have emerged as potent tools for predicting outcomes in graph-structured data. Despite their efficacy, a significant drawback of GNNs lies in their limited ability to provide robust uncertainty estimates, posing challenges to their reliability in contexts where errors carry significant consequences. Moreover, GNNs typically excel in in-distribution settings, assuming that training and test data follow identical distributions a condition often unmet in real world graph data scenarios. In this article, we leverage conformal prediction, a widely recognized statistical technique for quantifying uncertainty by transforming predictive model outputs into prediction sets, to address uncertainty quantification in GNN predictions amidst conditional shift\footnote{Representing the change in conditional probability distribution \(P(label|input)\) from source domain to target domain.} in graph-based semi-supervised learning (SSL). Additionally, we propose a novel loss function aimed at refining model predictions by minimizing conditional shift in latent stages. Termed Conditional Shift Robust (CondSR) conformal prediction for GNNs, our approach CondSR is model-agnostic and adaptable to various classification models. We validate the effectiveness of our method on standard graph benchmark datasets, integrating it with state-of-the-art GNNs in node classification tasks. Comprehensive evaluations demonstrate that our approach consistently achieves any predefined target marginal coverage, enhances the accuracy of state of the art GNN models by up to 12\% under conditional shift, and reduces the prediction set size by up to 48\%. The code implementation is publicly available for further exploration and experimentation.
comment: 15 pages, 3 figures, 4 tables
♻ ☆ VideoRefer Suite: Advancing Spatial-Temporal Object Understanding with Video LLM
Video Large Language Models (Video LLMs) have recently exhibited remarkable capabilities in general video understanding. However, they mainly focus on holistic comprehension and struggle with capturing fine-grained spatial and temporal details. Besides, the lack of high-quality object-level video instruction data and a comprehensive benchmark further hinders their advancements. To tackle these challenges, we introduce the VideoRefer Suite to empower Video LLM for finer-level spatial-temporal video understanding, i.e., enabling perception and reasoning on any objects throughout the video. Specially, we thoroughly develop VideoRefer Suite across three essential aspects: dataset, model, and benchmark. Firstly, we introduce a multi-agent data engine to meticulously curate a large-scale, high-quality object-level video instruction dataset, termed VideoRefer-700K. Next, we present the VideoRefer model, which equips a versatile spatial-temporal object encoder to capture precise regional and sequential representations. Finally, we meticulously create a VideoRefer-Bench to comprehensively assess the spatial-temporal understanding capability of a Video LLM, evaluating it across various aspects. Extensive experiments and analyses demonstrate that our VideoRefer model not only achieves promising performance on video referring benchmarks but also facilitates general video understanding capabilities.
comment: 17 pages, 14 figures, technical report
♻ ☆ Masking meets Supervision: A Strong Learning Alliance CVPR 2025
Pre-training with random masked inputs has emerged as a novel trend in self-supervised training. However, supervised learning still faces a challenge in adopting masking augmentations, primarily due to unstable training. In this paper, we propose a novel way to involve masking augmentations dubbed Masked Sub-branch (MaskSub). MaskSub consists of the main-branch and sub-branch, the latter being a part of the former. The main-branch undergoes conventional training recipes, while the sub-branch merits intensive masking augmentations, during training. MaskSub tackles the challenge by mitigating adverse effects through a relaxed loss function similar to a self-distillation loss. Our analysis shows that MaskSub improves performance, with the training loss converging faster than in standard training, which suggests our method stabilizes the training process. We further validate MaskSub across diverse training scenarios and models, including DeiT-III training, MAE finetuning, CLIP finetuning, BERT training, and hierarchical architectures (ResNet and Swin Transformer). Our results show that MaskSub consistently achieves impressive performance gains across all the cases. MaskSub provides a practical and effective solution for introducing additional regularization under various training recipes. Code available at https://github.com/naver-ai/augsub
comment: Accepted to CVPR 2025
♻ ☆ GFlowVLM: Enhancing Multi-step Reasoning in Vision-Language Models with Generative Flow Networks
Vision-Language Models (VLMs) have recently shown promising advancements in sequential decision-making tasks through task-specific fine-tuning. However, common fine-tuning methods, such as Supervised Fine-Tuning (SFT) and Reinforcement Learning (RL) techniques like Proximal Policy Optimization (PPO), present notable limitations: SFT assumes Independent and Identically Distributed (IID) data, while PPO focuses on maximizing cumulative rewards. These limitations often restrict solution diversity and hinder generalization in multi-step reasoning tasks. To address these challenges, we introduce a novel framework, GFlowVLM, a framework that fine-tune VLMs using Generative Flow Networks (GFlowNets) to promote generation of diverse solutions for complex reasoning tasks. GFlowVLM models the environment as a non-Markovian decision process, allowing it to capture long-term dependencies essential for real-world applications. It takes observations and task descriptions as inputs to prompt chain-of-thought (CoT) reasoning which subsequently guides action selection. We use task based rewards to fine-tune VLM with GFlowNets. This approach enables VLMs to outperform prior fine-tuning methods, including SFT and RL. Empirical results demonstrate the effectiveness of GFlowVLM on complex tasks such as card games (NumberLine, BlackJack) and embodied planning tasks (ALFWorld), showing enhanced training efficiency, solution diversity, and stronger generalization capabilities across both in-distribution and out-of-distribution scenarios.
♻ ☆ DistDNAS: Search Efficient Feature Interactions within 2 Hours
Search efficiency and serving efficiency are two major axes in building feature interactions and expediting the model development process in recommender systems. On large-scale benchmarks, searching for the optimal feature interaction design requires extensive cost due to the sequential workflow on the large volume of data. In addition, fusing interactions of various sources, orders, and mathematical operations introduces potential conflicts and additional redundancy toward recommender models, leading to sub-optimal trade-offs in performance and serving cost. In this paper, we present DistDNAS as a neat solution to brew swift and efficient feature interaction design. DistDNAS proposes a supernet to incorporate interaction modules of varying orders and types as a search space. To optimize search efficiency, DistDNAS distributes the search and aggregates the choice of optimal interaction modules on varying data dates, achieving over 25x speed-up and reducing search cost from 2 days to 2 hours. To optimize serving efficiency, DistDNAS introduces a differentiable cost-aware loss to penalize the selection of redundant interaction modules, enhancing the efficiency of discovered feature interactions in serving. We extensively evaluate the best models crafted by DistDNAS on a 1TB Criteo Terabyte dataset. Experimental evaluations demonstrate 0.001 AUC improvement and 60% FLOPs saving over current state-of-the-art CTR models.
♻ ☆ Unveil Inversion and Invariance in Flow Transformer for Versatile Image Editing CVPR 2025
Leveraging the large generative prior of the flow transformer for tuning-free image editing requires authentic inversion to project the image into the model's domain and a flexible invariance control mechanism to preserve non-target contents. However, the prevailing diffusion inversion performs deficiently in flow-based models, and the invariance control cannot reconcile diverse rigid and non-rigid editing tasks. To address these, we systematically analyze the \textbf{inversion and invariance} control based on the flow transformer. Specifically, we unveil that the Euler inversion shares a similar structure to DDIM yet is more susceptible to the approximation error. Thus, we propose a two-stage inversion to first refine the velocity estimation and then compensate for the leftover error, which pivots closely to the model prior and benefits editing. Meanwhile, we propose the invariance control that manipulates the text features within the adaptive layer normalization, connecting the changes in the text prompt to image semantics. This mechanism can simultaneously preserve the non-target contents while allowing rigid and non-rigid manipulation, enabling a wide range of editing types such as visual text, quantity, facial expression, etc. Experiments on versatile scenarios validate that our framework achieves flexible and accurate editing, unlocking the potential of the flow transformer for versatile image editing.
comment: CVPR 2025 Page: https://pengchengpcx.github.io/EditFT/
♻ ☆ BioMamba: Leveraging Spectro-Temporal Embedding in Bidirectional Mamba for Enhanced Biosignal Classification
Biological signals, such as electroencephalograms (EEGs) and electrocardiograms (ECGs), play a pivotal role in numerous clinical practices, such as diagnosing brain and cardiac arrhythmic diseases. Existing methods for biosignal classification rely on Attention-based frameworks with dense Feed Forward layers, which lead to inefficient learning, high computational overhead, and suboptimal performance. In this work, we introduce BioMamba, a Spectro-Temporal Embedding strategy applied to the Bidirectional Mamba framework with Sparse Feed Forward layers to enable effective learning of biosignal sequences. By integrating these three key components, BioMamba effectively addresses the limitations of existing methods. Extensive experiments demonstrate that BioMamba significantly outperforms state-of-the-art methods with marked improvement in classification performance. The advantages of the proposed BioMamba include (1) Reliability: BioMamba consistently delivers robust results, confirmed across six evaluation metrics. (2) Efficiency: We assess both model and training efficiency, the BioMamba demonstrates computational effectiveness by reducing model size and resource consumption compared to existing approaches. (3) Generality: With the capacity to effectively classify a diverse set of tasks, BioMamba demonstrates adaptability and effectiveness across various domains and applications.
comment: Biological signals
♻ ☆ RayFlow: Instance-Aware Diffusion Acceleration via Adaptive Flow Trajectories CVPR 2025
Diffusion models have achieved remarkable success across various domains. However, their slow generation speed remains a critical challenge. Existing acceleration methods, while aiming to reduce steps, often compromise sample quality, controllability, or introduce training complexities. Therefore, we propose RayFlow, a novel diffusion framework that addresses these limitations. Unlike previous methods, RayFlow guides each sample along a unique path towards an instance-specific target distribution. This method minimizes sampling steps while preserving generation diversity and stability. Furthermore, we introduce Time Sampler, an importance sampling technique to enhance training efficiency by focusing on crucial timesteps. Extensive experiments demonstrate RayFlow's superiority in generating high-quality images with improved speed, control, and training efficiency compared to existing acceleration techniques.
comment: 23 pages, 5 figures, CVPR 2025
♻ ☆ XXLTraffic: Expanding and Extremely Long Traffic forecasting beyond test adaptation
Traffic forecasting is crucial for smart cities and intelligent transportation initiatives, where deep learning has made significant progress in modeling complex spatio-temporal patterns in recent years. However, current public datasets have limitations in reflecting the distribution shift nature of real-world scenarios, characterized by continuously evolving infrastructures, varying temporal distributions, and long temporal gaps due to sensor downtimes or changes in traffic patterns. These limitations inevitably restrict the practical applicability of existing traffic forecasting datasets. To bridge this gap, we present XXLTraffic, largest available public traffic dataset with the longest timespan collected from Los Angeles, USA, and New South Wales, Australia, curated to support research in extremely long forecasting beyond test adaptation. Our benchmark includes both typical time-series forecasting settings with hourly and daily aggregated data and novel configurations that introduce gaps and down-sample the training size to better simulate practical constraints. We anticipate the new XXLTraffic will provide a fresh perspective for the time-series and traffic forecasting communities. It would also offer a robust platform for developing and evaluating models designed to tackle the extremely long forecasting problems beyond test adaptation. Our dataset supplements existing spatio-temporal data resources and leads to new research directions in this domain.
♻ ☆ Polysemanticity and Capacity in Neural Networks
Individual neurons in neural networks often represent a mixture of unrelated features. This phenomenon, called polysemanticity, can make interpreting neural networks more difficult and so we aim to understand its causes. We propose doing so through the lens of feature \emph{capacity}, which is the fractional dimension each feature consumes in the embedding space. We show that in a toy model the optimal capacity allocation tends to monosemantically represent the most important features, polysemantically represent less important features (in proportion to their impact on the loss), and entirely ignore the least important features. Polysemanticity is more prevalent when the inputs have higher kurtosis or sparsity and more prevalent in some architectures than others. Given an optimal allocation of capacity, we go on to study the geometry of the embedding space. We find a block-semi-orthogonal structure, with differing block sizes in different models, highlighting the impact of model architecture on the interpretability of its neurons.
comment: 22 pages, 7 figures. Improved notation and corrected an error in the description of the most general efficient matrices
Computation and Language 79
☆ Exploring Training and Inference Scaling Laws in Generative Retrieval
Generative retrieval has emerged as a novel paradigm that leverages large language models (LLMs) to autoregressively generate document identifiers. Although promising, the mechanisms that underpin its performance and scalability remain largely unclear. We conduct a systematic investigation of training and inference scaling laws in generative retrieval, exploring how model size, training data scale, and inference-time compute jointly influence retrieval performance. To address the lack of suitable metrics, we propose a novel evaluation measure inspired by contrastive entropy and generation loss, providing a continuous performance signal that enables robust comparisons across diverse generative retrieval methods. Our experiments show that n-gram-based methods demonstrate strong alignment with both training and inference scaling laws, especially when paired with larger LLMs. Furthermore, increasing inference computation yields substantial performance gains, revealing that generative retrieval can significantly benefit from higher compute budgets at inference. Across these settings, LLaMA models consistently outperform T5 models, suggesting a particular advantage for larger decoder-only models in generative retrieval. Taken together, our findings underscore that model sizes, data availability, and inference computation interact to unlock the full potential of generative retrieval, offering new insights for designing and optimizing future systems.
☆ xKV: Cross-Layer SVD for KV-Cache Compression
Large Language Models (LLMs) with long context windows enable powerful applications but come at the cost of high memory consumption to store the Key and Value states (KV-Cache). Recent studies attempted to merge KV-cache from multiple layers into shared representations, yet these approaches either require expensive pretraining or rely on assumptions of high per-token cosine similarity across layers which generally does not hold in practice. We find that the dominant singular vectors are remarkably well-aligned across multiple layers of the KV-Cache. Exploiting this insight, we propose xKV, a simple post-training method that applies Singular Value Decomposition (SVD) on the KV-Cache of grouped layers. xKV consolidates the KV-Cache of multiple layers into a shared low-rank subspace, significantly reducing KV-Cache sizes. Through extensive evaluations on the RULER long-context benchmark with widely-used LLMs (e.g., Llama-3.1 and Qwen2.5), xKV achieves up to 6.8x higher compression rates than state-of-the-art inter-layer technique while improving accuracy by 2.7%. Moreover, xKV is compatible with the emerging Multi-Head Latent Attention (MLA) (e.g., DeepSeek-Coder-V2), yielding a notable 3x compression rates on coding tasks without performance degradation. These results highlight xKV's strong capability and versatility in addressing memory bottlenecks for long-context LLM inference. Our code is publicly available at: https://github.com/abdelfattah-lab/xKV.
☆ SimpleRL-Zoo: Investigating and Taming Zero Reinforcement Learning for Open Base Models in the Wild
DeepSeek-R1 has shown that long chain-of-thought (CoT) reasoning can naturally emerge through a simple reinforcement learning (RL) framework with rule-based rewards, where the training may directly start from the base models-a paradigm referred to as zero RL training. Most recent efforts to reproduce zero RL training have primarily focused on the Qwen2.5 model series, which may not be representative as we find the base models already exhibit strong instruction-following and self-reflection abilities. In this work, we investigate zero RL training across 10 diverse base models, spanning different families and sizes including LLama3-8B, Mistral-7B/24B, DeepSeek-Math-7B, Qwen2.5-math-7B, and all Qwen2.5 models from 0.5B to 32B. Leveraging several key design strategies-such as adjusting format reward and controlling query difficulty-we achieve substantial improvements in both reasoning accuracy and response length across most settings. However, by carefully monitoring the training dynamics, we observe that different base models exhibit distinct patterns during training. For instance, the increased response length does not always correlate with the emergence of certain cognitive behaviors such as verification (i.e., the "aha moment"). Notably, we observe the "aha moment" for the first time in small models not from the Qwen family. We share the key designs that enable successful zero RL training, along with our findings and practices. To facilitate further research, we open-source the code, models, and analysis tools.
☆ AgentDropout: Dynamic Agent Elimination for Token-Efficient and High-Performance LLM-Based Multi-Agent Collaboration
Multi-agent systems (MAS) based on large language models (LLMs) have demonstrated significant potential in collaborative problem-solving. However, they still face substantial challenges of low communication efficiency and suboptimal task performance, making the careful design of the agents' communication topologies particularly important. Inspired by the management theory that roles in an efficient team are often dynamically adjusted, we propose AgentDropout, which identifies redundant agents and communication across different communication rounds by optimizing the adjacency matrices of the communication graphs and eliminates them to enhance both token efficiency and task performance. Compared to state-of-the-art methods, AgentDropout achieves an average reduction of 21.6% in prompt token consumption and 18.4% in completion token consumption, along with a performance improvement of 1.14 on the tasks. Furthermore, the extended experiments demonstrate that AgentDropout achieves notable domain transferability and structure robustness, revealing its reliability and effectiveness. We release our code at https://github.com/wangzx1219/AgentDropout.
☆ Toward building next-generation Geocoding systems: a systematic review
Geocoding systems are widely used in both scientific research for spatial analysis and everyday life through location-based services. The quality of geocoded data significantly impacts subsequent processes and applications, underscoring the need for next-generation systems. In response to this demand, this review first examines the evolving requirements for geocoding inputs and outputs across various scenarios these systems must address. It then provides a detailed analysis of how to construct such systems by breaking them down into key functional components and reviewing a broad spectrum of existing approaches, from traditional rule-based methods to advanced techniques in information retrieval, natural language processing, and large language models. Finally, we identify opportunities to improve next-generation geocoding systems in light of recent technological advances.
☆ I Have Covered All the Bases Here: Interpreting Reasoning Features in Large Language Models via Sparse Autoencoders
Large Language Models (LLMs) have achieved remarkable success in natural language processing. Recent advances have led to the developing of a new class of reasoning LLMs; for example, open-source DeepSeek-R1 has achieved state-of-the-art performance by integrating deep thinking and complex reasoning. Despite these impressive capabilities, the internal reasoning mechanisms of such models remain unexplored. In this work, we employ Sparse Autoencoders (SAEs), a method to learn a sparse decomposition of latent representations of a neural network into interpretable features, to identify features that drive reasoning in the DeepSeek-R1 series of models. First, we propose an approach to extract candidate ''reasoning features'' from SAE representations. We validate these features through empirical analysis and interpretability methods, demonstrating their direct correlation with the model's reasoning abilities. Crucially, we demonstrate that steering these features systematically enhances reasoning performance, offering the first mechanistic account of reasoning in LLMs. Code available at https://github.com/AIRI-Institute/SAE-Reasoning
☆ Reasoning to Learn from Latent Thoughts
Compute scaling for language model (LM) pretraining has outpaced the growth of human-written texts, leading to concerns that data will become the bottleneck to LM scaling. To continue scaling pretraining in this data-constrained regime, we propose that explicitly modeling and inferring the latent thoughts that underlie the text generation process can significantly improve pretraining data efficiency. Intuitively, our approach views web text as the compressed final outcome of a verbose human thought process and that the latent thoughts contain important contextual knowledge and reasoning steps that are critical to data-efficient learning. We empirically demonstrate the effectiveness of our approach through data-constrained continued pretraining for math. We first show that synthetic data approaches to inferring latent thoughts significantly improve data efficiency, outperforming training on the same amount of raw data (5.7\% $\rightarrow$ 25.4\% on MATH). Furthermore, we demonstrate latent thought inference without a strong teacher, where an LM bootstraps its own performance by using an EM algorithm to iteratively improve the capability of the trained LM and the quality of thought-augmented pretraining data. We show that a 1B LM can bootstrap its performance across at least three iterations and significantly outperform baselines trained on raw data, with increasing gains from additional inference compute when performing the E-step. The gains from inference scaling and EM iterations suggest new opportunities for scaling data-constrained pretraining.
☆ EconEvals: Benchmarks and Litmus Tests for LLM Agents in Unknown Environments
We develop benchmarks for LLM agents that act in, learn from, and strategize in unknown environments, the specifications of which the LLM agent must learn over time from deliberate exploration. Our benchmarks consist of decision-making tasks derived from key problems in economics. To forestall saturation, the benchmark tasks are synthetically generated with scalable difficulty levels. Additionally, we propose litmus tests, a new kind of quantitative measure for LLMs and LLM agents. Unlike benchmarks, litmus tests quantify differences in character, values, and tendencies of LLMs and LLM agents, by considering their behavior when faced with tradeoffs (e.g., efficiency versus equality) where there is no objectively right or wrong behavior. Overall, our benchmarks and litmus tests assess the abilities and tendencies of LLM agents in tackling complex economic problems in diverse settings spanning procurement, scheduling, task allocation, and pricing -- applications that should grow in importance as such agents are further integrated into the economy.
☆ REALM: A Dataset of Real-World LLM Use Cases
Large Language Models, such as the GPT series, have driven significant industrial applications, leading to economic and societal transformations. However, a comprehensive understanding of their real-world applications remains limited. To address this, we introduce REALM, a dataset of over 94,000 LLM use cases collected from Reddit and news articles. REALM captures two key dimensions: the diverse applications of LLMs and the demographics of their users. It categorizes LLM applications and explores how users' occupations relate to the types of applications they use. By integrating real-world data, REALM offers insights into LLM adoption across different domains, providing a foundation for future research on their evolving societal roles. A dedicated dashboard https://realm-e7682.web.app/ presents the data.
comment: 9 pages, 5 figures
☆ BitDecoding: Unlocking Tensor Cores for Long-Context LLMs Decoding with Low-Bit KV Cache
The growing adoption of long-context Large Language Models (LLMs) has introduced significant memory and computational challenges in autoregressive decoding due to the expanding Key-Value (KV) cache. KV cache quantization has emerged as a promising solution, with prior work showing that 4-bit or even 2-bit quantization can maintain model accuracy while reducing memory costs. However, despite these benefits, preliminary implementations for the low-bit KV cache struggle to deliver the expected speedup due to quantization and dequantization overheads and the lack of Tensor Cores utilization. In this work, we propose BitDecoding, a GPU-optimized framework that unlocks Tensor Cores for efficient decoding with low-bit KV cache. Efficiently leveraging Tensor Cores for low-bit KV cache is challenging due to the dynamic nature of KV cache generation at each decoding step. BitDecoding addresses these challenges with a Tensor Cores-Centric BitFusion Scheme that ensures data layout compatibility to enable high utilization of Tensor Cores. Additionally, BitDecoding incorporates a warp-efficient parallel decoding kernel and a fine-grained asynchronous pipeline, minimizing dequantization overhead and improving computational efficiency. Experiments show that BitDecoding achieves up to 7.5x speedup on RTX 4090, 4.8x on A100, and 8.9x on H100, compared to FP16 FlashDecoding-v2. It also outperforms the state-of-the-art low-bit KV cache implementation (QServe) by up to 4.3x. On LLaMA-3.1-8B with a 128K sequence length, BitDecoding reduces single-batch decoding latency by 3x, demonstrating its effectiveness in long-context generation scenarios. The code is available at https://github.com/DD-DuDa/BitDecoding.
☆ AlphaSpace: Enabling Robotic Actions through Semantic Tokenization and Symbolic Reasoning
This paper presents AlphaSpace, a novel methodology designed to enhance the spatial reasoning capabilities of large language models (LLMs) for 3D Cartesian space navigation. AlphaSpace employs a semantics-based tokenization strategy, encoding height information through specialized semantic tokens, and integrates primarily symbolic synthetic reasoning data. This approach enables LLMs to accurately manipulate objects by positioning them at specific [x, y, z] coordinates. Experimental results demonstrate that AlphaSpace significantly outperforms existing models on manipulation subtasks, achieving a total accuracy of 66.67%, compared to 37.5% for GPT-4o and 29.17% for Claude 3.5 Sonnet.
☆ Synthetic Function Demonstrations Improve Generation in Low-Resource Programming Languages
A key consideration when training an LLM is whether the target language is more or less resourced, whether this is English compared to Welsh, or Python compared to Excel. Typical training data for programming languages consist of real program demonstrations coupled with human-written comments. Here we present novel approaches to the creation of such data for low resource programming languages. We generate fully-synthetic, textbook-quality demonstrations of common library functions in an example domain of Excel formulas, using a teacher model. We then finetune an underperforming student model, and show improvement on 2 question-answering datasets recast into the Excel domain. We show advantages of finetuning over standard, off-the-shelf RAG approaches, which can offer only modest improvement due to the unfamiliar target domain.
☆ Construction Identification and Disambiguation Using BERT: A Case Study of NPN ACL
Construction Grammar hypothesizes that knowledge of a language consists chiefly of knowledge of form-meaning pairs (''constructions'') that include vocabulary, general grammar rules, and even idiosyncratic patterns. Recent work has shown that transformer language models represent at least some constructional patterns, including ones where the construction is rare overall. In this work, we probe BERT's representation of the form and meaning of a minor construction of English, the NPN (noun-preposition-noun) construction -- exhibited in such expressions as face to face and day to day -- which is known to be polysemous. We construct a benchmark dataset of semantically annotated corpus instances (including distractors that superficially resemble the construction). With this dataset, we train and evaluate probing classifiers. They achieve decent discrimination of the construction from distractors, as well as sense disambiguation among true instances of the construction, revealing that BERT embeddings carry indications of the construction's semantics. Moreover, artificially permuting the word order of true construction instances causes them to be rejected, indicating sensitivity to matters of form. We conclude that BERT does latently encode at least some knowledge of the NPN construction going beyond a surface syntactic pattern and lexical cues.
comment: 8 pages, ACL long-paper format (preprint)
☆ Predicting the Road Ahead: A Knowledge Graph based Foundation Model for Scene Understanding in Autonomous Driving
The autonomous driving field has seen remarkable advancements in various topics, such as object recognition, trajectory prediction, and motion planning. However, current approaches face limitations in effectively comprehending the complex evolutions of driving scenes over time. This paper proposes FM4SU, a novel methodology for training a symbolic foundation model (FM) for scene understanding in autonomous driving. It leverages knowledge graphs (KGs) to capture sensory observation along with domain knowledge such as road topology, traffic rules, or complex interactions between traffic participants. A bird's eye view (BEV) symbolic representation is extracted from the KG for each driving scene, including the spatio-temporal information among the objects across the scenes. The BEV representation is serialized into a sequence of tokens and given to pre-trained language models (PLMs) for learning an inherent understanding of the co-occurrence among driving scene elements and generating predictions on the next scenes. We conducted a number of experiments using the nuScenes dataset and KG in various scenarios. The results demonstrate that fine-tuned models achieve significantly higher accuracy in all tasks. The fine-tuned T5 model achieved a next scene prediction accuracy of 86.7%. This paper concludes that FM4SU offers a promising foundation for developing more comprehensive models for scene understanding in autonomous driving.
☆ Unsupervised Acquisition of Discrete Grammatical Categories
This article presents experiments performed using a computational laboratory environment for language acquisition experiments. It implements a multi-agent system consisting of two agents: an adult language model and a daughter language model that aims to learn the mother language. Crucially, the daughter agent does not have access to the internal knowledge of the mother language model but only to the language exemplars the mother agent generates. These experiments illustrate how this system can be used to acquire abstract grammatical knowledge. We demonstrate how statistical analyses of patterns in the input data corresponding to grammatical categories yield discrete grammatical rules. These rules are subsequently added to the grammatical knowledge of the daughter language model. To this end, hierarchical agglomerative cluster analysis was applied to the utterances consecutively generated by the mother language model. It is argued that this procedure can be used to acquire structures resembling grammatical categories proposed by linguists for natural languages. Thus, it is established that non-trivial grammatical knowledge has been acquired. Moreover, the parameter configuration of this computational laboratory environment determined using training data generated by the mother language model is validated in a second experiment with a test set similarly resulting in the acquisition of non-trivial categories.
comment: 34 pages, 3 figures, 7 tables
☆ Commander-GPT: Fully Unleashing the Sarcasm Detection Capability of Multi-Modal Large Language Models
Sarcasm detection, as a crucial research direction in the field of Natural Language Processing (NLP), has attracted widespread attention. Traditional sarcasm detection tasks have typically focused on single-modal approaches (e.g., text), but due to the implicit and subtle nature of sarcasm, such methods often fail to yield satisfactory results. In recent years, researchers have shifted the focus of sarcasm detection to multi-modal approaches. However, effectively leveraging multi-modal information to accurately identify sarcastic content remains a challenge that warrants further exploration. Leveraging the powerful integrated processing capabilities of Multi-Modal Large Language Models (MLLMs) for various information sources, we propose an innovative multi-modal Commander-GPT framework. Inspired by military strategy, we first decompose the sarcasm detection task into six distinct sub-tasks. A central commander (decision-maker) then assigns the best-suited large language model to address each specific sub-task. Ultimately, the detection results from each model are aggregated to identify sarcasm. We conducted extensive experiments on MMSD and MMSD 2.0, utilizing four multi-modal large language models and six prompting strategies. Our experiments demonstrate that our approach achieves state-of-the-art performance, with a 19.3% improvement in F1 score, without necessitating fine-tuning or ground-truth rationales.
☆ ArchSeek: Retrieving Architectural Case Studies Using Vision-Language Models
Efficiently searching for relevant case studies is critical in architectural design, as designers rely on precedent examples to guide or inspire their ongoing projects. However, traditional text-based search tools struggle to capture the inherently visual and complex nature of architectural knowledge, often leading to time-consuming and imprecise exploration. This paper introduces ArchSeek, an innovative case study search system with recommendation capability, tailored for architecture design professionals. Powered by the visual understanding capabilities from vision-language models and cross-modal embeddings, it enables text and image queries with fine-grained control, and interaction-based design case recommendations. It offers architects a more efficient, personalized way to discover design inspirations, with potential applications across other visually driven design fields. The source code is available at https://github.com/danruili/ArchSeek.
comment: 15 pages, 8 figures, 3 tables. Accepted by CAAD Futures 2025
☆ AgentSpec: Customizable Runtime Enforcement for Safe and Reliable LLM Agents
Agents built on LLMs are increasingly deployed across diverse domains, automating complex decision-making and task execution. However, their autonomy introduces safety risks, including security vulnerabilities, legal violations, and unintended harmful actions. Existing mitigation methods, such as model-based safeguards and early enforcement strategies, fall short in robustness, interpretability, and adaptability. To address these challenges, we propose AgentSpec, a lightweight domain-specific language for specifying and enforcing runtime constraints on LLM agents. With AgentSpec, users define structured rules that incorporate triggers, predicates, and enforcement mechanisms, ensuring agents operate within predefined safety boundaries. We implement AgentSpec across multiple domains, including code execution, embodied agents, and autonomous driving, demonstrating its adaptability and effectiveness. Our evaluation shows that AgentSpec successfully prevents unsafe executions in over 90% of code agent cases, eliminates all hazardous actions in embodied agent tasks, and enforces 100% compliance by autonomous vehicles (AVs). Despite its strong safety guarantees, AgentSpec remains computationally lightweight, with overheads in milliseconds. By combining interpretability, modularity, and efficiency, AgentSpec provides a practical and scalable solution for enforcing LLM agent safety across diverse applications. We also automate the generation of rules using LLMs and assess their effectiveness. Our evaluation shows that the rules generated by OpenAI o1 achieve a precision of 95.56% and recall of 70.96% for embodied agents, successfully identifying 87.26% of the risky code, and prevent AVs from breaking laws in 5 out of 8 scenarios.
☆ ZeroLM: Data-Free Transformer Architecture Search for Language Models
Neural architecture search (NAS) provides a systematic framework for automating the design of neural network architectures, yet its widespread adoption is hindered by prohibitive computational requirements. Existing zero-cost proxy methods, while reducing search overhead, demonstrate inadequate performance in architecture ranking tasks, particularly for Transformer-based models where they often underperform simple parameter counting metrics. Current automated proxy discovery approaches suffer from extended search times, susceptibility to data overfitting, and structural complexity. This paper introduces a novel zero-cost proxy methodology that quantifies model capacity through efficient weight statistics computation while decomposing Transformer architectures into functionally distinct sub-modules, thereby optimizing the balance of their contributions to overall performance. Our comprehensive evaluation demonstrates the superiority of this approach, achieving a Spearman's rho of 0.76 and Kendall's tau of 0.53 on the FlexiBERT benchmark. The proposed method exhibits exceptional computational efficiency while maintaining robust performance across diverse NAS benchmark tasks, offering a practical solution for large-scale architecture search.
☆ LANGALIGN: Enhancing Non-English Language Models via Cross-Lingual Embedding Alignment
While Large Language Models have gained attention, many service developers still rely on embedding-based models due to practical constraints. In such cases, the quality of fine-tuning data directly impacts performance, and English datasets are often used as seed data for training non-English models. In this study, we propose LANGALIGN, which enhances target language processing by aligning English embedding vectors with those of the target language at the interface between the language model and the task header. Experiments on Korean, Japanese, and Chinese demonstrate that LANGALIGN significantly improves performance across all three languages. Additionally, we show that LANGALIGN can be applied in reverse to convert target language data into a format that an English-based model can process.
comment: now preparing
☆ LinkAlign: Scalable Schema Linking for Real-World Large-Scale Multi-Database Text-to-SQL
Schema linking is a critical bottleneck in achieving human-level performance in Text-to-SQL tasks, particularly in real-world large-scale multi-database scenarios. Addressing schema linking faces two major challenges: (1) Database Retrieval: selecting the correct database from a large schema pool in multi-database settings, while filtering out irrelevant ones. (2) Schema Item Grounding: accurately identifying the relevant tables and columns from within a large and redundant schema for SQL generation. To address this, we introduce LinkAlign, a novel framework that can effectively adapt existing baselines to real-world environments by systematically addressing schema linking. Our framework comprises three key steps: multi-round semantic enhanced retrieval and irrelevant information isolation for Challenge 1, and schema extraction enhancement for Challenge 2. We evaluate our method performance of schema linking on the SPIDER and BIRD benchmarks, and the ability to adapt existing Text-to-SQL models to real-world environments on the SPIDER 2.0-lite benchmark. Experiments show that LinkAlign outperforms existing baselines in multi-database settings, demonstrating its effectiveness and robustness. On the other hand, our method ranks highest among models excluding those using long chain-of-thought reasoning LLMs. This work bridges the gap between current research and real-world scenarios, providing a practical solution for robust and scalable schema linking. The codes are available at https://github.com/Satissss/LinkAlign.
☆ ClinText-SP and RigoBERTa Clinical: a new set of open resources for Spanish Clinical NLP
We present a novel contribution to Spanish clinical natural language processing by introducing the largest publicly available clinical corpus, ClinText-SP, along with a state-of-the-art clinical encoder language model, RigoBERTa Clinical. Our corpus was meticulously curated from diverse open sources, including clinical cases from medical journals and annotated corpora from shared tasks, providing a rich and diverse dataset that was previously difficult to access. RigoBERTa Clinical, developed through domain-adaptive pretraining on this comprehensive dataset, significantly outperforms existing models on multiple clinical NLP benchmarks. By publicly releasing both the dataset and the model, we aim to empower the research community with robust resources that can drive further advancements in clinical NLP and ultimately contribute to improved healthcare applications.
☆ Dense Retrieval for Low Resource Languages -- the Case of Amharic Language
This paper reports some difficulties and some results when using dense retrievers on Amharic, one of the low-resource languages spoken by 120 millions populations. The efforts put and difficulties faced by University Addis Ababa toward Amharic Information Retrieval will be developed during the presentation.
comment: 4 pages, 2 figures
☆ Distil-xLSTM: Learning Attention Mechanisms through Recurrent Structures
The current era of Natural Language Processing (NLP) is dominated by Transformer models. However, novel architectures relying on recurrent mechanisms, such as xLSTM and Mamba, have been proposed as alternatives to attention-based models. Although computation is done differently than with the attention mechanism mechanism, these recurrent models yield good results and sometimes even outperform state-of-the-art attention-based models. In this work, we propose Distil-xLSTM, an xLSTM-based Small Language Model (SLM) trained by distilling knowledge from a Large Language Model (LLM) that shows promising results while being compute and scale efficient. Our Distil-xLSTM focuses on approximating a transformer-based model attention parametrization using its recurrent sequence mixing components and shows good results with minimal training.
☆ Self-Reported Confidence of Large Language Models in Gastroenterology: Analysis of Commercial, Open-Source, and Quantized Models
This study evaluated self-reported response certainty across several large language models (GPT, Claude, Llama, Phi, Mistral, Gemini, Gemma, and Qwen) using 300 gastroenterology board-style questions. The highest-performing models (GPT-o1 preview, GPT-4o, and Claude-3.5-Sonnet) achieved Brier scores of 0.15-0.2 and AUROC of 0.6. Although newer models demonstrated improved performance, all exhibited a consistent tendency towards overconfidence. Uncertainty estimation presents a significant challenge to the safe use of LLMs in healthcare. Keywords: Large Language Models; Confidence Elicitation; Artificial Intelligence; Gastroenterology; Uncertainty Quantification
comment: 35 pages, 5 figures, 1 table, 7 supplementary figures
☆ Instruction-Aligned Visual Attention for Mitigating Hallucinations in Large Vision-Language Models ICME2025
Despite the significant success of Large Vision-Language models(LVLMs), these models still suffer hallucinations when describing images, generating answers that include non-existent objects. It is reported that these models tend to over-focus on certain irrelevant image tokens that do not contain critical information for answering the question and distort the output. To address this, we propose an Instruction-Aligned Visual Attention(IAVA) approach, which identifies irrelevant tokens by comparing changes in attention weights under two different instructions. By applying contrastive decoding, we dynamically adjust the logits generated from original image tokens and irrelevant image tokens, reducing the model's over-attention to irrelevant information. The experimental results demonstrate that IAVA consistently outperforms existing decoding techniques on benchmarks such as MME, POPE, and TextVQA in mitigating object hallucinations. Our IAVA approach is available online at https://github.com/Lee-lab558/IAVA.
comment: Accepted by ICME2025
☆ Natural Language Processing for Electronic Health Records in Scandinavian Languages: Norwegian, Swedish, and Danish
Background: Clinical natural language processing (NLP) refers to the use of computational methods for extracting, processing, and analyzing unstructured clinical text data, and holds a huge potential to transform healthcare in various clinical tasks. Objective: The study aims to perform a systematic review to comprehensively assess and analyze the state-of-the-art NLP methods for the mainland Scandinavian clinical text. Method: A literature search was conducted in various online databases including PubMed, ScienceDirect, Google Scholar, ACM digital library, and IEEE Xplore between December 2022 and February 2024. Further, relevant references to the included articles were also used to solidify our search. The final pool includes articles that conducted clinical NLP in the mainland Scandinavian languages and were published in English between 2010 and 2024. Results: Out of the 113 articles, 18% (n=21) focus on Norwegian clinical text, 64% (n=72) on Swedish, 10% (n=11) on Danish, and 8% (n=9) focus on more than one language. Generally, the review identified positive developments across the region despite some observable gaps and disparities between the languages. There are substantial disparities in the level of adoption of transformer-based models. In essential tasks such as de-identification, there is significantly less research activity focusing on Norwegian and Danish compared to Swedish text. Further, the review identified a low level of sharing resources such as data, experimentation code, pre-trained models, and rate of adaptation and transfer learning in the region. Conclusion: The review presented a comprehensive assessment of the state-of-the-art Clinical NLP for electronic health records (EHR) text in mainland Scandinavian languages and, highlighted the potential barriers and challenges that hinder the rapid advancement of the field in the region.
comment: 45 pages including the appendix, 9 figures in the main manuscript and 11 figures in the Appendix
☆ SciClaims: An End-to-End Generative System for Biomedical Claim Analysis
Validating key claims in scientific literature, particularly in biomedical research, is essential for ensuring accuracy and advancing knowledge. This process is critical in sectors like the pharmaceutical industry, where rapid scientific progress requires automation and deep domain expertise. However, current solutions have significant limitations. They lack end-to-end pipelines encompassing all claim extraction, evidence retrieval, and verification steps; rely on complex NLP and information retrieval pipelines prone to multiple failure points; and often fail to provide clear, user-friendly justifications for claim verification outcomes. To address these challenges, we introduce SciClaims, an advanced system powered by state-of-the-art large language models (LLMs) that seamlessly integrates the entire scientific claim analysis process. SciClaims outperforms previous approaches in both claim extraction and verification without requiring additional fine-tuning, setting a new benchmark for automated scientific claim analysis.
comment: Pre-print version
☆ Autoregressive Language Models for Knowledge Base Population: A case study in the space mission domain
Knowledge base population KBP plays a crucial role in populating and maintaining knowledge bases up-to-date in organizations by leveraging domain corpora. Motivated by the increasingly large context windows supported by large language models, we propose to fine-tune an autoregressive language model for end-toend KPB. Our case study involves the population of a space mission knowledge graph. To fine-tune the model we generate a dataset for end-to-end KBP tapping into existing domain resources. Our case study shows that fine-tuned language models of limited size can achieve competitive and even higher accuracy than larger models in the KBP task. Smaller models specialized for KBP offer affordable deployment and lower-cost inference. Moreover, KBP specialist models do not require the ontology to be included in the prompt, allowing for more space in the context for additional input text or output serialization.
comment: Pre-print version
☆ Verbal Process Supervision Elicits Better Coding Agents
The emergence of large language models and their applications as AI agents have significantly advanced state-of-the-art code generation benchmarks, transforming modern software engineering tasks. However, even with test-time computed reasoning models, these systems still struggle with complex software engineering challenges. This work introduces CURA, a code understanding and reasoning agent system enhanced with verbal process supervision (VPS), achieving a 3.65\% improvement over baseline models on challenging benchmarks like BigCodeBench. Furthermore, CURA, when paired with the o3-mini model and VPS techniques, attains state-of-the-art performance. This work represents a step forward in integrating reasoning-driven architectures with LLM-based code generation, enabling agentic reasoning for language models to solve complex software engineering tasks.
☆ Safeguarding Mobile GUI Agent via Logic-based Action Verification
Large Foundation Models (LFMs) have unlocked new possibilities in human-computer interaction, particularly with the rise of mobile Graphical User Interface (GUI) Agents capable of interpreting GUIs. These agents promise to revolutionize mobile computing by allowing users to automate complex mobile tasks through simple natural language instructions. However, the inherent probabilistic nature of LFMs, coupled with the ambiguity and context-dependence of mobile tasks, makes LFM-based automation unreliable and prone to errors. To address this critical challenge, we introduce VeriSafe Agent (VSA): a formal verification system that serves as a logically grounded safeguard for Mobile GUI Agents. VSA is designed to deterministically ensure that an agent's actions strictly align with user intent before conducting an action. At its core, VSA introduces a novel autoformalization technique that translates natural language user instructions into a formally verifiable specification, expressed in our domain-specific language (DSL). This enables runtime, rule-based verification, allowing VSA to detect and prevent erroneous actions executing an action, either by providing corrective feedback or halting unsafe behavior. To the best of our knowledge, VSA is the first attempt to bring the rigor of formal verification to GUI agent. effectively bridging the gap between LFM-driven automation and formal software verification. We implement VSA using off-the-shelf LLM services (GPT-4o) and evaluate its performance on 300 user instructions across 18 widely used mobile apps. The results demonstrate that VSA achieves 94.3%-98.33% accuracy in verifying agent actions, representing a significant 20.4%-25.6% improvement over existing LLM-based verification methods, and consequently increases the GUI agent's task completion rate by 90%-130%.
☆ MAGIC-VQA: Multimodal And Grounded Inference with Commonsense Knowledge for Visual Question Answering
Visual Question Answering (VQA) requires reasoning across visual and textual modalities, yet Large Vision-Language Models (LVLMs) often lack integrated commonsense knowledge, limiting their robustness in real-world scenarios. To address this, we introduce MAGIC-VQA, a novel framework that enhances VQA by systematically integrating commonsense knowledge with LVLMs. MAGIC-VQA employs a three-stage process: (1) Explicit Knowledge Integration from external sources, (2) By-Type Post-Processing for contextual refinement, and (3) Implicit Knowledge Augmentation using a Graph Neural Network (GNN) for structured reasoning. While GNNs bring greater depth to structured inference, they enable superior relational inference beyond LVLMs. MAGIC-VQA bridges a key gap by unifying commonsensse knowledge with LVLM-driven reasoning, eliminating the need for extensive pre-training or complex prompt tuning. Our framework achieves state-of-the-art performance on benchmark datasets, significantly improving commonsense reasoning in VQA.
comment: 8 Pages, 5 figures
☆ Whispering in Amharic: Fine-tuning Whisper for Low-resource Language
This work explores fine-tuning OpenAI's Whisper automatic speech recognition (ASR) model for Amharic, a low-resource language, to improve transcription accuracy. While the foundational Whisper model struggles with Amharic due to limited representation in its training data, we fine-tune it using datasets like Mozilla Common Voice, FLEURS, and the BDU-speech dataset. The best-performing model, Whispersmall-am, significantly improves when finetuned on a mix of existing FLEURS data and new, unseen Amharic datasets. Training solely on new data leads to poor performance, but combining it with FLEURS data reinforces the model, enabling better specialization in Amharic. We also demonstrate that normalizing Amharic homophones significantly enhances Word Error Rate (WER) and Bilingual Evaluation Understudy (BLEU) scores. This study underscores the importance of fine-tuning strategies and dataset composition for improving ASR in low-resource languages, providing insights for future Amharic speech recognition research.
☆ PM4Bench: A Parallel Multilingual Multi-Modal Multi-task Benchmark for Large Vision Language Model
Existing multilingual benchmarks for Large Vision Language Models (LVLMs) suffer from limitations including language-specific content biases, disjointed multimodal input formats, and a lack of safety evaluation. To address these gaps, we propose PM4Bench, the first Parallel Multilingual Multi-Modal Multi-task Benchmark for LVLMs. PM4Bench features a parallel corpus design across 10 languages, enabling fair and accurate cross-lingual comparisons. It includes the vision setting where text and queries are embedded in images, requiring LVLMs to simultaneously "see", "read", and "think", aligning with real-world applications. Additionally, PM\textsuperscript{4}Bench incorporates safety evaluations, addressing critical oversight in existing multilingual benchmarks. Using PM4Bench, we evaluate 11 mainstream LVLMs, revealing significant cross-linguistic performance disparities, particularly in vision settings, and identifying OCR capability as a key determinant of these imbalances. We will release PM4Bench at https://github.com/opendatalab/PM4Bench .
comment: Equal contribution: Junyuan Gao, Jiahe Song, Jiang Wu; Corresponding author: Conghui He
☆ Global-Local Tree Search for Language Guided 3D Scene Generation CVPR 2025
Large Vision-Language Models (VLMs), such as GPT-4, have achieved remarkable success across various fields. However, there are few studies on 3D indoor scene generation with VLMs. This paper considers this task as a planning problem subject to spatial and layout common sense constraints. To solve the problem with a VLM, we propose a new global-local tree search algorithm. Globally, the method places each object sequentially and explores multiple placements during each placement process, where the problem space is represented as a tree. To reduce the depth of the tree, we decompose the scene structure hierarchically, i.e. room level, region level, floor object level, and supported object level. The algorithm independently generates the floor objects in different regions and supported objects placed on different floor objects. Locally, we also decompose the sub-task, the placement of each object, into multiple steps. The algorithm searches the tree of problem space. To leverage the VLM model to produce positions of objects, we discretize the top-down view space as a dense grid and fill each cell with diverse emojis to make to cells distinct. We prompt the VLM with the emoji grid and the VLM produces a reasonable location for the object by describing the position with the name of emojis. The quantitative and qualitative experimental results illustrate our approach generates more plausible 3D scenes than state-of-the-art approaches. Our source code is available at https://github.com/dw-dengwei/TreeSearchGen .
comment: Accepted by CVPR 2025
☆ Words as Bridges: Exploring Computational Support for Cross-Disciplinary Translation Work
Scholars often explore literature outside of their home community of study. This exploration process is frequently hampered by field-specific jargon. Past computational work often focuses on supporting translation work by removing jargon through simplification and summarization; here, we explore a different approach that preserves jargon as useful bridges to new conceptual spaces. Specifically, we cast different scholarly domains as different language-using communities, and explore how to adapt techniques from unsupervised cross-lingual alignment of word embeddings to explore conceptual alignments between domain-specific word embedding spaces.We developed a prototype cross-domain search engine that uses aligned domain-specific embeddings to support conceptual exploration, and tested this prototype in two case studies. We discuss qualitative insights into the promises and pitfalls of this approach to translation work, and suggest design insights for future interfaces that provide computational support for cross-domain information seeking.
comment: 26 pages, 8 tables, 6 figures
StableGS: A Floater-Free Framework for 3D Gaussian Splatting
Recent years have witnessed remarkable success of 3D Gaussian Splatting (3DGS) in novel view synthesis, surpassing prior differentiable rendering methods in both quality and efficiency. However, its training process suffers from coupled opacity-color optimization that frequently converges to local minima, producing floater artifacts that degrade visual fidelity. We present StableGS, a framework that eliminates floaters through cross-view depth consistency constraints while introducing a dual-opacity GS model to decouple geometry and material properties of translucent objects. To further enhance reconstruction quality in weakly-textured regions, we integrate DUSt3R depth estimation, significantly improving geometric stability. Our method fundamentally addresses 3DGS training instabilities, outperforming existing state-of-the-art methods across open-source datasets.
☆ On the Perception Bottleneck of VLMs for Chart Understanding
Chart understanding requires models to effectively analyze and reason about numerical data, textual elements, and complex visual components. Our observations reveal that the perception capabilities of existing large vision-language models (LVLMs) constitute a critical bottleneck in this process. In this study, we delve into this perception bottleneck by decomposing it into two components: the vision encoder bottleneck, where the visual representation may fail to encapsulate the correct information, and the extraction bottleneck, where the language model struggles to extract the necessary information from the provided visual representations. Through comprehensive experiments, we find that (1) the information embedded within visual representations is substantially richer than what is typically captured by linear extractors, such as the widely used retrieval accuracy metric; (2) While instruction tuning effectively enhances the extraction capability of LVLMs, the vision encoder remains a critical bottleneck, demanding focused attention and improvement. Therefore, we further enhance the visual encoder to mitigate the vision encoder bottleneck under a contrastive learning framework. Empirical results demonstrate that our approach significantly mitigates the perception bottleneck and improves the ability of LVLMs to comprehend charts. Code is publicly available at https://github.com/hkust-nlp/Vision4Chart.
☆ Teaching LLMs for Step-Level Automatic Math Correction via Reinforcement Learning
Automatic math correction aims to check students' solutions to mathematical problems via artificial intelligence technologies. Most existing studies focus on judging the final answer at the problem level, while they ignore detailed feedback on each step in a math problem-solving process, which requires abilities of semantic understanding and reasoning. In this paper, we propose a reinforcement learning (RL)-based method to boost large language model (LLM) for step-level automatic math correction, named StepAMC. Particularly, we convert the step-level automatic math correction within the text classification task into an RL problem to enhance the reasoning capabilities of LLMs. Then, we design a space-constrained policy network to improve the stability of RL. Then, we introduce a fine-grained reward network to convert the binary human feedback into a continuous value. We conduct extensive experiments over two benchmark datasets and the results show that our model outperforms the eleven strong baselines.
☆ Solving Situation Puzzles with Large Language Model and External Reformulation
In recent years, large language models (LLMs) have shown an impressive ability to perform arithmetic and symbolic reasoning tasks. However, we found that LLMs (e.g., ChatGPT) cannot perform well on reasoning that requires multiple rounds of dialogue, especially when solving situation puzzles. Specifically, LLMs intend to ask very detailed questions focusing on a specific aspect or same/similar questions after several rounds of Q&As. To help LLMs get out of the above dilemma, we propose a novel external reformulation methodology, where the situation puzzle will be reformulated after several rounds of Q&A or when the LLMs raise an incorrect guess. Experiments show superior performance (e.g., win rate, number of question/guess attempts) of our method than directly using LLMs for solving situation puzzles, highlighting the potential of strategic problem reformulation to enhance the reasoning capabilities of LLMs in complex interactive scenarios.
☆ J&H: Evaluating the Robustness of Large Language Models Under Knowledge-Injection Attacks in Legal Domain
As the scale and capabilities of Large Language Models (LLMs) increase, their applications in knowledge-intensive fields such as legal domain have garnered widespread attention. However, it remains doubtful whether these LLMs make judgments based on domain knowledge for reasoning. If LLMs base their judgments solely on specific words or patterns, rather than on the underlying logic of the language, the ''LLM-as-judges'' paradigm poses substantial risks in the real-world applications. To address this question, we propose a method of legal knowledge injection attacks for robustness testing, thereby inferring whether LLMs have learned legal knowledge and reasoning logic. In this paper, we propose J&H: an evaluation framework for detecting the robustness of LLMs under knowledge injection attacks in the legal domain. The aim of the framework is to explore whether LLMs perform deductive reasoning when accomplishing legal tasks. To further this aim, we have attacked each part of the reasoning logic underlying these tasks (major premise, minor premise, and conclusion generation). We have collected mistakes that legal experts might make in judicial decisions in the real world, such as typos, legal synonyms, inaccurate external legal statutes retrieval. However, in real legal practice, legal experts tend to overlook these mistakes and make judgments based on logic. However, when faced with these errors, LLMs are likely to be misled by typographical errors and may not utilize logic in their judgments. We conducted knowledge injection attacks on existing general and domain-specific LLMs. Current LLMs are not robust against the attacks employed in our experiments. In addition we propose and compare several methods to enhance the knowledge robustness of LLMs.
comment: 10 pages, 5 figures
☆ Bridging Writing Manner Gap in Visual Instruction Tuning by Creating LLM-aligned Instructions
In the realm of Large Multi-modal Models (LMMs), the instruction quality during the visual instruction tuning stage significantly influences the performance of modality alignment. In this paper, we assess the instruction quality from a unique perspective termed \textbf{Writing Manner}, which encompasses the selection of vocabulary, grammar and sentence structure to convey specific semantics. We argue that there exists a substantial writing manner gap between the visual instructions and the base Large Language Models (LLMs) within LMMs. This gap forces the pre-trained base LLMs to deviate from their original writing styles, leading to capability degradation of both base LLMs and LMMs. To bridge the writing manner gap while preserving the original semantics, we propose directly leveraging the base LLM to align the writing manner of soft-format visual instructions with that of the base LLM itself, resulting in novel LLM-aligned instructions. The manual writing manner evaluation results demonstrate that our approach successfully minimizes the writing manner gap. By utilizing LLM-aligned instructions, the baseline models LLaVA-7B and QwenVL demonstrate enhanced resistance to hallucinations and non-trivial comprehensive improvements across all $15$ visual and language benchmarks.
☆ Surgical Action Planning with Large Language Models
In robot-assisted minimally invasive surgery, we introduce the Surgical Action Planning (SAP) task, which generates future action plans from visual inputs to address the absence of intraoperative predictive planning in current intelligent applications. SAP shows great potential for enhancing intraoperative guidance and automating procedures. However, it faces challenges such as understanding instrument-action relationships and tracking surgical progress. Large Language Models (LLMs) show promise in understanding surgical video content but remain underexplored for predictive decision-making in SAP, as they focus mainly on retrospective analysis. Challenges like data privacy, computational demands, and modality-specific constraints further highlight significant research gaps. To tackle these challenges, we introduce LLM-SAP, a Large Language Models-based Surgical Action Planning framework that predicts future actions and generates text responses by interpreting natural language prompts of surgical goals. The text responses potentially support surgical education, intraoperative decision-making, procedure documentation, and skill analysis. LLM-SAP integrates two novel modules: the Near-History Focus Memory Module (NHF-MM) for modeling historical states and the prompts factory for action planning. We evaluate LLM-SAP on our constructed CholecT50-SAP dataset using models like Qwen2.5 and Qwen2-VL, demonstrating its effectiveness in next-action prediction. Pre-trained LLMs are tested zero-shot, and supervised fine-tuning (SFT) with LoRA is implemented to address data privacy concerns. Our experiments show that Qwen2.5-72B-SFT surpasses Qwen2.5-72B with a 19.3% higher accuracy.
comment: 10 pages,4 figures
☆ Fact-checking AI-generated news reports: Can LLMs catch their own lies?
In this paper, we evaluate the ability of Large Language Models (LLMs) to assess the veracity of claims in ''news reports'' generated by themselves or other LLMs. Our goal is to determine whether LLMs can effectively fact-check their own content, using methods similar to those used to verify claims made by humans. Our findings indicate that LLMs are more effective at assessing claims in national or international news stories than in local news stories, better at evaluating static information than dynamic information, and better at verifying true claims compared to false ones. We hypothesize that this disparity arises because the former types of claims are better represented in the training data. Additionally, we find that incorporating retrieved results from a search engine in a Retrieval-Augmented Generation (RAG) setting significantly reduces the number of claims an LLM cannot assess. However, this approach also increases the occurrence of incorrect assessments, partly due to irrelevant or low-quality search results. This diagnostic study highlights the need for future research on fact-checking machine-generated reports to prioritize improving the precision and relevance of retrieved information to better support fact-checking efforts. Furthermore, claims about dynamic events and local news may require human-in-the-loop fact-checking systems to ensure accuracy and reliability.
☆ When is dataset cartography ineffective? Using training dynamics does not improve robustness against Adversarial SQuAD
In this paper, I investigate the effectiveness of dataset cartography for extractive question answering on the SQuAD dataset. I begin by analyzing annotation artifacts in SQuAD and evaluate the impact of two adversarial datasets, AddSent and AddOneSent, on an ELECTRA-small model. Using training dynamics, I partition SQuAD into easy-to-learn, ambiguous, and hard-to-learn subsets. I then compare the performance of models trained on these subsets to those trained on randomly selected samples of equal size. Results show that training on cartography-based subsets does not improve generalization to the SQuAD validation set or the AddSent adversarial set. While the hard-to-learn subset yields a slightly higher F1 score on the AddOneSent dataset, the overall gains are limited. These findings suggest that dataset cartography provides little benefit for adversarial robustness in SQuAD-style QA tasks. I conclude by comparing these results to prior findings on SNLI and discuss possible reasons for the observed differences.
comment: 5 pages, 3 figures, 4 tables
☆ Sun-Shine: A Large Language Model for Tibetan Culture
Tibetan, a minority language in China, features a highly intricate grammatical structure, characterized by four verb tenses and a tense system with frequent irregularities, contributing to its extensive inflectional diversity. Recently, advances in Large Language Models (LLMs) have transformed the paradigm in many domains. Despite the success in other fields, current LLMs often fall short in catering to the needs of domain experts like Tibetans, and the potential of LLMs for Tibetan culture is under-explored. The intrinsic reasons are the immense and intricate nature of Tibetan culture as well as the necessity for higher granularity and richness in knowledge. Simultaneously, the complexity and uniqueness of its grammatical structure, coupled with its status as a minority ethnic language, contribute to data scarcity, which remains a fundamental challenge. To alleviate these issues, we introduce Llama-Sunshine (Sun-Shine), the first large language model for Tibetan culture, which is expert in various Tibetan language processing tasks. Sun-Shine incorporates state-of-the-art model architectures optimized for Tibetan's linguistic features. We also propose TIB-STC, a comprehensive dataset comprising diverse Tibetan texts such as literature, religious scripts, news, and conversational data, which is also the first large-scale dataset for Tibetan culture. Though comprehensive experiments, Sun-Shine not only demonstrates a higher level of knowledge expertise for Tibetan culture but also gains preliminary embodied intelligence capabilities in Tibetan language processing tasks, like language modeling, text classification, machine translation, and syntactic analysis. Moreover, it excels in low-resource scenarios, showcasing strong generalization capabilities.
☆ Bridging Emotions and Architecture: Sentiment Analysis in Modern Distributed Systems
Sentiment analysis is a field within NLP that has gained importance because it is applied in various areas such as; social media surveillance, customer feedback evaluation and market research. At the same time, distributed systems allow for effective processing of large amounts of data. Therefore, this paper examines how sentiment analysis converges with distributed systems by concentrating on different approaches, challenges and future investigations. Furthermore, we do an extensive experiment where we train sentiment analysis models using both single node configuration and distributed architecture to bring out the benefits and shortcomings of each method in terms of performance and accuracy.
comment: IEEE 3rd International Conference on Advancements in Smart, Secure and Intelligent Computing (ASSIC)
☆ Enhancing Multi-Label Emotion Analysis and Corresponding Intensities for Ethiopian Languages
In this digital world, people freely express their emotions using different social media platforms. As a result, modeling and integrating emotion-understanding models are vital for various human-computer interaction tasks such as decision-making, product and customer feedback analysis, political promotions, marketing research, and social media monitoring. As users express different emotions simultaneously in a single instance, annotating emotions in a multilabel setting such as the EthioEmo (Belay et al., 2025) dataset effectively captures this dynamic. Additionally, incorporating intensity, or the degree of emotion, is crucial, as emotions can significantly differ in their expressive strength and impact. This intensity is significant for assessing whether further action is necessary in decision-making processes, especially concerning negative emotions in applications such as healthcare and mental health studies. To enhance the EthioEmo dataset, we include annotations for the intensity of each labeled emotion. Furthermore, we evaluate various state-of-the-art encoder-only Pretrained Language Models (PLMs) and decoder-only Large Language Models (LLMs) to provide comprehensive benchmarking.
☆ PAD: Towards Efficient Data Generation for Transfer Learning Using Phrase Alignment
Transfer learning leverages the abundance of English data to address the scarcity of resources in modeling non-English languages, such as Korean. In this study, we explore the potential of Phrase Aligned Data (PAD) from standardized Statistical Machine Translation (SMT) to enhance the efficiency of transfer learning. Through extensive experiments, we demonstrate that PAD synergizes effectively with the syntactic characteristics of the Korean language, mitigating the weaknesses of SMT and significantly improving model performance. Moreover, we reveal that PAD complements traditional data construction methods and enhances their effectiveness when combined. This innovative approach not only boosts model performance but also suggests a cost-efficient solution for resource-scarce languages.
comment: Preparing for conference
☆ AfroXLMR-Social: Adapting Pre-trained Language Models for African Languages Social Media Text
Pretrained Language Models (PLMs) built from various sources are the foundation of today's NLP progress. Language representations learned by such models achieve strong performance across many tasks with datasets of varying sizes drawn from various sources. We explore a thorough analysis of domain and task adaptive continual pretraining approaches for low-resource African languages and a promising result is shown for the evaluated tasks. We create AfriSocial, a corpus designed for domain adaptive finetuning that passes through quality pre-processing steps. Continual pretraining PLMs using AfriSocial as domain adaptive pretraining (DAPT) data, consistently improves performance on fine-grained emotion classification task of 16 targeted languages from 1% to 28.27% macro F1 score. Likewise, using the task adaptive pertaining (TAPT) approach, further finetuning with small unlabeled but similar task data shows promising results. For example, unlabeled sentiment data (source) for fine-grained emotion classification task (target) improves the base model results by an F1 score ranging from 0.55% to 15.11%. Combining the two methods, DAPT + TAPT, achieves also better results than base models. All the resources will be available to improve low-resource NLP tasks, generally, as well as other similar domain tasks such as hate speech and sentiment tasks.
♻ ☆ Large Language Models Empowered Personalized Web Agents WWW 2025
Web agents have emerged as a promising direction to automate Web task completion based on user instructions, significantly enhancing user experience. Recently, Web agents have evolved from traditional agents to Large Language Models (LLMs)-based Web agents. Despite their success, existing LLM-based Web agents overlook the importance of personalized data (e.g., user profiles and historical Web behaviors) in assisting the understanding of users' personalized instructions and executing customized actions. To overcome the limitation, we first formulate the task of LLM-empowered personalized Web agents, which integrate personalized data and user instructions to personalize instruction comprehension and action execution. To address the absence of a comprehensive evaluation benchmark, we construct a Personalized Web Agent Benchmark (PersonalWAB), featuring user instructions, personalized user data, Web functions, and two evaluation paradigms across three personalized Web tasks. Moreover, we propose a Personalized User Memory-enhanced Alignment (PUMA) framework to adapt LLMs to the personalized Web agent task. PUMA utilizes a memory bank with a task-specific retrieval strategy to filter relevant historical Web behaviors. Based on the behaviors, PUMA then aligns LLMs for personalized action execution through fine-tuning and direct preference optimization. Extensive experiments validate the superiority of PUMA over existing Web agents on PersonalWAB.
comment: Accepted to WWW 2025. The code and data are available on the project website https://hongrucai.github.io/PersonalWAB/
♻ ☆ GroundCap: A Visually Grounded Image Captioning Dataset
Current image captioning systems lack the ability to link descriptive text to specific visual elements, making their outputs difficult to verify. While recent approaches offer some grounding capabilities, they cannot track object identities across multiple references or ground both actions and objects simultaneously. We propose a novel ID-based grounding system that enables consistent object reference tracking and action-object linking, and present GroundCap, a dataset containing 52,016 images from 77 movies, with 344 human-annotated and 52,016 automatically generated captions. Each caption is grounded on detected objects (132 classes) and actions (51 classes) using a tag system that maintains object identity while linking actions to the corresponding objects. Our approach features persistent object IDs for reference tracking, explicit action-object linking, and segmentation of background elements through K-means clustering. We propose gMETEOR, a metric combining caption quality with grounding accuracy, and establish baseline performance by fine-tuning Pixtral-12B. Human evaluation demonstrates our approach's effectiveness in producing verifiable descriptions with coherent object references.
comment: 37 pages
♻ ☆ Interleaved Scene Graphs for Interleaved Text-and-Image Generation Assessment ICLR 2025
Many real-world user queries (e.g. "How do to make egg fried rice?") could benefit from systems capable of generating responses with both textual steps with accompanying images, similar to a cookbook. Models designed to generate interleaved text and images face challenges in ensuring consistency within and across these modalities. To address these challenges, we present ISG, a comprehensive evaluation framework for interleaved text-and-image generation. ISG leverages a scene graph structure to capture relationships between text and image blocks, evaluating responses on four levels of granularity: holistic, structural, block-level, and image-specific. This multi-tiered evaluation allows for a nuanced assessment of consistency, coherence, and accuracy, and provides interpretable question-answer feedback. In conjunction with ISG, we introduce a benchmark, ISG-Bench, encompassing 1,150 samples across 8 categories and 21 subcategories. This benchmark dataset includes complex language-vision dependencies and golden answers to evaluate models effectively on vision-centric tasks such as style transfer, a challenging area for current models. Using ISG-Bench, we demonstrate that recent unified vision-language models perform poorly on generating interleaved content. While compositional approaches that combine separate language and image models show a 111% improvement over unified models at the holistic level, their performance remains suboptimal at both block and image levels. To facilitate future work, we develop ISG-Agent, a baseline agent employing a "plan-execute-refine" pipeline to invoke tools, achieving a 122% performance improvement.
comment: Accepted by ICLR 2025 as Spotlight. Project homepage: https://interleave-eval.github.io/
♻ ☆ HybridNorm: Towards Stable and Efficient Transformer Training via Hybrid Normalization
Transformers have become the de facto architecture for a wide range of machine learning tasks, particularly in large language models (LLMs). Despite their remarkable performance, challenges remain in training deep transformer networks, especially regarding the location of layer normalization. While Pre-Norm structures facilitate easier training due to their more prominent identity path, they often yield suboptimal performance compared to Post-Norm. In this paper, we propose $\textbf{HybridNorm}$, a straightforward yet effective hybrid normalization strategy that integrates the advantages of both Pre-Norm and Post-Norm approaches. Specifically, HybridNorm employs QKV normalization within the attention mechanism and Post-Norm in the feed-forward network (FFN) of each transformer block. This design not only stabilizes training but also enhances performance, particularly in the context of LLMs. Comprehensive experiments in both dense and sparse architectures show that HybridNorm consistently outperforms both Pre-Norm and Post-Norm approaches, achieving state-of-the-art results across various benchmarks. These findings highlight the potential of HybridNorm as a more stable and effective technique for improving the training and performance of deep transformer models. Code is available at https://github.com/BryceZhuo/HybridNorm.
♻ ☆ Federated Incremental Named Entity Recognition
Federated Named Entity Recognition (FNER) boosts model training within each local client by aggregating the model updates of decentralized local clients, without sharing their private data. However, existing FNER methods assume fixed entity types and local clients in advance, leading to their ineffectiveness in practical applications. In a more realistic scenario, local clients receive new entity types continuously, while new local clients collecting novel data may irregularly join the global FNER training. This challenging setup, referred to here as Federated Incremental NER, renders the global model suffering from heterogeneous forgetting of old entity types from both intra-client and inter-client perspectives. To overcome these challenges, we propose a Local-Global Forgetting Defense (LGFD) model. Specifically, to address intra-client forgetting, we develop a structural knowledge distillation loss to retain the latent space's feature structure and a pseudo-label-guided inter-type contrastive loss to enhance discriminative capability over different entity types, effectively preserving previously learned knowledge within local clients. To tackle inter-client forgetting, we propose a task switching monitor that can automatically identify new entity types under privacy protection and store the latest old global model for knowledge distillation and pseudo-labeling. Experiments demonstrate significant improvement of our LGFD model over comparison methods.
comment: Accepted by IEEE/ACM Transactions on Audio, Speech and Language Processing
♻ ☆ Assessing the Reliability and Validity of GPT-4 in Annotating Emotion Appraisal Ratings
Appraisal theories suggest that emotions arise from subjective evaluations of events, referred to as appraisals. The taxonomy of appraisals is quite diverse, and they are usually given ratings on a Likert scale to be annotated in an experiencer-annotator or reader-annotator paradigm. This paper studies GPT-4 as a reader-annotator of 21 specific appraisal ratings in different prompt settings, aiming to evaluate and improve its performance compared to human annotators. We found that GPT-4 is an effective reader-annotator that performs close to or even slightly better than human annotators, and its results can be significantly improved by using a majority voting of five completions. GPT-4 also effectively predicts appraisal ratings and emotion labels using a single prompt, but adding instruction complexity results in poorer performance. We also found that longer event descriptions lead to more accurate annotations for both model and human annotator ratings. This work contributes to the growing usage of LLMs in psychology and the strategies for improving GPT-4 performance in annotating appraisals.
comment: CLPsych 2025
♻ ☆ Image Over Text: Transforming Formula Recognition Evaluation with Character Detection Matching CVPR2025
Formula recognition presents significant challenges due to the complicated structure and varied notation of mathematical expressions. Despite continuous advancements in formula recognition models, the evaluation metrics employed by these models, such as BLEU and Edit Distance, still exhibit notable limitations. They overlook the fact that the same formula has diverse representations and is highly sensitive to the distribution of training data, thereby causing unfairness in formula recognition evaluation. To this end, we propose a Character Detection Matching (CDM) metric, ensuring the evaluation objectivity by designing an image-level rather than a LaTeX-level metric score. Specifically, CDM renders both the model-predicted LaTeX and the ground-truth LaTeX formulas into image-formatted formulas, then employs visual feature extraction and localization techniques for precise character-level matching, incorporating spatial position information. Such a spatially-aware and character-matching method offers a more accurate and equitable evaluation compared with previous BLEU and Edit Distance metrics that rely solely on text-based character matching. Experimentally, we evaluated various formula recognition models using CDM, BLEU, and ExpRate metrics. Their results demonstrate that the CDM aligns more closely with human evaluation standards and provides a fairer comparison across different models by eliminating discrepancies caused by diverse formula representations. Code is available at https://github.com/opendatalab/UniMERNet/tree/main/cdm.
comment: Accepted by CVPR2025
♻ ☆ How Much Knowledge Can You Pack into a LoRA Adapter without Harming LLM?
The performance of Large Language Models (LLMs) on many tasks is greatly limited by the knowledge learned during pre-training and stored in the model's parameters. Low-rank adaptation (LoRA) is a popular and efficient training technique for updating or domain-specific adaptation of LLMs. In this study, we investigate how new facts can be incorporated into the LLM using LoRA without compromising the previously learned knowledge. We fine-tuned Llama-3.1-8B-instruct using LoRA with varying amounts of new knowledge. Our experiments have shown that the best results are obtained when the training data contains a mixture of known and new facts. However, this approach is still potentially harmful because the model's performance on external question-answering benchmarks declines after such fine-tuning. When the training data is biased towards certain entities, the model tends to regress to few overrepresented answers. In addition, we found that the model becomes more confident and refuses to provide an answer in only few cases. These findings highlight the potential pitfalls of LoRA-based LLM updates and underscore the importance of training data composition and tuning parameters to balance new knowledge integration and general model capabilities.
♻ ☆ AutoTRIZ: Automating Engineering Innovation with TRIZ and Large Language Models
Various ideation methods, such as morphological analysis and design-by-analogy, have been developed to aid creative problem-solving and innovation. Among them, the Theory of Inventive Problem Solving (TRIZ) stands out as one of the best-known methods. However, the complexity of TRIZ and its reliance on users' knowledge, experience, and reasoning capabilities limit its practicality. To address this, we introduce AutoTRIZ, an artificial ideation system that integrates Large Language Models (LLMs) to automate and enhance the TRIZ methodology. By leveraging LLMs' vast pre-trained knowledge and advanced reasoning capabilities, AutoTRIZ offers a novel, generative, and interpretable approach to engineering innovation. AutoTRIZ takes a problem statement from the user as its initial input, automatically conduct the TRIZ reasoning process and generates a structured solution report. We demonstrate and evaluate the effectiveness of AutoTRIZ through comparative experiments with textbook cases and a real-world application in the design of a Battery Thermal Management System (BTMS). Moreover, the proposed LLM-based framework holds the potential for extension to automate other knowledge-based ideation methods, such as SCAMPER, Design Heuristics, and Design-by-Analogy, paving the way for a new era of AI-driven innovation tools.
comment: 28 pages, 12 figures
♻ ☆ GUI-World: A Video Benchmark and Dataset for Multimodal GUI-oriented Understanding ICLR 2025
Recently, Multimodal Large Language Models (MLLMs) have been used as agents to control keyboard and mouse inputs by directly perceiving the Graphical User Interface (GUI) and generating corresponding commands. However, current agents primarily demonstrate strong understanding capabilities in static environments and are mainly applied to relatively simple domains, such as Web or mobile interfaces. We argue that a robust GUI agent should be capable of perceiving temporal information on the GUI, including dynamic Web content and multi-step tasks. Additionally, it should possess a comprehensive understanding of various GUI scenarios, including desktop software and multi-window interactions. To this end, this paper introduces a new dataset, termed GUI-World, which features meticulously crafted Human-MLLM annotations, extensively covering six GUI scenarios and eight types of GUI-oriented questions in three formats. We evaluate the capabilities of current state-of-the-art MLLMs, including Image LLMs and Video LLMs, in understanding various types of GUI content, especially dynamic and sequential content. Our findings reveal that current models struggle with dynamic GUI content without manually annotated keyframes or operation history. On the other hand, Video LLMs fall short in all GUI-oriented tasks given the sparse GUI video dataset. Therefore, we take the initial step of leveraging a fine-tuned Video LLM, GUI-Vid, as a GUI-oriented assistant, demonstrating an improved understanding of various GUI tasks. However, due to the limitations in the performance of base LLMs, we conclude that using video LLMs as GUI agents remains a significant challenge. We believe our work provides valuable insights for future research in dynamic GUI content understanding. All the dataset and code are publicly available at: https://gui-world.github.io.
comment: Accepted by ICLR 2025
♻ ☆ ConCodeEval: Evaluating Large Language Models for Code Constraints in Domain-Specific Languages
Recent work shows Large Language Models (LLMs) struggle to understand natural language constraints for various text generation tasks in zero- and few-shot settings. While, in the code domain, there is wide usage of constraints in code format to maintain the integrity of code written in Domain-Specific Languages (DSLs) like JSON and YAML which are widely used for system-level programming tasks in enterprises. Given that LLMs are increasingly used for system-level code tasks, evaluating if they can comprehend these code constraints is crucial. However, no work has been done to evaluate their controllability over code constraints. Hence, we introduce ConCodeEval, a first-of-its-kind benchmark having two novel tasks for code constraints across five representations. Our findings suggest that language models struggle with code constraints. Code languages that perform excellently for normal code tasks do not perform well when the same languages represent fine-grained constraints.
♻ ☆ Toward a method for LLM-enabled Indoor Navigation
Indoor navigation presents unique challenges due to complex layouts, lack of GPS signals, and accessibility concerns. Existing solutions often struggle with real-time adaptability and user-specific needs. In this work, we explore the potential of a Large Language Model (LLM), i.e., ChatGPT, to generate natural, context-aware navigation instructions from indoor map images. We design and evaluate test cases across different real-world environments, analyzing the effectiveness of LLMs in interpreting spatial layouts, handling user constraints, and planning efficient routes. Our findings demonstrate the potential of LLMs for supporting personalized indoor navigation, with an average of 50.54% correct indications and a maximum of 77.78%. The results do not appear to depend on the complexity of the layout or the complexity of the expected path, but rather on the number of points of interest and the abundance of visual information, which negatively affect the performance.
comment: 7 pages, 3 figures, 5 tables
♻ ☆ LLM Post-Training: A Deep Dive into Reasoning Large Language Models
Large Language Models (LLMs) have transformed the natural language processing landscape and brought to life diverse applications. Pretraining on vast web-scale data has laid the foundation for these models, yet the research community is now increasingly shifting focus toward post-training techniques to achieve further breakthroughs. While pretraining provides a broad linguistic foundation, post-training methods enable LLMs to refine their knowledge, improve reasoning, enhance factual accuracy, and align more effectively with user intents and ethical considerations. Fine-tuning, reinforcement learning, and test-time scaling have emerged as critical strategies for optimizing LLMs performance, ensuring robustness, and improving adaptability across various real-world tasks. This survey provides a systematic exploration of post-training methodologies, analyzing their role in refining LLMs beyond pretraining, addressing key challenges such as catastrophic forgetting, reward hacking, and inference-time trade-offs. We highlight emerging directions in model alignment, scalable adaptation, and inference-time reasoning, and outline future research directions. We also provide a public repository to continually track developments in this fast-evolving field: https://github.com/mbzuai-oryx/Awesome-LLM-Post-training.
comment: 32 pages, 7 figures, 3 tables, 377 references. Github Repo: https://github.com/mbzuai-oryx/Awesome-LLM-Post-training
♻ ☆ ThreatCrawl: A BERT-based Focused Crawler for the Cybersecurity Domain
Publicly available information contains valuable information for Cyber Threat Intelligence (CTI). This can be used to prevent attacks that have already taken place on other systems. Ideally, only the initial attack succeeds and all subsequent ones are detected and stopped. But while there are different standards to exchange this information, a lot of it is shared in articles or blog posts in non-standardized ways. Manually scanning through multiple online portals and news pages to discover new threats and extracting them is a time-consuming task. To automize parts of this scanning process, multiple papers propose extractors that use Natural Language Processing (NLP) to extract Indicators of Compromise (IOCs) from documents. However, while this already solves the problem of extracting the information out of documents, the search for these documents is rarely considered. In this paper, a new focused crawler is proposed called ThreatCrawl, which uses Bidirectional Encoder Representations from Transformers (BERT)-based models to classify documents and adapt its crawling path dynamically. While ThreatCrawl has difficulties to classify the specific type of Open Source Intelligence (OSINT) named in texts, e.g., IOC content, it can successfully find relevant documents and modify its path accord ingly. It yields harvest rates of up to 52%, which are, to the best of our knowledge, better than the current state of the art. The results and source code will be made publicly available upon acceptance.
comment: 11 pages, 9 figures, 5 tables
♻ ☆ Human-like conceptual representations emerge from language prediction
People acquire concepts through rich physical and social experiences and use them to understand the world. In contrast, large language models (LLMs), trained exclusively through next-token prediction over language data, exhibit remarkably human-like behaviors. Are these models developing concepts akin to humans, and if so, how are such concepts represented and organized? To address these questions, we reframed the classic reverse dictionary task to simulate human concept inference in context and investigated the emergence of human-like conceptual representations within LLMs. Our results demonstrate that LLMs can flexibly derive concepts from linguistic descriptions in relation to contextual cues about other concepts. The derived representations converged towards a shared, context-independent structure that effectively predicted human behavior across key psychological phenomena, including computation of similarities, categories and semantic scales. Moreover, these representations aligned well with neural activity patterns in the human brain, even in response to visual rather than linguistic stimuli, providing evidence for biological plausibility. These findings establish that structured, human-like conceptual representations can naturally emerge from language prediction without real-world grounding. More broadly, our work positions LLMs as promising computational tools for understanding complex human cognition and paves the way for better alignment between artificial and human intelligence.
comment: 51 pages
♻ ☆ Enhancing Multimodal Sentiment Analysis for Missing Modality through Self-Distillation and Unified Modality Cross-Attention
In multimodal sentiment analysis, collecting text data is often more challenging than video or audio due to higher annotation costs and inconsistent automatic speech recognition (ASR) quality. To address this challenge, our study has developed a robust model that effectively integrates multimodal sentiment information, even in the absence of text modality. Specifically, we have developed a Double-Flow Self-Distillation Framework, including Unified Modality Cross-Attention (UMCA) and Modality Imagination Autoencoder (MIA), which excels at processing both scenarios with complete modalities and those with missing text modality. In detail, when the text modality is missing, our framework uses the LLM-based model to simulate the text representation from the audio modality, while the MIA module supplements information from the other two modalities to make the simulated text representation similar to the real text representation. To further align the simulated and real representations, and to enable the model to capture the continuous nature of sample orders in sentiment valence regression tasks, we have also introduced the Rank-N Contrast (RNC) loss function. When testing on the CMU-MOSEI, our model achieved outstanding performance on MAE and significantly outperformed other models when text modality is missing. The code is available at: https://github.com/WarmCongee/SDUMC
♻ ☆ JanusFlow: Harmonizing Autoregression and Rectified Flow for Unified Multimodal Understanding and Generation CVPR 2025
We present JanusFlow, a powerful framework that unifies image understanding and generation in a single model. JanusFlow introduces a minimalist architecture that integrates autoregressive language models with rectified flow, a state-of-the-art method in generative modeling. Our key finding demonstrates that rectified flow can be straightforwardly trained within the large language model framework, eliminating the need for complex architectural modifications. To further improve the performance of our unified model, we adopt two key strategies: (i) decoupling the understanding and generation encoders, and (ii) aligning their representations during unified training. Extensive experiments show that JanusFlow achieves comparable or superior performance to specialized models in their respective domains, while significantly outperforming existing unified approaches across standard benchmarks. This work represents a step toward more efficient and versatile vision-language models.
comment: Accepted by CVPR 2025
♻ ☆ CSCE: Boosting LLM Reasoning by Simultaneous Enhancing of Causal Significance and Consistency
Chain-based reasoning methods like chain of thought (CoT) play a rising role in solving reasoning tasks for large language models (LLMs). However, the causal hallucinations between a step of reasoning and corresponding state transitions are becoming a significant obstacle to advancing LLMs' reasoning capabilities, especially in long-range reasoning tasks. This paper proposes a non-chain-based reasoning framework for simultaneous consideration of causal significance and consistency, i.e., the Causal Significance and Consistency Enhancer (CSCE). We customize LLM's loss function utilizing treatment effect assessments to enhance its reasoning ability from two aspects: causal significance and consistency. This ensures that the model captures essential causal relationships and maintains robust and consistent performance across various scenarios. Additionally, we transform the reasoning process from the cascading multiple one-step reasoning commonly used in Chain-Based methods, like CoT, to a causal-enhanced method that outputs the entire reasoning process in one go, further improving the model's reasoning efficiency. Extensive experiments show that our method improves both the reasoning success rate and speed. These improvements further demonstrate that non-chain-based methods can also aid LLMs in completing reasoning tasks.
comment: 6 pages,4 figures. This paper has been accepted for presentation at IEEE International Conference on Multimedia & Expo 2025
♻ ☆ TriG-NER: Triplet-Grid Framework for Discontinuous Named Entity Recognition WWW'25
Discontinuous Named Entity Recognition (DNER) presents a challenging problem where entities may be scattered across multiple non-adjacent tokens, making traditional sequence labelling approaches inadequate. Existing methods predominantly rely on custom tagging schemes to handle these discontinuous entities, resulting in models tightly coupled to specific tagging strategies and lacking generalisability across diverse datasets. To address these challenges, we propose TriG-NER, a novel Triplet-Grid Framework that introduces a generalisable approach to learning robust token-level representations for discontinuous entity extraction. Our framework applies triplet loss at the token level, where similarity is defined by word pairs existing within the same entity, effectively pulling together similar and pushing apart dissimilar ones. This approach enhances entity boundary detection and reduces the dependency on specific tagging schemes by focusing on word-pair relationships within a flexible grid structure. We evaluate TriG-NER on three benchmark DNER datasets and demonstrate significant improvements over existing grid-based architectures. These results underscore our framework's effectiveness in capturing complex entity structures and its adaptability to various tagging schemes, setting a new benchmark for discontinuous entity extraction.
comment: Accepted at The ACM Web Conference WWW'25. Code available at https://github.com/adlnlp/trig_ner
♻ ☆ EPO: Explicit Policy Optimization for Strategic Reasoning in LLMs via Reinforcement Learning
Large Language Models (LLMs) have shown impressive reasoning capabilities in well-defined problems with clear solutions, such as mathematics and coding. However, they still struggle with complex real-world scenarios like business negotiations, which require strategic reasoning-an ability to navigate dynamic environments and align long-term goals amidst uncertainty. Existing methods for strategic reasoning face challenges in adaptability, scalability, and transferring strategies to new contexts. To address these issues, we propose explicit policy optimization (EPO) for strategic reasoning, featuring an LLM that provides strategies in open-ended action space and can be plugged into arbitrary LLM agents to motivate goal-directed behavior. To improve adaptability and policy transferability, we train the strategic reasoning model via multi-turn reinforcement learning (RL) using process rewards and iterative self-play, without supervised fine-tuning (SFT) as a preliminary step. Experiments across social and physical domains demonstrate EPO's ability of long-term goal alignment through enhanced strategic reasoning, achieving state-of-the-art performance on social dialogue and web navigation tasks. Our findings reveal various collaborative reasoning mechanisms emergent in EPO and its effectiveness in generating novel strategies, underscoring its potential for strategic reasoning in real-world applications.
comment: 22 pages, 4 figures
♻ ☆ Leveraging Allophony in Self-Supervised Speech Models for Atypical Pronunciation Assessment NAACL 2025
Allophony refers to the variation in the phonetic realization of a phoneme based on its phonetic environment. Modeling allophones is crucial for atypical pronunciation assessment, which involves distinguishing atypical from typical pronunciations. However, recent phoneme classifier-based approaches often simplify this by treating various realizations as a single phoneme, bypassing the complexity of modeling allophonic variation. Motivated by the acoustic modeling capabilities of frozen self-supervised speech model (S3M) features, we propose MixGoP, a novel approach that leverages Gaussian mixture models to model phoneme distributions with multiple subclusters. Our experiments show that MixGoP achieves state-of-the-art performance across four out of five datasets, including dysarthric and non-native speech. Our analysis further suggests that S3M features capture allophonic variation more effectively than MFCCs and Mel spectrograms, highlighting the benefits of integrating MixGoP with S3M features.
comment: Accepted to NAACL 2025. Codebase available at https://github.com/juice500ml/acoustic-units-for-ood
♻ ☆ A Survey on Personalized Alignment -- The Missing Piece for Large Language Models in Real-World Applications
Large Language Models (LLMs) have demonstrated remarkable capabilities, yet their transition to real-world applications reveals a critical limitation: the inability to adapt to individual preferences while maintaining alignment with universal human values. Current alignment techniques adopt a one-size-fits-all approach that fails to accommodate users' diverse backgrounds and needs. This paper presents the first comprehensive survey of personalized alignment-a paradigm that enables LLMs to adapt their behavior within ethical boundaries based on individual preferences. We propose a unified framework comprising preference memory management, personalized generation, and feedback-based alignment, systematically analyzing implementation approaches and evaluating their effectiveness across various scenarios. By examining current techniques, potential risks, and future challenges, this survey provides a structured foundation for developing more adaptable and ethically-aligned LLMs.
comment: 9 pages
♻ ☆ Similarity-Dissimilarity Loss for Multi-label Supervised Contrastive Learning
Supervised contrastive learning has been explored in making use of label information for multi-label classification, but determining positive samples in multi-label scenario remains challenging. Previous studies have examined strategies for identifying positive samples, considering label overlap proportion between anchors and samples. However, they ignore various relations between given anchors and samples, as well as how to dynamically adjust the weights in contrastive loss functions based on different relations, leading to great ambiguity. In this paper, we introduce five distinct relations between multi-label samples and propose a Similarity-Dissimilarity Loss with contrastive learning for multi-label classification. Our loss function re-weights the loss by computing the similarity and dissimilarity between positive samples and a given anchor based on the introduced relations. We mainly conduct experiments for multi-label text classification on MIMIC datasets, then further extend the evaluation on MS-COCO. The Experimental results show that our proposed loss effectively improves the performance on all encoders under supervised contrastive learning paradigm, demonstrating its effectiveness and robustness.
♻ ☆ R2-KG: General-Purpose Dual-Agent Framework for Reliable Reasoning on Knowledge Graphs
Recent studies have combined Large Language Models (LLMs) with Knowledge Graphs (KGs) to enhance reasoning, improving inference accuracy without additional training while mitigating hallucination. However, existing frameworks are often rigid, struggling to adapt to KG or task changes. They also rely heavily on powerful LLMs for reliable (i.e., trustworthy) reasoning. To address this, We introduce R2-KG, a plug-and-play, dual-agent framework that separates reasoning into two roles: an Operator (a low-capacity LLM) that gathers evidence and a Supervisor (a high-capacity LLM) that makes final judgments. This design is cost-efficient for LLM inference while still maintaining strong reasoning accuracy. Additionally, R2-KG employs an Abstention mechanism, generating answers only when sufficient evidence is collected from KG, which significantly enhances reliability. Experiments across multiple KG-based reasoning tasks show that R2-KG consistently outperforms baselines in both accuracy and reliability, regardless of the inherent capability of LLMs used as the Operator. Further experiments reveal that the single-agent version of R2-KG, equipped with a strict self-consistency strategy, achieves significantly higher-than-baseline reliability while reducing inference cost. However, it also leads to a higher abstention rate in complex KGs. Our findings establish R2-KG as a flexible and cost-effective solution for KG-based reasoning. It reduces reliance on high-capacity LLMs while ensuring trustworthy inference.
♻ ☆ Design and Implementation of an FPGA-Based Tiled Matrix Multiplication Accelerator for Transformer Self-Attention on the Xilinx KV260 SoM
Transformer-based large language models (LLMs) rely heavily on intensive matrix multiplications for attention and feed-forward layers, with the Q, K, and V linear projections in the Multi-Head Self-Attention (MHA) module constituting a decisive performance bottleneck. In this work, we introduce a highly optimized tiled matrix multiplication accelerator on a resource-constrained Xilinx KV260 FPGA that not only addresses this challenge but sets a new standard for efficiency and performance. Our design exploits persistent on-chip storage, a robust two-level tiling strategy for maximal data reuse, and a systolic-like unrolled compute engine that together deliver unparalleled speed and energy efficiency. Integrated with DistilBERT for Q, K, and V projections, our accelerator achieves an unequivocal 7x speedup over ARM CPU implementations (PyTorch) and an extraordinary 200x improvement over naive NumPy, reaching a throughput of up to 3.1~GFLOPs for matrix multiplications on (64,768) x (768,3072) matrices while operating at a conservative 100 MHz. These results decisively demonstrate the transformative potential of FPGA-based acceleration for critical Transformer operations, paving the way for scalable and energy-efficient deep learning inference on edge devices.
comment: 7 pages, 4 figures, 2 tables. Prepared in ACM conference style. Preprint under review
♻ ☆ Advancing Mathematical Reasoning in Language Models: The Impact of Problem-Solving Data, Data Synthesis Methods, and Training Stages ICLR 2025
Mathematical reasoning remains a challenging area for large language models (LLMs), prompting the development of math-specific LLMs such as LLEMMA, DeepSeekMath, and Qwen2-Math, among others. These models typically follow a two-stage training paradigm: pre-training with math-related corpora and post-training with problem datasets for supervised fine-tuning (SFT). Despite these efforts, the improvements in mathematical reasoning achieved through continued pre-training (CPT) are often less significant compared to those obtained via SFT. This study addresses this discrepancy by exploring alternative strategies during the pre-training phase, focusing on the use of problem-solving data over general mathematical corpora. We investigate three primary research questions: (1) Can problem-solving data enhance the model's mathematical reasoning capabilities more effectively than general mathematical corpora during CPT? (2) Are synthetic data from the same source equally effective, and which synthesis methods are most efficient? (3) How do the capabilities developed from the same problem-solving data differ between the CPT and SFT stages, and what factors contribute to these differences? Our findings indicate that problem-solving data significantly enhances the model's mathematical capabilities compared to general mathematical corpora. We also identify effective data synthesis methods, demonstrating that the tutorship amplification synthesis method achieves the best performance. Furthermore, while SFT facilitates instruction-following abilities, it underperforms compared to CPT with the same data, which can be partially attributed to its poor learning capacity for more challenging problem-solving data. These insights provide valuable guidance for optimizing the mathematical reasoning capabilities of LLMs, culminating in our development of a powerful mathematical base model called MathGPT-8B.
comment: ICLR 2025
♻ ☆ ReTaKe: Reducing Temporal and Knowledge Redundancy for Long Video Understanding
Video Large Language Models (VideoLLMs) have made significant strides in video understanding but struggle with long videos due to the limitations of their backbone LLMs. Existing solutions rely on length extrapolation, which is memory-constrained, or visual token compression, which primarily leverages low-level temporal redundancy while overlooking the more effective high-level knowledge redundancy. To address this, we propose $\textbf{ReTaKe}$, a training-free method with two novel modules DPSelect and PivotKV, to jointly reduce both temporal visual redundancy and knowledge redundancy for video compression. To align with the way of human temporal perception, DPSelect identifies keyframes based on inter-frame distance peaks. To leverage LLMs' learned prior knowledge, PivotKV marks the keyframes as pivots and compress non-pivot frames by pruning low-attention tokens in their KV cache. ReTaKe enables VideoLLMs to process 8 times longer frames (up to 2048), outperforming similar-sized models by 3-5% and even rivaling much larger ones on VideoMME, MLVU, LongVideoBench, and LVBench. Moreover, by overlapping compression operations with prefilling, ReTaKe introduces only ~10% prefilling latency overhead while reducing decoding latency by ~20%. Our code is available at https://github.com/SCZwangxiao/video-ReTaKe.
comment: Rewrite the methods section. Add more ablation studies and results in LongVideoBench. Update metadata
♻ ☆ Inside-Out: Hidden Factual Knowledge in LLMs
This work presents a framework for assessing whether large language models (LLMs) encode more factual knowledge in their parameters than what they express in their outputs. While a few studies hint at this possibility, none has clearly defined or demonstrated this phenomenon. We first propose a formal definition of knowledge, quantifying it for a given question as the fraction of correct-incorrect answer pairs where the correct one is ranked higher. This gives rise to external and internal knowledge, depending on the information used to score individual answer candidates: either the model's observable token-level probabilities or its intermediate computations. Hidden knowledge arises when internal knowledge exceeds external knowledge. We then present a case study, applying this framework to three popular open-weights LLMs in a closed-book QA setup. Our results indicate that: (1) LLMs consistently encode more factual knowledge internally than what they express externally, with an average relative gap of 40%. (2) Surprisingly, some knowledge is so deeply hidden that a model can internally know an answer perfectly, yet fail to generate it even once, despite large-scale repeated sampling of 1,000 answers. This reveals fundamental limitations in the generation capabilities of LLMs, which (3) put a practical constraint on scaling test-time compute via repeated answer sampling in closed-book QA: significant performance improvements remain inaccessible because some answers are practically never sampled, yet if they were, we would be guaranteed to rank them first.
♻ ☆ Training and Evaluating with Human Label Variation: An Empirical Study
Human label variation (HLV) challenges the standard assumption that a labelled instance has a single ground truth, instead embracing the natural variation in human annotation to train and evaluate models. While various training methods and metrics for HLV have been proposed, it is still unclear which methods and metrics perform best in what settings. We propose new evaluation metrics for HLV leveraging fuzzy set theory. Since these new proposed metrics are differentiable, we then in turn experiment with employing these metrics as training objectives. We conduct an extensive study over 6 HLV datasets testing 14 training methods and 6 evaluation metrics. We find that training on either disaggregated annotations or soft labels performs best across metrics, outperforming training using the proposed training objectives with differentiable metrics. We also show that our proposed soft metric is more interpretable and correlates best with human preference.
comment: 25 pages
Computer Vision and Pattern Recognition 256
☆ Target-Aware Video Diffusion Models
We present a target-aware video diffusion model that generates videos from an input image in which an actor interacts with a specified target while performing a desired action. The target is defined by a segmentation mask and the desired action is described via a text prompt. Unlike existing controllable image-to-video diffusion models that often rely on dense structural or motion cues to guide the actor's movements toward the target, our target-aware model requires only a simple mask to indicate the target, leveraging the generalization capabilities of pretrained models to produce plausible actions. This makes our method particularly effective for human-object interaction (HOI) scenarios, where providing precise action guidance is challenging, and further enables the use of video diffusion models for high-level action planning in applications such as robotics. We build our target-aware model by extending a baseline model to incorporate the target mask as an additional input. To enforce target awareness, we introduce a special token that encodes the target's spatial information within the text prompt. We then fine-tune the model with our curated dataset using a novel cross-attention loss that aligns the cross-attention maps associated with this token with the input target mask. To further improve performance, we selectively apply this loss to the most semantically relevant transformer blocks and attention regions. Experimental results show that our target-aware model outperforms existing solutions in generating videos where actors interact accurately with the specified targets. We further demonstrate its efficacy in two downstream applications: video content creation and zero-shot 3D HOI motion synthesis.
comment: The project page is available at https://taeksuu.github.io/tavid/
☆ Equivariant Image Modeling
Current generative models, such as autoregressive and diffusion approaches, decompose high-dimensional data distribution learning into a series of simpler subtasks. However, inherent conflicts arise during the joint optimization of these subtasks, and existing solutions fail to resolve such conflicts without sacrificing efficiency or scalability. We propose a novel equivariant image modeling framework that inherently aligns optimization targets across subtasks by leveraging the translation invariance of natural visual signals. Our method introduces (1) column-wise tokenization which enhances translational symmetry along the horizontal axis, and (2) windowed causal attention which enforces consistent contextual relationships across positions. Evaluated on class-conditioned ImageNet generation at 256x256 resolution, our approach achieves performance comparable to state-of-the-art AR models while using fewer computational resources. Systematic analysis demonstrates that enhanced equivariance reduces inter-task conflicts, significantly improving zero-shot generalization and enabling ultra-long image synthesis. This work establishes the first framework for task-aligned decomposition in generative modeling, offering insights into efficient parameter sharing and conflict-free optimization. The code and models are publicly available at https://github.com/drx-code/EquivariantModeling.
☆ Tuning-Free Amodal Segmentation via the Occlusion-Free Bias of Inpainting Models
Amodal segmentation aims to predict segmentation masks for both the visible and occluded regions of an object. Most existing works formulate this as a supervised learning problem, requiring manually annotated amodal masks or synthetic training data. Consequently, their performance depends on the quality of the datasets, which often lack diversity and scale. This work introduces a tuning-free approach that repurposes pretrained diffusion-based inpainting models for amodal segmentation. Our approach is motivated by the "occlusion-free bias" of inpainting models, i.e., the inpainted objects tend to be complete objects without occlusions. Specifically, we reconstruct the occluded regions of an object via inpainting and then apply segmentation, all without additional training or fine-tuning. Experiments on five datasets demonstrate the generalizability and robustness of our approach. On average, our approach achieves 5.3% more accurate masks over the state-of-the-art.
☆ Aether: Geometric-Aware Unified World Modeling
The integration of geometric reconstruction and generative modeling remains a critical challenge in developing AI systems capable of human-like spatial reasoning. This paper proposes Aether, a unified framework that enables geometry-aware reasoning in world models by jointly optimizing three core capabilities: (1) 4D dynamic reconstruction, (2) action-conditioned video prediction, and (3) goal-conditioned visual planning. Through task-interleaved feature learning, Aether achieves synergistic knowledge sharing across reconstruction, prediction, and planning objectives. Building upon video generation models, our framework demonstrates unprecedented synthetic-to-real generalization despite never observing real-world data during training. Furthermore, our approach achieves zero-shot generalization in both action following and reconstruction tasks, thanks to its intrinsic geometric modeling. Remarkably, even without real-world data, its reconstruction performance far exceeds that of domain-specific models. Additionally, Aether leverages a geometry-informed action space to seamlessly translate predictions into actions, enabling effective autonomous trajectory planning. We hope our work inspires the community to explore new frontiers in physically-reasonable world modeling and its applications.
comment: Project Page: https://aether-world.github.io/
☆ DINO in the Room: Leveraging 2D Foundation Models for 3D Segmentation
Vision foundation models (VFMs) trained on large-scale image datasets provide high-quality features that have significantly advanced 2D visual recognition. However, their potential in 3D vision remains largely untapped, despite the common availability of 2D images alongside 3D point cloud datasets. While significant research has been dedicated to 2D-3D fusion, recent state-of-the-art 3D methods predominantly focus on 3D data, leaving the integration of VFMs into 3D models underexplored. In this work, we challenge this trend by introducing DITR, a simple yet effective approach that extracts 2D foundation model features, projects them to 3D, and finally injects them into a 3D point cloud segmentation model. DITR achieves state-of-the-art results on both indoor and outdoor 3D semantic segmentation benchmarks. To enable the use of VFMs even when images are unavailable during inference, we further propose to distill 2D foundation models into a 3D backbone as a pretraining task. By initializing the 3D backbone with knowledge distilled from 2D VFMs, we create a strong basis for downstream 3D segmentation tasks, ultimately boosting performance across various datasets.
comment: Project page at https://vision.rwth-aachen.de/DITR
☆ SlowFast-LLaVA-1.5: A Family of Token-Efficient Video Large Language Models for Long-Form Video Understanding
We introduce SlowFast-LLaVA-1.5 (abbreviated as SF-LLaVA-1.5), a family of video large language models (LLMs) offering a token-efficient solution for long-form video understanding. This model family employs the two-stream SlowFast mechanism, enabling efficient modeling of long-range temporal context to meet the demand for lightweight, mobile-friendly Video LLMs. We provide models ranging from 1B to 7B parameters, optimized through a streamlined training pipeline and a high-quality data mixture composed of publicly available datasets. Experimental results demonstrate that SF-LLaVA-1.5 achieves competitive performance on a wide range of video and image benchmarks, with robust results across all model sizes. Notably, SF-LLaVA-1.5 achieves state-of-the-art results in long-form video understanding (e.g., LongVideoBench and MLVU) and excels at small scales (1B and 3B) across various video benchmarks.
comment: Technical report
☆ Video-T1: Test-Time Scaling for Video Generation
With the scale capability of increasing training data, model size, and computational cost, video generation has achieved impressive results in digital creation, enabling users to express creativity across various domains. Recently, researchers in Large Language Models (LLMs) have expanded the scaling to test-time, which can significantly improve LLM performance by using more inference-time computation. Instead of scaling up video foundation models through expensive training costs, we explore the power of Test-Time Scaling (TTS) in video generation, aiming to answer the question: if a video generation model is allowed to use non-trivial amount of inference-time compute, how much can it improve generation quality given a challenging text prompt. In this work, we reinterpret the test-time scaling of video generation as a searching problem to sample better trajectories from Gaussian noise space to the target video distribution. Specifically, we build the search space with test-time verifiers to provide feedback and heuristic algorithms to guide searching process. Given a text prompt, we first explore an intuitive linear search strategy by increasing noise candidates at inference time. As full-step denoising all frames simultaneously requires heavy test-time computation costs, we further design a more efficient TTS method for video generation called Tree-of-Frames (ToF) that adaptively expands and prunes video branches in an autoregressive manner. Extensive experiments on text-conditioned video generation benchmarks demonstrate that increasing test-time compute consistently leads to significant improvements in the quality of videos. Project page: https://liuff19.github.io/Video-T1
comment: Project page: https://liuff19.github.io/Video-T1
☆ Training-free Diffusion Acceleration with Bottleneck Sampling
Diffusion models have demonstrated remarkable capabilities in visual content generation but remain challenging to deploy due to their high computational cost during inference. This computational burden primarily arises from the quadratic complexity of self-attention with respect to image or video resolution. While existing acceleration methods often compromise output quality or necessitate costly retraining, we observe that most diffusion models are pre-trained at lower resolutions, presenting an opportunity to exploit these low-resolution priors for more efficient inference without degrading performance. In this work, we introduce Bottleneck Sampling, a training-free framework that leverages low-resolution priors to reduce computational overhead while preserving output fidelity. Bottleneck Sampling follows a high-low-high denoising workflow: it performs high-resolution denoising in the initial and final stages while operating at lower resolutions in intermediate steps. To mitigate aliasing and blurring artifacts, we further refine the resolution transition points and adaptively shift the denoising timesteps at each stage. We evaluate Bottleneck Sampling on both image and video generation tasks, where extensive experiments demonstrate that it accelerates inference by up to 3$\times$ for image generation and 2.5$\times$ for video generation, all while maintaining output quality comparable to the standard full-resolution sampling process across multiple evaluation metrics. Code is available at: https://github.com/tyfeld/Bottleneck-Sampling
comment: Code Repo: https://github.com/tyfeld/Bottleneck-Sampling ,Project Page: https://tyfeld.github.io/BottleneckSampling.github.io/
AdaWorld: Learning Adaptable World Models with Latent Actions
World models aim to learn action-controlled prediction models and have proven essential for the development of intelligent agents. However, most existing world models rely heavily on substantial action-labeled data and costly training, making it challenging to adapt to novel environments with heterogeneous actions through limited interactions. This limitation can hinder their applicability across broader domains. To overcome this challenge, we propose AdaWorld, an innovative world model learning approach that enables efficient adaptation. The key idea is to incorporate action information during the pretraining of world models. This is achieved by extracting latent actions from videos in a self-supervised manner, capturing the most critical transitions between frames. We then develop an autoregressive world model that conditions on these latent actions. This learning paradigm enables highly adaptable world models, facilitating efficient transfer and learning of new actions even with limited interactions and finetuning. Our comprehensive experiments across multiple environments demonstrate that AdaWorld achieves superior performance in both simulation quality and visual planning.
comment: Project page: https://adaptable-world-model.github.io/
☆ SyncVP: Joint Diffusion for Synchronous Multi-Modal Video Prediction
Predicting future video frames is essential for decision-making systems, yet RGB frames alone often lack the information needed to fully capture the underlying complexities of the real world. To address this limitation, we propose a multi-modal framework for Synchronous Video Prediction (SyncVP) that incorporates complementary data modalities, enhancing the richness and accuracy of future predictions. SyncVP builds on pre-trained modality-specific diffusion models and introduces an efficient spatio-temporal cross-attention module to enable effective information sharing across modalities. We evaluate SyncVP on standard benchmark datasets, such as Cityscapes and BAIR, using depth as an additional modality. We furthermore demonstrate its generalization to other modalities on SYNTHIA with semantic information and ERA5-Land with climate data. Notably, SyncVP achieves state-of-the-art performance, even in scenarios where only one modality is present, demonstrating its robustness and potential for a wide range of applications.
☆ CoMP: Continual Multimodal Pre-training for Vision Foundation Models
Pre-trained Vision Foundation Models (VFMs) provide strong visual representations for a wide range of applications. In this paper, we continually pre-train prevailing VFMs in a multimodal manner such that they can effortlessly process visual inputs of varying sizes and produce visual representations that are more aligned with language representations, regardless of their original pre-training process. To this end, we introduce CoMP, a carefully designed multimodal pre-training pipeline. CoMP uses a Continual Rotary Position Embedding to support native resolution continual pre-training, and an Alignment Loss between visual and textual features through language prototypes to align multimodal representations. By three-stage training, our VFMs achieve remarkable improvements not only in multimodal understanding but also in other downstream tasks such as classification and segmentation. Remarkably, CoMP-SigLIP achieves scores of 66.7 on ChartQA and 75.9 on DocVQA with a 0.5B LLM, while maintaining an 87.4% accuracy on ImageNet-1K and a 49.5 mIoU on ADE20K under frozen chunk evaluation.
comment: Code is available in https://github.com/SliMM-X/CoMP-MM
☆ Video SimpleQA: Towards Factuality Evaluation in Large Video Language Models
Recent advancements in Large Video Language Models (LVLMs) have highlighted their potential for multi-modal understanding, yet evaluating their factual grounding in video contexts remains a critical unsolved challenge. To address this gap, we introduce Video SimpleQA, the first comprehensive benchmark tailored for factuality evaluation of LVLMs. Our work distinguishes from existing video benchmarks through the following key features: 1) Knowledge required: demanding integration of external knowledge beyond the explicit narrative; 2) Fact-seeking question: targeting objective, undisputed events or relationships, avoiding subjective interpretation; 3) Definitive & short-form answer: Answers are crafted as unambiguous and definitively correct in a short format, enabling automated evaluation through LLM-as-a-judge frameworks with minimal scoring variance; 4) External-source verified: All annotations undergo rigorous validation against authoritative external references to ensure the reliability; 5) Temporal reasoning required: The annotated question types encompass both static single-frame understanding and dynamic temporal reasoning, explicitly evaluating LVLMs factuality under the long-context dependencies. We extensively evaluate 41 state-of-the-art LVLMs and summarize key findings as follows: 1) Current LVLMs exhibit notable deficiencies in factual adherence, particularly for open-source models. The best-performing model Gemini-1.5-Pro achieves merely an F-score of 54.4%; 2) Test-time compute paradigms show insignificant performance gains, revealing fundamental constraints for enhancing factuality through post-hoc computation; 3) Retrieval-Augmented Generation demonstrates consistent improvements at the cost of additional inference time overhead, presenting a critical efficiency-performance trade-off.
comment: 24 pages
☆ Building Blocks for Robust and Effective Semi-Supervised Real-World Object Detection
Semi-supervised object detection (SSOD) based on pseudo-labeling significantly reduces dependence on large labeled datasets by effectively leveraging both labeled and unlabeled data. However, real-world applications of SSOD often face critical challenges, including class imbalance, label noise, and labeling errors. We present an in-depth analysis of SSOD under real-world conditions, uncovering causes of suboptimal pseudo-labeling and key trade-offs between label quality and quantity. Based on our findings, we propose four building blocks that can be seamlessly integrated into an SSOD framework. Rare Class Collage (RCC): a data augmentation method that enhances the representation of rare classes by creating collages of rare objects. Rare Class Focus (RCF): a stratified batch sampling strategy that ensures a more balanced representation of all classes during training. Ground Truth Label Correction (GLC): a label refinement method that identifies and corrects false, missing, and noisy ground truth labels by leveraging the consistency of teacher model predictions. Pseudo-Label Selection (PLS): a selection method for removing low-quality pseudo-labeled images, guided by a novel metric estimating the missing detection rate while accounting for class rarity. We validate our methods through comprehensive experiments on autonomous driving datasets, resulting in up to 6% increase in SSOD performance. Overall, our investigation and novel, data-centric, and broadly applicable building blocks enable robust and effective SSOD in complex, real-world scenarios. Code is available at https://mos-ks.github.io/publications.
comment: Accepted to Transactions on Machine Learning Research (TMLR). OpenReview: https://openreview.net/forum?id=vRYt8QLKqK
☆ Online 3D Scene Reconstruction Using Neural Object Priors 3DV 2025
This paper addresses the problem of reconstructing a scene online at the level of objects given an RGB-D video sequence. While current object-aware neural implicit representations hold promise, they are limited in online reconstruction efficiency and shape completion. Our main contributions to alleviate the above limitations are twofold. First, we propose a feature grid interpolation mechanism to continuously update grid-based object-centric neural implicit representations as new object parts are revealed. Second, we construct an object library with previously mapped objects in advance and leverage the corresponding shape priors to initialize geometric object models in new videos, subsequently completing them with novel views as well as synthesized past views to avoid losing original object details. Extensive experiments on synthetic environments from the Replica dataset, real-world ScanNet sequences and videos captured in our laboratory demonstrate that our approach outperforms state-of-the-art neural implicit models for this task in terms of reconstruction accuracy and completeness.
comment: 3DV 2025. Project page: https://www.di.ens.fr/willow/research/online-scene-reconstruction/
☆ CFG-Zero*: Improved Classifier-Free Guidance for Flow Matching Models
Classifier-Free Guidance (CFG) is a widely adopted technique in diffusion/flow models to improve image fidelity and controllability. In this work, we first analytically study the effect of CFG on flow matching models trained on Gaussian mixtures where the ground-truth flow can be derived. We observe that in the early stages of training, when the flow estimation is inaccurate, CFG directs samples toward incorrect trajectories. Building on this observation, we propose CFG-Zero*, an improved CFG with two contributions: (a) optimized scale, where a scalar is optimized to correct for the inaccuracies in the estimated velocity, hence the * in the name; and (b) zero-init, which involves zeroing out the first few steps of the ODE solver. Experiments on both text-to-image (Lumina-Next, Stable Diffusion 3, and Flux) and text-to-video (Wan-2.1) generation demonstrate that CFG-Zero* consistently outperforms CFG, highlighting its effectiveness in guiding Flow Matching models. (Code is available at github.com/WeichenFan/CFG-Zero-star)
☆ Efficient and Accurate Scene Text Recognition with Cascaded-Transformers
In recent years, vision transformers with text decoder have demonstrated remarkable performance on Scene Text Recognition (STR) due to their ability to capture long-range dependencies and contextual relationships with high learning capacity. However, the computational and memory demands of these models are significant, limiting their deployment in resource-constrained applications. To address this challenge, we propose an efficient and accurate STR system. Specifically, we focus on improving the efficiency of encoder models by introducing a cascaded-transformers structure. This structure progressively reduces the vision token size during the encoding step, effectively eliminating redundant tokens and reducing computational cost. Our experimental results confirm that our STR system achieves comparable performance to state-of-the-art baselines while substantially decreasing computational requirements. In particular, for large-models, the accuracy remains same, 92.77 to 92.68, while computational complexity is almost halved with our structure.
comment: Accepted to ACM-MMSys2025
☆ Seeing Speech and Sound: Distinguishing and Locating Audios in Visual Scenes CVPR 2025
We present a unified model capable of simultaneously grounding both spoken language and non-speech sounds within a visual scene, addressing key limitations in current audio-visual grounding models. Existing approaches are typically limited to handling either speech or non-speech sounds independently, or at best, together but sequentially without mixing. This limitation prevents them from capturing the complexity of real-world audio sources that are often mixed. Our approach introduces a 'mix-and-separate' framework with audio-visual alignment objectives that jointly learn correspondence and disentanglement using mixed audio. Through these objectives, our model learns to produce distinct embeddings for each audio type, enabling effective disentanglement and grounding across mixed audio sources. Additionally, we created a new dataset to evaluate simultaneous grounding of mixed audio sources, demonstrating that our model outperforms prior methods. Our approach also achieves comparable or better performance in standard segmentation and cross-modal retrieval tasks, highlighting the benefits of our mix-and-separate approach.
comment: CVPR 2025
☆ A semantic communication-based workload-adjustable transceiver for wireless AI-generated content (AIGC) delivery
With the significant advances in generative AI (GAI) and the proliferation of mobile devices, providing high-quality AI-generated content (AIGC) services via wireless networks is becoming the future direction. However, the primary challenges of AIGC service delivery in wireless networks lie in unstable channels, limited bandwidth resources, and unevenly distributed computational resources. In this paper, we employ semantic communication (SemCom) in diffusion-based GAI models to propose a Resource-aware wOrkload-adjUstable TransceivEr (ROUTE) for AIGC delivery in dynamic wireless networks. Specifically, to relieve the communication resource bottleneck, SemCom is utilized to prioritize semantic information of the generated content. Then, to improve computational resource utilization in both edge and local and reduce AIGC semantic distortion in transmission, modified diffusion-based models are applied to adjust the computing workload and semantic density in cooperative content generation. Simulations verify the superiority of our proposed ROUTE in terms of latency and content quality compared to conventional AIGC approaches.
☆ Efficient Self-Supervised Adaptation for Medical Image Analysis
Self-supervised adaptation (SSA) improves foundation model transfer to medical domains but is computationally prohibitive. Although parameter efficient fine-tuning methods such as LoRA have been explored for supervised adaptation, their effectiveness for SSA remains unknown. In this work, we introduce efficient self-supervised adaptation (ESSA), a framework that applies parameter-efficient fine-tuning techniques to SSA with the aim of reducing computational cost and improving adaptation performance. Among the methods tested, Attention Projection Layer Adaptation (APLA) sets a new state-of-the-art, consistently surpassing full-parameter SSA and supervised fine-tuning across diverse medical tasks, while reducing GPU memory by up to 40.1% and increasing training throughput by 25.2%, all while maintaining inference efficiency.
☆ Curriculum Coarse-to-Fine Selection for High-IPC Dataset Distillation CVPR2025
Dataset distillation (DD) excels in synthesizing a small number of images per class (IPC) but struggles to maintain its effectiveness in high-IPC settings. Recent works on dataset distillation demonstrate that combining distilled and real data can mitigate the effectiveness decay. However, our analysis of the combination paradigm reveals that the current one-shot and independent selection mechanism induces an incompatibility issue between distilled and real images. To address this issue, we introduce a novel curriculum coarse-to-fine selection (CCFS) method for efficient high-IPC dataset distillation. CCFS employs a curriculum selection framework for real data selection, where we leverage a coarse-to-fine strategy to select appropriate real data based on the current synthetic dataset in each curriculum. Extensive experiments validate CCFS, surpassing the state-of-the-art by +6.6\% on CIFAR-10, +5.8\% on CIFAR-100, and +3.4\% on Tiny-ImageNet under high-IPC settings. Notably, CCFS achieves 60.2\% test accuracy on ResNet-18 with a 20\% compression ratio of Tiny-ImageNet, closely matching full-dataset training with only 0.3\% degradation. Code: https://github.com/CYDaaa30/CCFS.
comment: Accepted by CVPR2025
☆ Exploring the Integration of Key-Value Attention Into Pure and Hybrid Transformers for Semantic Segmentation
While CNNs were long considered state of the art for image processing, the introduction of Transformer architectures has challenged this position. While achieving excellent results in image classification and segmentation, Transformers remain inherently reliant on large training datasets and remain computationally expensive. A newly introduced Transformer derivative named KV Transformer shows promising results in synthetic, NLP, and image classification tasks, while reducing complexity and memory usage. This is especially conducive to use cases where local inference is required, such as medical screening applications. We endeavoured to further evaluate the merit of KV Transformers on semantic segmentation tasks, specifically in the domain of medical imaging. By directly comparing traditional and KV variants of the same base architectures, we provide further insight into the practical tradeoffs of reduced model complexity. We observe a notable reduction in parameter count and multiply accumulate operations, while achieving similar performance from most of the KV variant models when directly compared to their QKV implementation.
comment: 6 pages, 3 figures, Preprint. Final version published in: Bildverarbeitung f\"ur die Medizin 2025, Springer. DOI: https://doi.org/10.1007/978-3-658-47422-5_71
☆ HunyuanPortrait: Implicit Condition Control for Enhanced Portrait Animation CVPR 2025
We introduce HunyuanPortrait, a diffusion-based condition control method that employs implicit representations for highly controllable and lifelike portrait animation. Given a single portrait image as an appearance reference and video clips as driving templates, HunyuanPortrait can animate the character in the reference image by the facial expression and head pose of the driving videos. In our framework, we utilize pre-trained encoders to achieve the decoupling of portrait motion information and identity in videos. To do so, implicit representation is adopted to encode motion information and is employed as control signals in the animation phase. By leveraging the power of stable video diffusion as the main building block, we carefully design adapter layers to inject control signals into the denoising unet through attention mechanisms. These bring spatial richness of details and temporal consistency. HunyuanPortrait also exhibits strong generalization performance, which can effectively disentangle appearance and motion under different image styles. Our framework outperforms existing methods, demonstrating superior temporal consistency and controllability. Our project is available at https://kkakkkka.github.io/HunyuanPortrait.
comment: CVPR 2025
☆ MC-LLaVA: Multi-Concept Personalized Vision-Language Model
Current vision-language models (VLMs) show exceptional abilities across diverse tasks, such as visual question answering. To enhance user experience, recent studies investigate VLM personalization to understand user-provided concepts. However, they mainly focus on single-concept personalization, neglecting the existence and interplay of multiple concepts, which limits real-world applicability. This paper proposes the first multi-concept personalization paradigm, MC-LLaVA. Specifically, MC-LLaVA employs a multi-concept instruction tuning strategy, effectively integrating multiple concepts in a single training step. To reduce the costs related to joint training, we propose a personalized textual prompt that uses visual token information to initialize concept tokens. Additionally, we introduce a personalized visual prompt during inference, aggregating location confidence maps for enhanced recognition and grounding capabilities. To advance multi-concept personalization research, we further contribute a high-quality instruction tuning dataset. We carefully collect images with multiple characters and objects from movies and manually generate question-answer samples for multi-concept scenarios, featuring superior diversity. Comprehensive qualitative and quantitative experiments demonstrate that MC-LLaVA can achieve impressive multi-concept personalized responses, paving the way for VLMs to become better user-specific assistants. The code and dataset will be publicly available at $\href{https://github.com/arctanxarc/MC-LLaVA}{https://github.com/arctanxarc/MC-LLaVA}$.
comment: The code and dataset will be publicly available at $\href{https://github.com/arctanxarc/MC-LLaVA}{https://github.com/arctanxarc/MC-LLaVA}$
3DSwapping: Texture Swapping For 3D Object From Single Reference Image
3D texture swapping allows for the customization of 3D object textures, enabling efficient and versatile visual transformations in 3D editing. While no dedicated method exists, adapted 2D editing and text-driven 3D editing approaches can serve this purpose. However, 2D editing requires frame-by-frame manipulation, causing inconsistencies across views, while text-driven 3D editing struggles to preserve texture characteristics from reference images. To tackle these challenges, we introduce 3DSwapping, a 3D texture swapping method that integrates: 1) progressive generation, 2) view-consistency gradient guidance, and 3) prompt-tuned gradient guidance. To ensure view consistency, our progressive generation process starts by editing a single reference image and gradually propagates the edits to adjacent views. Our view-consistency gradient guidance further reinforces consistency by conditioning the generation model on feature differences between consistent and inconsistent outputs. To preserve texture characteristics, we introduce prompt-tuning-based gradient guidance, which learns a token that precisely captures the difference between the reference image and the 3D object. This token then guides the editing process, ensuring more consistent texture preservation across views. Overall, 3DSwapping integrates these novel strategies to achieve higher-fidelity texture transfer while preserving structural coherence across multiple viewpoints. Extensive qualitative and quantitative evaluations confirm that our three novel components enable convincing and effective 2D texture swapping for 3D objects. Code will be available upon acceptance.
☆ Learning to segment anatomy and lesions from disparately labeled sources in brain MRI
Segmenting healthy tissue structures alongside lesions in brain Magnetic Resonance Images (MRI) remains a challenge for today's algorithms due to lesion-caused disruption of the anatomy and lack of jointly labeled training datasets, where both healthy tissues and lesions are labeled on the same images. In this paper, we propose a method that is robust to lesion-caused disruptions and can be trained from disparately labeled training sets, i.e., without requiring jointly labeled samples, to automatically segment both. In contrast to prior work, we decouple healthy tissue and lesion segmentation in two paths to leverage multi-sequence acquisitions and merge information with an attention mechanism. During inference, an image-specific adaptation reduces adverse influences of lesion regions on healthy tissue predictions. During training, the adaptation is taken into account through meta-learning and co-training is used to learn from disparately labeled training images. Our model shows an improved performance on several anatomical structures and lesions on a publicly available brain glioblastoma dataset compared to the state-of-the-art segmentation methods.
☆ Dual-domain Multi-path Self-supervised Diffusion Model for Accelerated MRI Reconstruction
Magnetic resonance imaging (MRI) is a vital diagnostic tool, but its inherently long acquisition times reduce clinical efficiency and patient comfort. Recent advancements in deep learning, particularly diffusion models, have improved accelerated MRI reconstruction. However, existing diffusion models' training often relies on fully sampled data, models incur high computational costs, and often lack uncertainty estimation, limiting their clinical applicability. To overcome these challenges, we propose a novel framework, called Dual-domain Multi-path Self-supervised Diffusion Model (DMSM), that integrates a self-supervised dual-domain diffusion model training scheme, a lightweight hybrid attention network for the reconstruction diffusion model, and a multi-path inference strategy, to enhance reconstruction accuracy, efficiency, and explainability. Unlike traditional diffusion-based models, DMSM eliminates the dependency on training from fully sampled data, making it more practical for real-world clinical settings. We evaluated DMSM on two human MRI datasets, demonstrating that it achieves favorable performance over several supervised and self-supervised baselines, particularly in preserving fine anatomical structures and suppressing artifacts under high acceleration factors. Additionally, our model generates uncertainty maps that correlate reasonably well with reconstruction errors, offering valuable clinically interpretable guidance and potentially enhancing diagnostic confidence.
comment: 10 pages, 8 figures, 5 tables
☆ DAGait: Generalized Skeleton-Guided Data Alignment for Gait Recognition
Gait recognition is emerging as a promising and innovative area within the field of computer vision, widely applied to remote person identification. Although existing gait recognition methods have achieved substantial success in controlled laboratory datasets, their performance often declines significantly when transitioning to wild datasets.We argue that the performance gap can be primarily attributed to the spatio-temporal distribution inconsistencies present in wild datasets, where subjects appear at varying angles, positions, and distances across the frames. To achieve accurate gait recognition in the wild, we propose a skeleton-guided silhouette alignment strategy, which uses prior knowledge of the skeletons to perform affine transformations on the corresponding silhouettes.To the best of our knowledge, this is the first study to explore the impact of data alignment on gait recognition. We conducted extensive experiments across multiple datasets and network architectures, and the results demonstrate the significant advantages of our proposed alignment strategy.Specifically, on the challenging Gait3D dataset, our method achieved an average performance improvement of 7.9% across all evaluated networks. Furthermore, our method achieves substantial improvements on cross-domain datasets, with accuracy improvements of up to 24.0%.
☆ Enhanced OoD Detection through Cross-Modal Alignment of Multi-Modal Representations CVPR 2025
Prior research on out-of-distribution detection (OoDD) has primarily focused on single-modality models. Recently, with the advent of large-scale pretrained vision-language models such as CLIP, OoDD methods utilizing such multi-modal representations through zero-shot and prompt learning strategies have emerged. However, these methods typically involve either freezing the pretrained weights or only partially tuning them, which can be suboptimal for downstream datasets. In this paper, we highlight that multi-modal fine-tuning (MMFT) can achieve notable OoDD performance. Despite some recent works demonstrating the impact of fine-tuning methods for OoDD, there remains significant potential for performance improvement. We investigate the limitation of na\"ive fine-tuning methods, examining why they fail to fully leverage the pretrained knowledge. Our empirical analysis suggests that this issue could stem from the modality gap within in-distribution (ID) embeddings. To address this, we propose a training objective that enhances cross-modal alignment by regularizing the distances between image and text embeddings of ID data. This adjustment helps in better utilizing pretrained textual information by aligning similar semantics from different modalities (i.e., text and image) more closely in the hyperspherical representation space. We theoretically demonstrate that the proposed regularization corresponds to the maximum likelihood estimation of an energy-based model on a hypersphere. Utilizing ImageNet-1k OoD benchmark datasets, we show that our method, combined with post-hoc OoDD approaches leveraging pretrained knowledge (e.g., NegLabel), significantly outperforms existing methods, achieving state-of-the-art OoDD performance and leading ID accuracy.
comment: CVPR 2025
☆ SKDU at De-Factify 4.0: Vision Transformer with Data Augmentation for AI-Generated Image Detection AAAI
The aim of this work is to explore the potential of pre-trained vision-language models, e.g. Vision Transformers (ViT), enhanced with advanced data augmentation strategies for the detection of AI-generated images. Our approach leverages a fine-tuned ViT model trained on the Defactify-4.0 dataset, which includes images generated by state-of-the-art models such as Stable Diffusion 2.1, Stable Diffusion XL, Stable Diffusion 3, DALL-E 3, and MidJourney. We employ perturbation techniques like flipping, rotation, Gaussian noise injection, and JPEG compression during training to improve model robustness and generalisation. The experimental results demonstrate that our ViT-based pipeline achieves state-of-the-art performance, significantly outperforming competing methods on both validation and test datasets.
comment: De-Factify 4.0 workshop at the 39th Annual AAAI Conference on Artificial Intelligence (AAAI 2025)
☆ CRCL: Causal Representation Consistency Learning for Anomaly Detection in Surveillance Videos
Video Anomaly Detection (VAD) remains a fundamental yet formidable task in the video understanding community, with promising applications in areas such as information forensics and public safety protection. Due to the rarity and diversity of anomalies, existing methods only use easily collected regular events to model the inherent normality of normal spatial-temporal patterns in an unsupervised manner. Previous studies have shown that existing unsupervised VAD models are incapable of label-independent data offsets (e.g., scene changes) in real-world scenarios and may fail to respond to light anomalies due to the overgeneralization of deep neural networks. Inspired by causality learning, we argue that there exist causal factors that can adequately generalize the prototypical patterns of regular events and present significant deviations when anomalous instances occur. In this regard, we propose Causal Representation Consistency Learning (CRCL) to implicitly mine potential scene-robust causal variable in unsupervised video normality learning. Specifically, building on the structural causal models, we propose scene-debiasing learning and causality-inspired normality learning to strip away entangled scene bias in deep representations and learn causal video normality, respectively. Extensive experiments on benchmarks validate the superiority of our method over conventional deep representation learning. Moreover, ablation studies and extension validation show that the CRCL can cope with label-independent biases in multi-scene settings and maintain stable performance with only limited training data available.
comment: Accepted for publication by IEEE Transactions on Image Processing
☆ Change3D: Revisiting Change Detection and Captioning from A Video Modeling Perspective CVPR 2025
In this paper, we present Change3D, a framework that reconceptualizes the change detection and captioning tasks through video modeling. Recent methods have achieved remarkable success by regarding each pair of bi-temporal images as separate frames. They employ a shared-weight image encoder to extract spatial features and then use a change extractor to capture differences between the two images. However, image feature encoding, being a task-agnostic process, cannot attend to changed regions effectively. Furthermore, different change extractors designed for various change detection and captioning tasks make it difficult to have a unified framework. To tackle these challenges, Change3D regards the bi-temporal images as comprising two frames akin to a tiny video. By integrating learnable perception frames between the bi-temporal images, a video encoder enables the perception frames to interact with the images directly and perceive their differences. Therefore, we can get rid of the intricate change extractors, providing a unified framework for different change detection and captioning tasks. We verify Change3D on multiple tasks, encompassing change detection (including binary change detection, semantic change detection, and building damage assessment) and change captioning, across eight standard benchmarks. Without bells and whistles, this simple yet effective framework can achieve superior performance with an ultra-light video model comprising only ~6%-13% of the parameters and ~8%-34% of the FLOPs compared to state-of-the-art methods. We hope that Change3D could be an alternative to 2D-based models and facilitate future research.
comment: conference paper, accepted by CVPR 2025
☆ NexusGS: Sparse View Synthesis with Epipolar Depth Priors in 3D Gaussian Splatting CVPR 2025
Neural Radiance Field (NeRF) and 3D Gaussian Splatting (3DGS) have noticeably advanced photo-realistic novel view synthesis using images from densely spaced camera viewpoints. However, these methods struggle in few-shot scenarios due to limited supervision. In this paper, we present NexusGS, a 3DGS-based approach that enhances novel view synthesis from sparse-view images by directly embedding depth information into point clouds, without relying on complex manual regularizations. Exploiting the inherent epipolar geometry of 3DGS, our method introduces a novel point cloud densification strategy that initializes 3DGS with a dense point cloud, reducing randomness in point placement while preventing over-smoothing and overfitting. Specifically, NexusGS comprises three key steps: Epipolar Depth Nexus, Flow-Resilient Depth Blending, and Flow-Filtered Depth Pruning. These steps leverage optical flow and camera poses to compute accurate depth maps, while mitigating the inaccuracies often associated with optical flow. By incorporating epipolar depth priors, NexusGS ensures reliable dense point cloud coverage and supports stable 3DGS training under sparse-view conditions. Experiments demonstrate that NexusGS significantly enhances depth accuracy and rendering quality, surpassing state-of-the-art methods by a considerable margin. Furthermore, we validate the superiority of our generated point clouds by substantially boosting the performance of competing methods. Project page: https://usmizuki.github.io/NexusGS/.
comment: This paper is accepted by CVPR 2025
☆ LGI-DETR: Local-Global Interaction for UAV Object Detection
UAV has been widely used in various fields. However, most of the existing object detectors used in drones are not end-to-end and require the design of various complex components and careful fine-tuning. Most of the existing end-to-end object detectors are designed for natural scenes. It is not ideal to apply them directly to UAV images. In order to solve the above challenges, we design an local-global information interaction DETR for UAVs, namely LGI-DETR. Cross-layer bidirectional low-level and high-level feature information enhancement, this fusion method is effective especially in the field of small objection detection. At the initial stage of encoder, we propose a local spatial enhancement module (LSE), which enhances the low-level rich local spatial information into the high-level feature, and reduces the loss of local information in the transmission process of high-level information. At the final stage of the encoder, we propose a novel global information injection module (GII) designed to integrate rich high-level global semantic representations with low-level feature maps. This hierarchical fusion mechanism effectively addresses the inherent limitations of local receptive fields by propagating contextual information across the feature hierarchy. Experimental results on two challenging UAV image object detection benchmarks, VisDrone2019 and UAVDT, show that our proposed model outperforms the SOTA model. Compared to the baseline model, AP and AP50 improved by 1.9% and 2.4%, respectively.
comment: 14 pages
☆ Leveraging Perturbation Robustness to Enhance Out-of-Distribution Detection
Out-of-distribution (OOD) detection is the task of identifying inputs that deviate from the training data distribution. This capability is essential for safely deploying deep computer vision models in open-world environments. In this work, we propose a post-hoc method, Perturbation-Rectified OOD detection (PRO), based on the insight that prediction confidence for OOD inputs is more susceptible to reduction under perturbation than in-distribution (IND) inputs. Based on the observation, we propose an adversarial score function that searches for the local minimum scores near the original inputs by applying gradient descent. This procedure enhances the separability between IND and OOD samples. Importantly, the approach improves OOD detection performance without complex modifications to the underlying model architectures. We conduct extensive experiments using the OpenOOD benchmark~\cite{yang2022openood}. Our approach further pushes the limit of softmax-based OOD detection and is the leading post-hoc method for small-scale models. On a CIFAR-10 model with adversarial training, PRO effectively detects near-OOD inputs, achieving a reduction of more than 10\% on FPR@95 compared to state-of-the-art methods.
☆ Frequency Dynamic Convolution for Dense Image Prediction CVPR 2025
While Dynamic Convolution (DY-Conv) has shown promising performance by enabling adaptive weight selection through multiple parallel weights combined with an attention mechanism, the frequency response of these weights tends to exhibit high similarity, resulting in high parameter costs but limited adaptability. In this work, we introduce Frequency Dynamic Convolution (FDConv), a novel approach that mitigates these limitations by learning a fixed parameter budget in the Fourier domain. FDConv divides this budget into frequency-based groups with disjoint Fourier indices, enabling the construction of frequency-diverse weights without increasing the parameter cost. To further enhance adaptability, we propose Kernel Spatial Modulation (KSM) and Frequency Band Modulation (FBM). KSM dynamically adjusts the frequency response of each filter at the spatial level, while FBM decomposes weights into distinct frequency bands in the frequency domain and modulates them dynamically based on local content. Extensive experiments on object detection, segmentation, and classification validate the effectiveness of FDConv. We demonstrate that when applied to ResNet-50, FDConv achieves superior performance with a modest increase of +3.6M parameters, outperforming previous methods that require substantial increases in parameter budgets (e.g., CondConv +90M, KW +76.5M). Moreover, FDConv seamlessly integrates into a variety of architectures, including ConvNeXt, Swin-Transformer, offering a flexible and efficient solution for modern vision tasks. The code is made publicly available at https://github.com/Linwei-Chen/FDConv.
comment: Accepted by CVPR 2025
☆ Good Keypoints for the Two-View Geometry Estimation Problem CVPR 2025
Local features are essential to many modern downstream applications. Therefore, it is of interest to determine the properties of local features that contribute to the downstream performance for a better design of feature detectors and descriptors. In our work, we propose a new theoretical model for scoring feature points (keypoints) in the context of the two-view geometry estimation problem. The model determines two properties that a good keypoint for solving the homography estimation problem should have: be repeatable and have a small expected measurement error. This result provides key insights into why maximizing the number of correspondences doesn't always lead to better homography estimation accuracy. We use the developed model to design a method that detects keypoints that benefit the homography estimation introducing the Bounded NeSS-ST (BoNeSS-ST) keypoint detector. The novelty of BoNeSS-ST comes from strong theoretical foundations, a more accurate keypoint scoring due to subpixel refinement and a cost designed for superior robustness to low saliency keypoints. As a result, BoNeSS-ST outperforms prior self-supervised local feature detectors in both planar homography and epipolar geometry estimation problems.
comment: Camera-ready version of the CVPR 2025 paper
☆ EgoSurgery-HTS: A Dataset for Egocentric Hand-Tool Segmentation in Open Surgery Videos
Egocentric open-surgery videos capture rich, fine-grained details essential for accurately modeling surgical procedures and human behavior in the operating room. A detailed, pixel-level understanding of hands and surgical tools is crucial for interpreting a surgeon's actions and intentions. We introduce EgoSurgery-HTS, a new dataset with pixel-wise annotations and a benchmark suite for segmenting surgical tools, hands, and interacting tools in egocentric open-surgery videos. Specifically, we provide a labeled dataset for (1) tool instance segmentation of 14 distinct surgical tools, (2) hand instance segmentation, and (3) hand-tool segmentation to label hands and the tools they manipulate. Using EgoSurgery-HTS, we conduct extensive evaluations of state-of-the-art segmentation methods and demonstrate significant improvements in the accuracy of hand and hand-tool segmentation in egocentric open-surgery videos compared to existing datasets. The dataset will be released at https://github.com/Fujiry0/EgoSurgery.
☆ Self-Supervised Learning based on Transformed Image Reconstruction for Equivariance-Coherent Feature Representation
The equivariant behaviour of features is essential in many computer vision tasks, yet popular self-supervised learning (SSL) methods tend to constrain equivariance by design. We propose a self-supervised learning approach where the system learns transformations independently by reconstructing images that have undergone previously unseen transformations. Specifically, the model is tasked to reconstruct intermediate transformed images, e.g. translated or rotated images, without prior knowledge of these transformations. This auxiliary task encourages the model to develop equivariance-coherent features without relying on predefined transformation rules. To this end, we apply transformations to the input image, generating an image pair, and then split the extracted features into two sets per image. One set is used with a usual SSL loss encouraging invariance, the other with our loss based on the auxiliary task to reconstruct the intermediate transformed images. Our loss and the SSL loss are linearly combined with weighted terms. Evaluating on synthetic tasks with natural images, our proposed method strongly outperforms all competitors, regardless of whether they are designed to learn equivariance. Furthermore, when trained alongside augmentation-based methods as the invariance tasks, such as iBOT or DINOv2, we successfully learn a balanced combination of invariant and equivariant features. Our approach performs strong on a rich set of realistic computer vision downstream tasks, almost always improving over all baselines.
☆ Robust Tube-based Control Strategy for Vision-guided Autonomous Vehicles
A robust control strategy for autonomous vehicles can improve system stability, enhance riding comfort, and prevent driving accidents. This paper presents a novel interpolation tube-based constrained iterative linear quadratic regulator (itube-CILQR) algorithm for autonomous computer-vision-based vehicle lane-keeping. The goal of the algorithm is to enhance robustness during high-speed cornering on tight turns. The advantages of itube-CILQR over the standard tube-approach include reduced system conservatism and increased computational speed. Numerical and vision-based experiments were conducted to examine the feasibility of the proposed algorithm. The proposed itube-CILQR algorithm is better suited to vehicle lane-keeping than variational CILQR-based methods and model predictive control (MPC) approaches using a classical interior-point solver. Specifically, in evaluation experiments, itube-CILQR achieved an average runtime of 3.16 ms to generate a control signal to guide a self-driving vehicle; itube-MPC typically required a 4.67-times longer computation time to complete the same task. Moreover, the influence of conservatism on system behavior was investigated by exploring the interpolation variable trajectories derived from the proposed itube-CILQR algorithm during lane-keeping maneuvers.
comment: 13 pages, 14 figures
☆ Linguistics-aware Masked Image Modeling for Self-supervised Scene Text Recognition CVPR 2025
Text images are unique in their dual nature, encompassing both visual and linguistic information. The visual component encompasses structural and appearance-based features, while the linguistic dimension incorporates contextual and semantic elements. In scenarios with degraded visual quality, linguistic patterns serve as crucial supplements for comprehension, highlighting the necessity of integrating both aspects for robust scene text recognition (STR). Contemporary STR approaches often use language models or semantic reasoning modules to capture linguistic features, typically requiring large-scale annotated datasets. Self-supervised learning, which lacks annotations, presents challenges in disentangling linguistic features related to the global context. Typically, sequence contrastive learning emphasizes the alignment of local features, while masked image modeling (MIM) tends to exploit local structures to reconstruct visual patterns, resulting in limited linguistic knowledge. In this paper, we propose a Linguistics-aware Masked Image Modeling (LMIM) approach, which channels the linguistic information into the decoding process of MIM through a separate branch. Specifically, we design a linguistics alignment module to extract vision-independent features as linguistic guidance using inputs with different visual appearances. As features extend beyond mere visual structures, LMIM must consider the global context to achieve reconstruction. Extensive experiments on various benchmarks quantitatively demonstrate our state-of-the-art performance, and attention visualizations qualitatively show the simultaneous capture of both visual and linguistic information.
comment: CVPR 2025
☆ SFDLA: Source-Free Document Layout Analysis
Document Layout Analysis (DLA) is a fundamental task in document understanding. However, existing DLA and adaptation methods often require access to large-scale source data and target labels. This requirements severely limiting their real-world applicability, particularly in privacy-sensitive and resource-constrained domains, such as financial statements, medical records, and proprietary business documents. According to our observation, directly transferring source-domain fine-tuned models on target domains often results in a significant performance drop (Avg. -32.64%). In this work, we introduce Source-Free Document Layout Analysis (SFDLA), aiming for adapting a pre-trained source DLA models to an unlabeled target domain, without access to any source data. To address this challenge, we establish the first SFDLA benchmark, covering three major DLA datasets for geometric- and content-aware adaptation. Furthermore, we propose Document Layout Analysis Adapter (DLAdapter), a novel framework that is designed to improve source-free adaptation across document domains. Our method achieves a +4.21% improvement over the source-only baseline and a +2.26% gain over existing source-free methods from PubLayNet to DocLayNet. We believe this work will inspire the DLA community to further investigate source-free document understanding. To support future research of the community, the benchmark, models, and code will be publicly available at https://github.com/s3setewe/sfdla-DLAdapter.
comment: The benchmark, models, and code will be publicly available at https://github.com/s3setewe/sfdla-DLAdapter
☆ FG$^2$: Fine-Grained Cross-View Localization by Fine-Grained Feature Matching
We propose a novel fine-grained cross-view localization method that estimates the 3 Degrees of Freedom pose of a ground-level image in an aerial image of the surroundings by matching fine-grained features between the two images. The pose is estimated by aligning a point plane generated from the ground image with a point plane sampled from the aerial image. To generate the ground points, we first map ground image features to a 3D point cloud. Our method then learns to select features along the height dimension to pool the 3D points to a Bird's-Eye-View (BEV) plane. This selection enables us to trace which feature in the ground image contributes to the BEV representation. Next, we sample a set of sparse matches from computed point correspondences between the two point planes and compute their relative pose using Procrustes alignment. Compared to the previous state-of-the-art, our method reduces the mean localization error by 28% on the VIGOR cross-area test set. Qualitative results show that our method learns semantically consistent matches across ground and aerial views through weakly supervised learning from the camera pose.
☆ Boosting Resolution Generalization of Diffusion Transformers with Randomized Positional Encodings
Resolution generalization in image generation tasks enables the production of higher-resolution images with lower training resolution overhead. However, a significant challenge in resolution generalization, particularly in the widely used Diffusion Transformers, lies in the mismatch between the positional encodings encountered during testing and those used during training. While existing methods have employed techniques such as interpolation, extrapolation, or their combinations, none have fully resolved this issue. In this paper, we propose a novel two-dimensional randomized positional encodings (RPE-2D) framework that focuses on learning positional order of image patches instead of the specific distances between them, enabling seamless high- and low-resolution image generation without requiring high- and low-resolution image training. Specifically, RPE-2D independently selects positions over a broader range along both the horizontal and vertical axes, ensuring that all position encodings are trained during the inference phase, thus improving resolution generalization. Additionally, we propose a random data augmentation technique to enhance the modeling of position order. To address the issue of image cropping caused by the augmentation, we introduce corresponding micro-conditioning to enable the model to perceive the specific cropping patterns. On the ImageNet dataset, our proposed RPE-2D achieves state-of-the-art resolution generalization performance, outperforming existing competitive methods when trained at a resolution of $256 \times 256$ and inferred at $384 \times 384$ and $512 \times 512$, as well as when scaling from $512 \times 512$ to $768 \times 768$ and $1024 \times 1024$. And it also exhibits outstanding capabilities in low-resolution image generation, multi-stage training acceleration and multi-resolution inheritance.
☆ GS-Marker: Generalizable and Robust Watermarking for 3D Gaussian Splatting
In the Generative AI era, safeguarding 3D models has become increasingly urgent. While invisible watermarking is well-established for 2D images with encoder-decoder frameworks, generalizable and robust solutions for 3D remain elusive. The main difficulty arises from the renderer between the 3D encoder and 2D decoder, which disrupts direct gradient flow and complicates training. Existing 3D methods typically rely on per-scene iterative optimization, resulting in time inefficiency and limited generalization. In this work, we propose a single-pass watermarking approach for 3D Gaussian Splatting (3DGS), a well-known yet underexplored representation for watermarking. We identify two major challenges: (1) ensuring effective training generalized across diverse 3D models, and (2) reliably extracting watermarks from free-view renderings, even under distortions. Our framework, named GS-Marker, incorporates a 3D encoder to embed messages, distortion layers to enhance resilience against various distortions, and a 2D decoder to extract watermarks from renderings. A key innovation is the Adaptive Marker Control mechanism that adaptively perturbs the initially optimized 3DGS, escaping local minima and improving both training stability and convergence. Extensive experiments show that GS-Marker outperforms per-scene training approaches in terms of decoding accuracy and model fidelity, while also significantly reducing computation time.
☆ LLaVAction: evaluating and training multi-modal large language models for action recognition
Understanding human behavior requires measuring behavioral actions. Due to its complexity, behavior is best mapped onto a rich, semantic structure such as language. The recent development of multi-modal large language models (MLLMs) is a promising candidate for a wide range of action understanding tasks. In this work, we focus on evaluating and then improving MLLMs to perform action recognition. We reformulate EPIC-KITCHENS-100, one of the largest and most challenging egocentric action datasets, to the form of video multiple question answering (EPIC-KITCHENS-100-MQA). We show that when we sample difficult incorrect answers as distractors, leading MLLMs struggle to recognize the correct actions. We propose a series of methods that greatly improve the MLLMs' ability to perform action recognition, achieving state-of-the-art on both the EPIC-KITCHENS-100 validation set, as well as outperforming GPT-4o by 21 points in accuracy on EPIC-KITCHENS-100-MQA. Lastly, we show improvements on other action-related video benchmarks such as EgoSchema, PerceptionTest, LongVideoBench, VideoMME and MVBench, suggesting that MLLMs are a promising path forward for complex action tasks. Code and models are available at: https://github.com/AdaptiveMotorControlLab/LLaVAction.
comment: https://github.com/AdaptiveMotorControlLab/LLaVAction
☆ Accenture-NVS1: A Novel View Synthesis Dataset
This paper introduces ACC-NVS1, a specialized dataset designed for research on Novel View Synthesis specifically for airborne and ground imagery. Data for ACC-NVS1 was collected in Austin, TX and Pittsburgh, PA in 2023 and 2024. The collection encompasses six diverse real-world scenes captured from both airborne and ground cameras, resulting in a total of 148,000 images. ACC-NVS1 addresses challenges such as varying altitudes and transient objects. This dataset is intended to supplement existing datasets, providing additional resources for comprehensive research, rather than serving as a benchmark.
comment: 6 pages, 7 figures
☆ Revisiting Automatic Data Curation for Vision Foundation Models in Digital Pathology MICCAI 2025
Vision foundation models (FMs) are accelerating the development of digital pathology algorithms and transforming biomedical research. These models learn, in a self-supervised manner, to represent histological features in highly heterogeneous tiles extracted from whole-slide images (WSIs) of real-world patient samples. The performance of these FMs is significantly influenced by the size, diversity, and balance of the pre-training data. However, data selection has been primarily guided by expert knowledge at the WSI level, focusing on factors such as disease classification and tissue types, while largely overlooking the granular details available at the tile level. In this paper, we investigate the potential of unsupervised automatic data curation at the tile-level, taking into account 350 million tiles. Specifically, we apply hierarchical clustering trees to pre-extracted tile embeddings, allowing us to sample balanced datasets uniformly across the embedding space of the pretrained FM. We further identify these datasets are subject to a trade-off between size and balance, potentially compromising the quality of representations learned by FMs, and propose tailored batch sampling strategies to mitigate this effect. We demonstrate the effectiveness of our method through improved performance on a diverse range of clinically relevant downstream tasks.
comment: MICCAI 2025
☆ Benchmarking Burst Super-Resolution for Polarization Images: Noise Dataset and Analysis
Snapshot polarization imaging calculates polarization states from linearly polarized subimages. To achieve this, a polarization camera employs a double Bayer-patterned sensor to capture both color and polarization. It demonstrates low light efficiency and low spatial resolution, resulting in increased noise and compromised polarization measurements. Although burst super-resolution effectively reduces noise and enhances spatial resolution, applying it to polarization imaging poses challenges due to the lack of tailored datasets and reliable ground truth noise statistics. To address these issues, we introduce PolarNS and PolarBurstSR, two innovative datasets developed specifically for polarization imaging. PolarNS provides characterization of polarization noise statistics, facilitating thorough analysis, while PolarBurstSR functions as a benchmark for burst super-resolution in polarization images. These datasets, collected under various real-world conditions, enable comprehensive evaluation. Additionally, we present a model for analyzing polarization noise to quantify noise propagation, tested on a large dataset captured in a darkroom environment. As part of our application, we compare the latest burst super-resolution models, highlighting the advantages of training tailored to polarization compared to RGB-based methods. This work establishes a benchmark for polarization burst super-resolution and offers critical insights into noise propagation, thereby enhancing polarization image reconstruction.
☆ Channel Consistency Prior and Self-Reconstruction Strategy Based Unsupervised Image Deraining CVPR2025
Recently, deep image deraining models based on paired datasets have made a series of remarkable progress. However, they cannot be well applied in real-world applications due to the difficulty of obtaining real paired datasets and the poor generalization performance. In this paper, we propose a novel Channel Consistency Prior and Self-Reconstruction Strategy Based Unsupervised Image Deraining framework, CSUD, to tackle the aforementioned challenges. During training with unpaired data, CSUD is capable of generating high-quality pseudo clean and rainy image pairs which are used to enhance the performance of deraining network. Specifically, to preserve more image background details while transferring rain streaks from rainy images to the unpaired clean images, we propose a novel Channel Consistency Loss (CCLoss) by introducing the Channel Consistency Prior (CCP) of rain streaks into training process, thereby ensuring that the generated pseudo rainy images closely resemble the real ones. Furthermore, we propose a novel Self-Reconstruction (SR) strategy to alleviate the redundant information transfer problem of the generator, further improving the deraining performance and the generalization capability of our method. Extensive experiments on multiple synthetic and real-world datasets demonstrate that the deraining performance of CSUD surpasses other state-of-the-art unsupervised methods and CSUD exhibits superior generalization capability.
comment: Accepted to CVPR2025
☆ OCRT: Boosting Foundation Models in the Open World with Object-Concept-Relation Triad CVPR 2025
Although foundation models (FMs) claim to be powerful, their generalization ability significantly decreases when faced with distribution shifts, weak supervision, or malicious attacks in the open world. On the other hand, most domain generalization or adversarial fine-tuning methods are task-related or model-specific, ignoring the universality in practical applications and the transferability between FMs. This paper delves into the problem of generalizing FMs to the out-of-domain data. We propose a novel framework, the Object-Concept-Relation Triad (OCRT), that enables FMs to extract sparse, high-level concepts and intricate relational structures from raw visual inputs. The key idea is to bind objects in visual scenes and a set of object-centric representations through unsupervised decoupling and iterative refinement. To be specific, we project the object-centric representations onto a semantic concept space that the model can readily interpret and estimate their importance to filter out irrelevant elements. Then, a concept-based graph, which has a flexible degree, is constructed to incorporate the set of concepts and their corresponding importance, enabling the extraction of high-order factors from informative concepts and facilitating relational reasoning among these concepts. Extensive experiments demonstrate that OCRT can substantially boost the generalizability and robustness of SAM and CLIP across multiple downstream tasks.
comment: Accepted by CVPR 2025
☆ Hardware-Rasterized Ray-Based Gaussian Splatting
We present a novel, hardware rasterized rendering approach for ray-based 3D Gaussian Splatting (RayGS), obtaining both fast and high-quality results for novel view synthesis. Our work contains a mathematically rigorous and geometrically intuitive derivation about how to efficiently estimate all relevant quantities for rendering RayGS models, structured with respect to standard hardware rasterization shaders. Our solution is the first enabling rendering RayGS models at sufficiently high frame rates to support quality-sensitive applications like Virtual and Mixed Reality. Our second contribution enables alias-free rendering for RayGS, by addressing MIP-related issues arising when rendering diverging scales during training and testing. We demonstrate significant performance gains, across different benchmark scenes, while retaining state-of-the-art appearance quality of RayGS.
☆ NullSwap: Proactive Identity Cloaking Against Deepfake Face Swapping
Suffering from performance bottlenecks in passively detecting high-quality Deepfake images due to the advancement of generative models, proactive perturbations offer a promising approach to disabling Deepfake manipulations by inserting signals into benign images. However, existing proactive perturbation approaches remain unsatisfactory in several aspects: 1) visual degradation due to direct element-wise addition; 2) limited effectiveness against face swapping manipulation; 3) unavoidable reliance on white- and grey-box settings to involve generative models during training. In this study, we analyze the essence of Deepfake face swapping and argue the necessity of protecting source identities rather than target images, and we propose NullSwap, a novel proactive defense approach that cloaks source image identities and nullifies face swapping under a pure black-box scenario. We design an Identity Extraction module to obtain facial identity features from the source image, while a Perturbation Block is then devised to generate identity-guided perturbations accordingly. Meanwhile, a Feature Block extracts shallow-level image features, which are then fused with the perturbation in the Cloaking Block for image reconstruction. Furthermore, to ensure adaptability across different identity extractors in face swapping algorithms, we propose Dynamic Loss Weighting to adaptively balance identity losses. Experiments demonstrate the outstanding ability of our approach to fool various identity recognition models, outperforming state-of-the-art proactive perturbations in preventing face swapping models from generating images with correct source identities.
☆ Human Motion Unlearning
We introduce the task of human motion unlearning to prevent the synthesis of toxic animations while preserving the general text-to-motion generative performance. Unlearning toxic motions is challenging as those can be generated from explicit text prompts and from implicit toxic combinations of safe motions (e.g., ``kicking" is ``loading and swinging a leg"). We propose the first motion unlearning benchmark by filtering toxic motions from the large and recent text-to-motion datasets of HumanML3D and Motion-X. We propose baselines, by adapting state-of-the-art image unlearning techniques to process spatio-temporal signals. Finally, we propose a novel motion unlearning model based on Latent Code Replacement, which we dub LCR. LCR is training-free and suitable to the discrete latent spaces of state-of-the-art text-to-motion diffusion models. LCR is simple and consistently outperforms baselines qualitatively and quantitatively. Project page: \href{https://www.pinlab.org/hmu}{https://www.pinlab.org/hmu}.
☆ Any6D: Model-free 6D Pose Estimation of Novel Objects CVPR 2025
We introduce Any6D, a model-free framework for 6D object pose estimation that requires only a single RGB-D anchor image to estimate both the 6D pose and size of unknown objects in novel scenes. Unlike existing methods that rely on textured 3D models or multiple viewpoints, Any6D leverages a joint object alignment process to enhance 2D-3D alignment and metric scale estimation for improved pose accuracy. Our approach integrates a render-and-compare strategy to generate and refine pose hypotheses, enabling robust performance in scenarios with occlusions, non-overlapping views, diverse lighting conditions, and large cross-environment variations. We evaluate our method on five challenging datasets: REAL275, Toyota-Light, HO3D, YCBINEOAT, and LM-O, demonstrating its effectiveness in significantly outperforming state-of-the-art methods for novel object pose estimation. Project page: https://taeyeop.com/any6d
comment: CVPR 2025, Project Page: https://taeyeop.com/any6d
☆ Feature Calibration enhanced Parameter Synthesis for CLIP-based Class-incremental Learning
Class-incremental Learning (CIL) enables models to continuously learn new class knowledge while memorizing previous classes, facilitating their adaptation and evolution in dynamic environments. Traditional CIL methods are mainly based on visual features, which limits their ability to handle complex scenarios. In contrast, Vision-Language Models (VLMs) show promising potential to promote CIL by integrating pretrained knowledge with textual features. However, previous methods make it difficult to overcome catastrophic forgetting while preserving the generalization capabilities of VLMs. To tackle these challenges, we propose Feature Calibration enhanced Parameter Synthesis (FCPS) in this paper. Specifically, our FCPS employs a specific parameter adjustment mechanism to iteratively refine the proportion of original visual features participating in the final class determination, ensuring the model's foundational generalization capabilities. Meanwhile, parameter integration across different tasks achieves a balance between learning new class knowledge and retaining old knowledge. Experimental results on popular benchmarks (e.g., CIFAR100 and ImageNet100) validate the superiority of the proposed method.
☆ Structure-Aware Correspondence Learning for Relative Pose Estimation CVPR2025
Relative pose estimation provides a promising way for achieving object-agnostic pose estimation. Despite the success of existing 3D correspondence-based methods, the reliance on explicit feature matching suffers from small overlaps in visible regions and unreliable feature estimation for invisible regions. Inspired by humans' ability to assemble two object parts that have small or no overlapping regions by considering object structure, we propose a novel Structure-Aware Correspondence Learning method for Relative Pose Estimation, which consists of two key modules. First, a structure-aware keypoint extraction module is designed to locate a set of kepoints that can represent the structure of objects with different shapes and appearance, under the guidance of a keypoint based image reconstruction loss. Second, a structure-aware correspondence estimation module is designed to model the intra-image and inter-image relationships between keypoints to extract structure-aware features for correspondence estimation. By jointly leveraging these two modules, the proposed method can naturally estimate 3D-3D correspondences for unseen objects without explicit feature matching for precise relative pose estimation. Experimental results on the CO3D, Objaverse and LineMOD datasets demonstrate that the proposed method significantly outperforms prior methods, i.e., with 5.7{\deg}reduction in mean angular error on the CO3D dataset.
comment: CVPR2025
☆ Boosting Virtual Agent Learning and Reasoning: A Step-wise, Multi-dimensional, and Generalist Reward Model with Benchmark
The development of Generalist Virtual Agents (GVAs) powered by Multimodal Large Language Models (MLLMs) has shown significant promise in autonomous task execution. However, current training paradigms face critical limitations, including reliance on outcome supervision and labor-intensive human annotations. To address these challenges, we propose Similar, a Step-wise Multi-dimensional Generalist Reward Model, which offers fine-grained signals for agent training and can choose better action for inference-time scaling. Specifically, we begin by systematically defining five dimensions for evaluating agent actions. Building on this framework, we design an MCTS-P algorithm to automatically collect and annotate step-wise, five-dimensional agent execution data. Using this data, we train Similar with the Triple-M strategy. Furthermore, we introduce the first benchmark in the virtual agent domain for step-wise, multi-dimensional reward model training and evaluation, named SRM. This benchmark consists of two components: SRMTrain, which serves as the training set for Similar, and SRMEval, a manually selected test set for evaluating the reward model. Experimental results demonstrate that Similar, through its step-wise, multi-dimensional assessment and synergistic gain, provides GVAs with effective intermediate signals during both training and inference-time scaling. The code is available at https://github.com/Galery23/Similar-v1.
☆ Leveraging Land Cover Priors for Isoprene Emission Super-Resolution
Remote sensing plays a crucial role in monitoring Earth's ecosystems, yet satellite-derived data often suffer from limited spatial resolution, restricting their applicability in atmospheric modeling and climate research. In this work, we propose a deep learning-based Super-Resolution (SR) framework that leverages land cover information to enhance the spatial accuracy of Biogenic Volatile Organic Compounds (BVOCs) emissions, with a particular focus on isoprene. Our approach integrates land cover priors as emission drivers, capturing spatial patterns more effectively than traditional methods. We evaluate the model's performance across various climate conditions and analyze statistical correlations between isoprene emissions and key environmental information such as cropland and tree cover data. Additionally, we assess the generalization capabilities of our SR model by applying it to unseen climate zones and geographical regions. Experimental results demonstrate that incorporating land cover data significantly improves emission SR accuracy, particularly in heterogeneous landscapes. This study contributes to atmospheric chemistry and climate modeling by providing a cost-effective, data-driven approach to refining BVOC emission maps. The proposed method enhances the usability of satellite-based emissions data, supporting applications in air quality forecasting, climate impact assessments, and environmental studies.
comment: 17 pages, 16 figures, 4 tables
☆ Robust face recognition based on the wing loss and the $\ell_1$ regularization
In recent years, sparse sampling techniques based on regression analysis have witnessed extensive applications in face recognition research. Presently, numerous sparse sampling models based on regression analysis have been explored by various researchers. Nevertheless, the recognition rates of the majority of these models would be significantly decreased when confronted with highly occluded and highly damaged face images. In this paper, a new wing-constrained sparse coding model(WCSC) and its weighted version(WWCSC) are introduced, so as to deal with the face recognition problem in complex circumstances, where the alternating direction method of multipliers (ADMM) algorithm is employed to solve the corresponding minimization problems. In addition, performances of the proposed method are examined based on the four well-known facial databases, namely the ORL facial database, the Yale facial database, the AR facial database and the FERET facial database. Also, compared to the other methods in the literatures, the WWCSC has a very high recognition rate even in complex situations where face images have high occlusion or high damage, which illustrates the robustness of the WWCSC method in facial recognition.
comment: 10 pages, 3 figures
☆ Rethinking Glaucoma Calibration: Voting-Based Binocular and Metadata Integration
Glaucoma is an incurable ophthalmic disease that damages the optic nerve, leads to vision loss, and ranks among the leading causes of blindness worldwide. Diagnosing glaucoma typically involves fundus photography, optical coherence tomography (OCT), and visual field testing. However, the high cost of OCT often leads to reliance on fundus photography and visual field testing, both of which exhibit inherent inter-observer variability. This stems from glaucoma being a multifaceted disease that influenced by various factors. As a result, glaucoma diagnosis is highly subjective, emphasizing the necessity of calibration, which aligns predicted probabilities with actual disease likelihood. Proper calibration is essential to prevent overdiagnosis or misdiagnosis, which are critical concerns for high-risk diseases. Although AI has significantly improved diagnostic accuracy, overconfidence in models have worsen calibration performance. Recent study has begun focusing on calibration for glaucoma. Nevertheless, previous study has not fully considered glaucoma's systemic nature and the high subjectivity in its diagnostic process. To overcome these limitations, we propose V-ViT (Voting-based ViT), a novel framework that enhances calibration by incorporating disease-specific characteristics. V-ViT integrates binocular data and metadata, reflecting the multi-faceted nature of glaucoma diagnosis. Additionally, we introduce a MC dropout-based Voting System to address high subjectivity. Our approach achieves state-of-the-art performance across all metrics, including accuracy, demonstrating that our proposed methods are effective in addressing calibration issues. We validate our method using a custom dataset including binocular data.
☆ LLGS: Unsupervised Gaussian Splatting for Image Enhancement and Reconstruction in Pure Dark Environment
3D Gaussian Splatting has shown remarkable capabilities in novel view rendering tasks and exhibits significant potential for multi-view optimization.However, the original 3D Gaussian Splatting lacks color representation for inputs in low-light environments. Simply using enhanced images as inputs would lead to issues with multi-view consistency, and current single-view enhancement systems rely on pre-trained data, lacking scene generalization. These problems limit the application of 3D Gaussian Splatting in low-light conditions in the field of robotics, including high-fidelity modeling and feature matching. To address these challenges, we propose an unsupervised multi-view stereoscopic system based on Gaussian Splatting, called Low-Light Gaussian Splatting (LLGS). This system aims to enhance images in low-light environments while reconstructing the scene. Our method introduces a decomposable Gaussian representation called M-Color, which separately characterizes color information for targeted enhancement. Furthermore, we propose an unsupervised optimization method with zero-knowledge priors, using direction-based enhancement to ensure multi-view consistency. Experiments conducted on real-world datasets demonstrate that our system outperforms state-of-the-art methods in both low-light enhancement and 3D Gaussian Splatting.
☆ Unbiasing through Textual Descriptions: Mitigating Representation Bias in Video Benchmarks CVPR 2025
We propose a new "Unbiased through Textual Description (UTD)" video benchmark based on unbiased subsets of existing video classification and retrieval datasets to enable a more robust assessment of video understanding capabilities. Namely, we tackle the problem that current video benchmarks may suffer from different representation biases, e.g., object bias or single-frame bias, where mere recognition of objects or utilization of only a single frame is sufficient for correct prediction. We leverage VLMs and LLMs to analyze and debias benchmarks from such representation biases. Specifically, we generate frame-wise textual descriptions of videos, filter them for specific information (e.g. only objects) and leverage them to examine representation biases across three dimensions: 1) concept bias - determining if a specific concept (e.g., objects) alone suffice for prediction; 2) temporal bias - assessing if temporal information contributes to prediction; and 3) common sense vs. dataset bias - evaluating whether zero-shot reasoning or dataset correlations contribute to prediction. We conduct a systematic analysis of 12 popular video classification and retrieval datasets and create new object-debiased test splits for these datasets. Moreover, we benchmark 30 state-of-the-art video models on original and debiased splits and analyze biases in the models. To facilitate the future development of more robust video understanding benchmarks and models, we release: "UTD-descriptions", a dataset with our rich structured descriptions for each dataset, and "UTD-splits", a dataset of object-debiased test splits.
comment: To be published at CVPR 2025, project webpage https://utd-project.github.io/
☆ OCCO: LVM-guided Infrared and Visible Image Fusion Framework based on Object-aware and Contextual COntrastive Learning
Image fusion is a crucial technique in the field of computer vision, and its goal is to generate high-quality fused images and improve the performance of downstream tasks. However, existing fusion methods struggle to balance these two factors. Achieving high quality in fused images may result in lower performance in downstream visual tasks, and vice versa. To address this drawback, a novel LVM (large vision model)-guided fusion framework with Object-aware and Contextual COntrastive learning is proposed, termed as OCCO. The pre-trained LVM is utilized to provide semantic guidance, allowing the network to focus solely on fusion tasks while emphasizing learning salient semantic features in form of contrastive learning. Additionally, a novel feature interaction fusion network is also designed to resolve information conflicts in fusion images caused by modality differences. By learning the distinction between positive samples and negative samples in the latent feature space (contextual space), the integrity of target information in fused image is improved, thereby benefiting downstream performance. Finally, compared with eight state-of-the-art methods on four datasets, the effectiveness of the proposed method is validated, and exceptional performance is also demonstrated on downstream visual task.
☆ Robust Lane Detection with Wavelet-Enhanced Context Modeling and Adaptive Sampling
Lane detection is critical for autonomous driving and ad-vanced driver assistance systems (ADAS). While recent methods like CLRNet achieve strong performance, they struggle under adverse con-ditions such as extreme weather, illumination changes, occlusions, and complex curves. We propose a Wavelet-Enhanced Feature Pyramid Net-work (WE-FPN) to address these challenges. A wavelet-based non-local block is integrated before the feature pyramid to improve global context modeling, especially for occluded and curved lanes. Additionally, we de-sign an adaptive preprocessing module to enhance lane visibility under poor lighting. An attention-guided sampling strategy further reffnes spa-tial features, boosting accuracy on distant and curved lanes. Experiments on CULane and TuSimple demonstrate that our approach signiffcantly outperforms baselines in challenging scenarios, achieving better robust-ness and accuracy in real-world driving conditions.
☆ Towards Human-Understandable Multi-Dimensional Concept Discovery
Concept-based eXplainable AI (C-XAI) aims to overcome the limitations of traditional saliency maps by converting pixels into human-understandable concepts that are consistent across an entire dataset. A crucial aspect of C-XAI is completeness, which measures how well a set of concepts explains a model's decisions. Among C-XAI methods, Multi-Dimensional Concept Discovery (MCD) effectively improves completeness by breaking down the CNN latent space into distinct and interpretable concept subspaces. However, MCD's explanations can be difficult for humans to understand, raising concerns about their practical utility. To address this, we propose Human-Understandable Multi-dimensional Concept Discovery (HU-MCD). HU-MCD uses the Segment Anything Model for concept identification and implements a CNN-specific input masking technique to reduce noise introduced by traditional masking methods. These changes to MCD, paired with the completeness relation, enable HU-MCD to enhance concept understandability while maintaining explanation faithfulness. Our experiments, including human subject studies, show that HU-MCD provides more precise and reliable explanations than existing C-XAI methods. The code is available at https://github.com/grobruegge/hu-mcd.
☆ Dig2DIG: Dig into Diffusion Information Gains for Image Fusion
Image fusion integrates complementary information from multi-source images to generate more informative results. Recently, the diffusion model, which demonstrates unprecedented generative potential, has been explored in image fusion. However, these approaches typically incorporate predefined multimodal guidance into diffusion, failing to capture the dynamically changing significance of each modality, while lacking theoretical guarantees. To address this issue, we reveal a significant spatio-temporal imbalance in image denoising; specifically, the diffusion model produces dynamic information gains in different image regions with denoising steps. Based on this observation, we Dig into the Diffusion Information Gains (Dig2DIG) and theoretically derive a diffusion-based dynamic image fusion framework that provably reduces the upper bound of the generalization error. Accordingly, we introduce diffusion information gains (DIG) to quantify the information contribution of each modality at different denoising steps, thereby providing dynamic guidance during the fusion process. Extensive experiments on multiple fusion scenarios confirm that our method outperforms existing diffusion-based approaches in terms of both fusion quality and inference efficiency.
☆ Generative Dataset Distillation using Min-Max Diffusion Model ECCV2024
In this paper, we address the problem of generative dataset distillation that utilizes generative models to synthesize images. The generator may produce any number of images under a preserved evaluation time. In this work, we leverage the popular diffusion model as the generator to compute a surrogate dataset, boosted by a min-max loss to control the dataset's diversity and representativeness during training. However, the diffusion model is time-consuming when generating images, as it requires an iterative generation process. We observe a critical trade-off between the number of image samples and the image quality controlled by the diffusion steps and propose Diffusion Step Reduction to achieve optimal performance. This paper details our comprehensive method and its performance. Our model achieved $2^{nd}$ place in the generative track of \href{https://www.dd-challenge.com/#/}{The First Dataset Distillation Challenge of ECCV2024}, demonstrating its superior performance.
comment: The paper is accepted as the ECCV2024 workshop paper and achieved second place in the generative track of The First Dataset Distillation Challenge of ECCV2024, https://www.dd-challenge.com/#/
☆ Training-Free Personalization via Retrieval and Reasoning on Fingerprints
Vision Language Models (VLMs) have lead to major improvements in multimodal reasoning, yet they still struggle to understand user-specific concepts. Existing personalization methods address this limitation but heavily rely on training procedures, that can be either costly or unpleasant to individual users. We depart from existing work, and for the first time explore the training-free setting in the context of personalization. We propose a novel method, Retrieval and Reasoning for Personalization (R2P), leveraging internal knowledge of VLMs. First, we leverage VLMs to extract the concept fingerprint, i.e., key attributes uniquely defining the concept within its semantic class. When a query arrives, the most similar fingerprints are retrieved and scored via chain-of-thought-reasoning. To reduce the risk of hallucinations, the scores are validated through cross-modal verification at the attribute level: in case of a discrepancy between the scores, R2P refines the concept association via pairwise multimodal matching, where the retrieved fingerprints and their images are directly compared with the query. We validate R2P on two publicly available benchmarks and a newly introduced dataset, Personal Concepts with Visual Ambiguity (PerVA), for concept identification highlighting challenges in visual ambiguity. R2P consistently outperforms state-of-the-art approaches on various downstream tasks across all benchmarks. Code will be available upon acceptance.
☆ Unified Uncertainty-Aware Diffusion for Multi-Agent Trajectory Modeling CVPR 2025
Multi-agent trajectory modeling has primarily focused on forecasting future states, often overlooking broader tasks like trajectory completion, which are crucial for real-world applications such as correcting tracking data. Existing methods also generally predict agents' states without offering any state-wise measure of uncertainty. Moreover, popular multi-modal sampling methods lack any error probability estimates for each generated scene under the same prior observations, making it difficult to rank the predictions during inference time. We introduce U2Diff, a \textbf{unified} diffusion model designed to handle trajectory completion while providing state-wise \textbf{uncertainty} estimates jointly. This uncertainty estimation is achieved by augmenting the simple denoising loss with the negative log-likelihood of the predicted noise and propagating latent space uncertainty to the real state space. Additionally, we incorporate a Rank Neural Network in post-processing to enable \textbf{error probability} estimation for each generated mode, demonstrating a strong correlation with the error relative to ground truth. Our method outperforms the state-of-the-art solutions in trajectory completion and forecasting across four challenging sports datasets (NBA, Basketball-U, Football-U, Soccer-U), highlighting the effectiveness of uncertainty and error probability estimation. Video at https://youtu.be/ngw4D4eJToE
comment: Accepted to CVPR 2025 conference
☆ Adapting Video Diffusion Models for Time-Lapse Microscopy
We present a domain adaptation of video diffusion models to generate highly realistic time-lapse microscopy videos of cell division in HeLa cells. Although state-of-the-art generative video models have advanced significantly for natural videos, they remain underexplored in microscopy domains. To address this gap, we fine-tune a pretrained video diffusion model on microscopy-specific sequences, exploring three conditioning strategies: (1) text prompts derived from numeric phenotypic measurements (e.g., proliferation rates, migration speeds, cell-death frequencies), (2) direct numeric embeddings of phenotype scores, and (3) image-conditioned generation, where an initial microscopy frame is extended into a complete video sequence. Evaluation using biologically meaningful morphological, proliferation, and migration metrics demonstrates that fine-tuning substantially improves realism and accurately captures critical cellular behaviors such as mitosis and migration. Notably, the fine-tuned model also generalizes beyond the training horizon, generating coherent cell dynamics even in extended sequences. However, precisely controlling specific phenotypic characteristics remains challenging, highlighting opportunities for future work to enhance conditioning methods. Our results demonstrate the potential for domain-specific fine-tuning of generative video models to produce biologically plausible synthetic microscopy data, supporting applications such as in-silico hypothesis testing and data augmentation.
☆ Galaxy Walker: Geometry-aware VLMs For Galaxy-scale Understanding
Modern vision-language models (VLMs) develop patch embedding and convolution backbone within vector space, especially Euclidean ones, at the very founding. When expanding VLMs to a galaxy scale for understanding astronomical phenomena, the integration of spherical space for planetary orbits and hyperbolic spaces for black holes raises two formidable challenges. a) The current pre-training model is confined to Euclidean space rather than a comprehensive geometric embedding. b) The predominant architecture lacks suitable backbones for anisotropic physical geometries. In this paper, we introduced Galaxy-Walker, a geometry-aware VLM, for the universe-level vision understanding tasks. We proposed the geometry prompt that generates geometry tokens by random walks across diverse spaces on a multi-scale physical graph, along with a geometry adapter that compresses and reshapes the space anisotropy in a mixture-of-experts manner. Extensive experiments demonstrate the effectiveness of our approach, with Galaxy-Walker achieving state-of-the-art performance in both galaxy property estimation ($R^2$ scores up to $0.91$) and morphology classification tasks (up to $+0.17$ F1 improvement in challenging features), significantly outperforming both domain-specific models and general-purpose VLMs.
☆ Advancing Cross-Organ Domain Generalization with Test-Time Style Transfer and Diversity Enhancement
Deep learning has made significant progress in addressing challenges in various fields including computational pathology (CPath). However, due to the complexity of the domain shift problem, the performance of existing models will degrade, especially when it comes to multi-domain or cross-domain tasks. In this paper, we propose a Test-time style transfer (T3s) that uses a bidirectional mapping mechanism to project the features of the source and target domains into a unified feature space, enhancing the generalization ability of the model. To further increase the style expression space, we introduce a Cross-domain style diversification module (CSDM) to ensure the orthogonality between style bases. In addition, data augmentation and low-rank adaptation techniques are used to improve feature alignment and sensitivity, enabling the model to adapt to multi-domain inputs effectively. Our method has demonstrated effectiveness on three unseen datasets.
comment: 2025 IEEE International Symposium on Biomedical Imaging (ISBI)
☆ AMD-Hummingbird: Towards an Efficient Text-to-Video Model
Text-to-Video (T2V) generation has attracted significant attention for its ability to synthesize realistic videos from textual descriptions. However, existing models struggle to balance computational efficiency and high visual quality, particularly on resource-limited devices, e.g.,iGPUs and mobile phones. Most prior work prioritizes visual fidelity while overlooking the need for smaller, more efficient models suitable for real-world deployment. To address this challenge, we propose a lightweight T2V framework, termed Hummingbird, which prunes existing models and enhances visual quality through visual feedback learning. Our approach reduces the size of the U-Net from 1.4 billion to 0.7 billion parameters, significantly improving efficiency while preserving high-quality video generation. Additionally, we introduce a novel data processing pipeline that leverages Large Language Models (LLMs) and Video Quality Assessment (VQA) models to enhance the quality of both text prompts and video data. To support user-driven training and style customization, we publicly release the full training code, including data processing and model training. Extensive experiments show that our method achieves a 31X speedup compared to state-of-the-art models such as VideoCrafter2, while also attaining the highest overall score on VBench. Moreover, our method supports the generation of videos with up to 26 frames, addressing the limitations of existing U-Net-based methods in long video generation. Notably, the entire training process requires only four GPUs, yet delivers performance competitive with existing leading methods. Hummingbird presents a practical and efficient solution for T2V generation, combining high performance, scalability, and flexibility for real-world applications.
comment: Homepage: https://www.amd.com/en/developer/resources/technical-articles/amd-hummingbird-0-9b-text-to-video-diffusion-model-with-4-step-inferencing.html| GitHub: https://github.com/AMD-AIG-AIMA/AMD-Hummingbird-T2V
☆ LeanStereo: A Leaner Backbone based Stereo Network
Recently, end-to-end deep networks based stereo matching methods, mainly because of their performance, have gained popularity. However, this improvement in performance comes at the cost of increased computational and memory bandwidth requirements, thus necessitating specialized hardware (GPUs); even then, these methods have large inference times compared to classical methods. This limits their applicability in real-world applications. Although we desire high accuracy stereo methods albeit with reasonable inference time. To this end, we propose a fast end-to-end stereo matching method. Majority of this speedup comes from integrating a leaner backbone. To recover the performance lost because of a leaner backbone, we propose to use learned attention weights based cost volume combined with LogL1 loss for stereo matching. Using LogL1 loss not only improves the overall performance of the proposed network but also leads to faster convergence. We do a detailed empirical evaluation of different design choices and show that our method requires 4x less operations and is also about 9 to 14x faster compared to the state of the art methods like ACVNet [1], LEAStereo [2] and CFNet [3] while giving comparable performance.
comment: 8 pages, 4 figures
☆ Instruction-Aligned Visual Attention for Mitigating Hallucinations in Large Vision-Language Models ICME2025
Despite the significant success of Large Vision-Language models(LVLMs), these models still suffer hallucinations when describing images, generating answers that include non-existent objects. It is reported that these models tend to over-focus on certain irrelevant image tokens that do not contain critical information for answering the question and distort the output. To address this, we propose an Instruction-Aligned Visual Attention(IAVA) approach, which identifies irrelevant tokens by comparing changes in attention weights under two different instructions. By applying contrastive decoding, we dynamically adjust the logits generated from original image tokens and irrelevant image tokens, reducing the model's over-attention to irrelevant information. The experimental results demonstrate that IAVA consistently outperforms existing decoding techniques on benchmarks such as MME, POPE, and TextVQA in mitigating object hallucinations. Our IAVA approach is available online at https://github.com/Lee-lab558/IAVA.
comment: Accepted by ICME2025
☆ ATARS: An Aerial Traffic Atomic Activity Recognition and Temporal Segmentation Dataset
Traffic Atomic Activity which describes traffic patterns for topological intersection dynamics is a crucial topic for the advancement of intelligent driving systems. However, existing atomic activity datasets are collected from an egocentric view, which cannot support the scenarios where traffic activities in an entire intersection are required. Moreover, existing datasets only provide video-level atomic activity annotations, which require exhausting efforts to manually trim the videos for recognition and limit their applications to untrimmed videos. To bridge this gap, we introduce the Aerial Traffic Atomic Activity Recognition and Segmentation (ATARS) dataset, the first aerial dataset designed for multi-label atomic activity analysis. We offer atomic activity labels for each frame, which accurately record the intervals for traffic activities. Moreover, we propose a novel task, Multi-label Temporal Atomic Activity Recognition, enabling the study of accurate temporal localization for atomic activity and easing the burden of manual video trimming for recognition. We conduct extensive experiments to evaluate existing state-of-the-art models on both atomic activity recognition and temporal atomic activity segmentation. The results highlight the unique challenges of our ATARS dataset, such as recognizing extremely small objects' activities. We further provide comprehensive discussion analyzing these challenges and offer valuable insights for future direction to improve recognizing atomic activity in aerial view. Our source code and dataset are available at https://github.com/magecliff96/ATARS/
☆ EvAnimate: Event-conditioned Image-to-Video Generation for Human Animation
Conditional human animation transforms a static reference image into a dynamic sequence by applying motion cues such as poses. These motion cues are typically derived from video data but are susceptible to limitations including low temporal resolution, motion blur, overexposure, and inaccuracies under low-light conditions. In contrast, event cameras provide data streams with exceptionally high temporal resolution, a wide dynamic range, and inherent resistance to motion blur and exposure issues. In this work, we propose EvAnimate, a framework that leverages event streams as motion cues to animate static human images. Our approach employs a specialized event representation that transforms asynchronous event streams into 3-channel slices with controllable slicing rates and appropriate slice density, ensuring compatibility with diffusion models. Subsequently, a dual-branch architecture generates high-quality videos by harnessing the inherent motion dynamics of the event streams, thereby enhancing both video quality and temporal consistency. Specialized data augmentation strategies further enhance cross-person generalization. Finally, we establish a new benchmarking, including simulated event data for training and validation, and a real-world event dataset capturing human actions under normal and extreme scenarios. The experiment results demonstrate that EvAnimate achieves high temporal fidelity and robust performance in scenarios where traditional video-derived cues fall short.
☆ Benchmarking Post-Hoc Unknown-Category Detection in Food Recognition
Food recognition models often struggle to distinguish between seen and unseen samples, frequently misclassifying samples from unseen categories by assigning them an in-distribution (ID) label. This misclassification presents significant challenges when deploying these models in real-world applications, particularly within automatic dietary assessment systems, where incorrect labels can lead to cascading errors throughout the system. Ideally, such models should prompt the user when an unknown sample is encountered, allowing for corrective action. Given no prior research exploring food recognition in real-world settings, in this work we conduct an empirical analysis of various post-hoc out-of-distribution (OOD) detection methods for fine-grained food recognition. Our findings indicate that virtual logit matching (ViM) performed the best overall, likely due to its combination of logits and feature-space representations. Additionally, our work reinforces prior notions in the OOD domain, noting that models with higher ID accuracy performed better across the evaluated OOD detection methods. Furthermore, transformer-based architectures consistently outperformed convolution-based models in detecting OOD samples across various methods.
☆ Distilling Stereo Networks for Performant and Efficient Leaner Networks IJCNN
Knowledge distillation has been quite popular in vision for tasks like classification and segmentation however not much work has been done for distilling state-of-the-art stereo matching methods despite their range of applications. One of the reasons for its lack of use in stereo matching networks is due to the inherent complexity of these networks, where a typical network is composed of multiple two- and three-dimensional modules. In this work, we systematically combine the insights from state-of-the-art stereo methods with general knowledge-distillation techniques to develop a joint framework for stereo networks distillation with competitive results and faster inference. Moreover, we show, via a detailed empirical analysis, that distilling knowledge from the stereo network requires careful design of the complete distillation pipeline starting from backbone to the right selection of distillation points and corresponding loss functions. This results in the student networks that are not only leaner and faster but give excellent performance . For instance, our student network while performing better than the performance oriented methods like PSMNet [1], CFNet [2], and LEAStereo [3]) on benchmark SceneFlow dataset, is 8x, 5x, and 8x faster respectively. Furthermore, compared to speed oriented methods having inference time less than 100ms, our student networks perform better than all the tested methods. In addition, our student network also shows better generalization capabilities when tested on unseen datasets like ETH3D and Middlebury.
comment: 8 pages, 3 figures. Published in: 2023 International Joint Conference on Neural Networks (IJCNN)
☆ UniPCGC: Towards Practical Point Cloud Geometry Compression via an Efficient Unified Approach AAAI 2025
Learning-based point cloud compression methods have made significant progress in terms of performance. However, these methods still encounter challenges including high complexity, limited compression modes, and a lack of support for variable rate, which restrict the practical application of these methods. In order to promote the development of practical point cloud compression, we propose an efficient unified point cloud geometry compression framework, dubbed as UniPCGC. It is a lightweight framework that supports lossy compression, lossless compression, variable rate and variable complexity. First, we introduce the Uneven 8-Stage Lossless Coder (UELC) in the lossless mode, which allocates more computational complexity to groups with higher coding difficulty, and merges groups with lower coding difficulty. Second, Variable Rate and Complexity Module (VRCM) is achieved in the lossy mode through joint adoption of a rate modulation module and dynamic sparse convolution. Finally, through the dynamic combination of UELC and VRCM, we achieve lossy compression, lossless compression, variable rate and complexity within a unified framework. Compared to the previous state-of-the-art method, our method achieves a compression ratio (CR) gain of 8.1\% on lossless compression, and a Bjontegaard Delta Rate (BD-Rate) gain of 14.02\% on lossy compression, while also supporting variable rate and variable complexity.
comment: Accepted to AAAI 2025
☆ HiRes-FusedMIM: A High-Resolution RGB-DSM Pre-trained Model for Building-Level Remote Sensing Applications
Recent advances in self-supervised learning have led to the development of foundation models that have significantly advanced performance in various computer vision tasks. However, despite their potential, these models often overlook the crucial role of high-resolution digital surface models (DSMs) in understanding urban environments, particularly for building-level analysis, which is essential for applications like digital twins. To address this gap, we introduce HiRes-FusedMIM, a novel pre-trained model specifically designed to leverage the rich information contained within high-resolution RGB and DSM data. HiRes-FusedMIM utilizes a dual-encoder simple masked image modeling (SimMIM) architecture with a multi-objective loss function that combines reconstruction and contrastive objectives, enabling it to learn powerful, joint representations from both modalities. We conducted a comprehensive evaluation of HiRes-FusedMIM on a diverse set of downstream tasks, including classification, semantic segmentation, and instance segmentation. Our results demonstrate that: 1) HiRes-FusedMIM outperforms previous state-of-the-art geospatial methods on several building-related datasets, including WHU Aerial and LoveDA, demonstrating its effectiveness in capturing and leveraging fine-grained building information; 2) Incorporating DSMs during pre-training consistently improves performance compared to using RGB data alone, highlighting the value of elevation information for building-level analysis; 3) The dual-encoder architecture of HiRes-FusedMIM, with separate encoders for RGB and DSM data, significantly outperforms a single-encoder model on the Vaihingen segmentation task, indicating the benefits of learning specialized representations for each modality. To facilitate further research and applications in this direction, we will publicly release the trained model weights.
☆ DiN: Diffusion Model for Robust Medical VQA with Semantic Noisy Labels
Medical Visual Question Answering (Med-VQA) systems benefit the interpretation of medical images containing critical clinical information. However, the challenge of noisy labels and limited high-quality datasets remains underexplored. To address this, we establish the first benchmark for noisy labels in Med-VQA by simulating human mislabeling with semantically designed noise types. More importantly, we introduce the DiN framework, which leverages a diffusion model to handle noisy labels in Med-VQA. Unlike the dominant classification-based VQA approaches that directly predict answers, our Answer Diffuser (AD) module employs a coarse-to-fine process, refining answer candidates with a diffusion model for improved accuracy. The Answer Condition Generator (ACG) further enhances this process by generating task-specific conditional information via integrating answer embeddings with fused image-question features. To address label noise, our Noisy Label Refinement(NLR) module introduces a robust loss function and dynamic answer adjustment to further boost the performance of the AD module.
☆ k-NN as a Simple and Effective Estimator of Transferability
How well can one expect transfer learning to work in a new setting where the domain is shifted, the task is different, and the architecture changes? Many transfer learning metrics have been proposed to answer this question. But how accurate are their predictions in a realistic new setting? We conducted an extensive evaluation involving over 42,000 experiments comparing 23 transferability metrics across 16 different datasets to assess their ability to predict transfer performance. Our findings reveal that none of the existing metrics perform well across the board. However, we find that a simple k-nearest neighbor evaluation -- as is commonly used to evaluate feature quality for self-supervision -- not only surpasses existing metrics, but also offers better computational efficiency and ease of implementation.
☆ AIM2PC: Aerial Image to 3D Building Point Cloud Reconstruction SP
Three-dimensional urban reconstruction of buildings from single-view images has attracted significant attention over the past two decades. However, recent methods primarily focus on rooftops from aerial images, often overlooking essential geometrical details. Additionally, there is a notable lack of datasets containing complete 3D point clouds for entire buildings, along with challenges in obtaining reliable camera pose information for aerial images. This paper addresses these challenges by presenting a novel methodology, AIM2PC , which utilizes our generated dataset that includes complete 3D point clouds and determined camera poses. Our approach takes features from a single aerial image as input and concatenates them with essential additional conditions, such as binary masks and Sobel edge maps, to enable more edge-aware reconstruction. By incorporating a point cloud diffusion model based on Centered denoising Diffusion Probabilistic Models (CDPM), we project these concatenated features onto the partially denoised point cloud using our camera poses at each diffusion step. The proposed method is able to reconstruct the complete 3D building point cloud, including wall information and demonstrates superior performance compared to existing baseline techniques. To allow further comparisons with our methodology the dataset has been made available at https://github.com/Soulaimene/AIM2PCDataset
comment: Accepted to ISPRS Geospatial Week 2025
☆ LookCloser: Frequency-aware Radiance Field for Tiny-Detail Scene CVPR 2025
Humans perceive and comprehend their surroundings through information spanning multiple frequencies. In immersive scenes, people naturally scan their environment to grasp its overall structure while examining fine details of objects that capture their attention. However, current NeRF frameworks primarily focus on modeling either high-frequency local views or the broad structure of scenes with low-frequency information, which is limited to balancing both. We introduce FA-NeRF, a novel frequency-aware framework for view synthesis that simultaneously captures the overall scene structure and high-definition details within a single NeRF model. To achieve this, we propose a 3D frequency quantification method that analyzes the scene's frequency distribution, enabling frequency-aware rendering. Our framework incorporates a frequency grid for fast convergence and querying, a frequency-aware feature re-weighting strategy to balance features across different frequency contents. Extensive experiments show that our method significantly outperforms existing approaches in modeling entire scenes while preserving fine details.
comment: 8 pages, 6 figures. Accepted to CVPR 2025
☆ Uncertainty-guided Perturbation for Image Super-Resolution Diffusion Model CVPR 2025
Diffusion-based image super-resolution methods have demonstrated significant advantages over GAN-based approaches, particularly in terms of perceptual quality. Building upon a lengthy Markov chain, diffusion-based methods possess remarkable modeling capacity, enabling them to achieve outstanding performance in real-world scenarios. Unlike previous methods that focus on modifying the noise schedule or sampling process to enhance performance, our approach emphasizes the improved utilization of LR information. We find that different regions of the LR image can be viewed as corresponding to different timesteps in a diffusion process, where flat areas are closer to the target HR distribution but edge and texture regions are farther away. In these flat areas, applying a slight noise is more advantageous for the reconstruction. We associate this characteristic with uncertainty and propose to apply uncertainty estimate to guide region-specific noise level control, a technique we refer to as Uncertainty-guided Noise Weighting. Pixels with lower uncertainty (i.e., flat regions) receive reduced noise to preserve more LR information, therefore improving performance. Furthermore, we modify the network architecture of previous methods to develop our Uncertainty-guided Perturbation Super-Resolution (UPSR) model. Extensive experimental results demonstrate that, despite reduced model size and training overhead, the proposed UWSR method outperforms current state-of-the-art methods across various datasets, both quantitatively and qualitatively.
comment: Accepted to CVPR 2025
☆ Can Text-to-Video Generation help Video-Language Alignment? CVPR 2025
Recent video-language alignment models are trained on sets of videos, each with an associated positive caption and a negative caption generated by large language models. A problem with this procedure is that negative captions may introduce linguistic biases, i.e., concepts are seen only as negatives and never associated with a video. While a solution would be to collect videos for the negative captions, existing databases lack the fine-grained variations needed to cover all possible negatives. In this work, we study whether synthetic videos can help to overcome this issue. Our preliminary analysis with multiple generators shows that, while promising on some tasks, synthetic videos harm the performance of the model on others. We hypothesize this issue is linked to noise (semantic and visual) in the generated videos and develop a method, SynViTA, that accounts for those. SynViTA dynamically weights the contribution of each synthetic video based on how similar its target caption is w.r.t. the real counterpart. Moreover, a semantic consistency loss makes the model focus on fine-grained differences across captions, rather than differences in video appearance. Experiments show that, on average, SynViTA improves over existing methods on VideoCon test sets and SSv2-Temporal, SSv2-Events, and ATP-Hard benchmarks, being a first promising step for using synthetic videos when learning video-language models.
comment: CVPR 2025. Project website at https://lucazanella.github.io/synvita/
☆ PM4Bench: A Parallel Multilingual Multi-Modal Multi-task Benchmark for Large Vision Language Model
Existing multilingual benchmarks for Large Vision Language Models (LVLMs) suffer from limitations including language-specific content biases, disjointed multimodal input formats, and a lack of safety evaluation. To address these gaps, we propose PM4Bench, the first Parallel Multilingual Multi-Modal Multi-task Benchmark for LVLMs. PM4Bench features a parallel corpus design across 10 languages, enabling fair and accurate cross-lingual comparisons. It includes the vision setting where text and queries are embedded in images, requiring LVLMs to simultaneously "see", "read", and "think", aligning with real-world applications. Additionally, PM\textsuperscript{4}Bench incorporates safety evaluations, addressing critical oversight in existing multilingual benchmarks. Using PM4Bench, we evaluate 11 mainstream LVLMs, revealing significant cross-linguistic performance disparities, particularly in vision settings, and identifying OCR capability as a key determinant of these imbalances. We will release PM4Bench at https://github.com/opendatalab/PM4Bench .
comment: Equal contribution: Junyuan Gao, Jiahe Song, Jiang Wu; Corresponding author: Conghui He
☆ Explaining Domain Shifts in Language: Concept erasing for Interpretable Image Classification CVPR2025
Concept-based models can map black-box representations to human-understandable concepts, which makes the decision-making process more transparent and then allows users to understand the reason behind predictions. However, domain-specific concepts often impact the final predictions, which subsequently undermine the model generalization capabilities, and prevent the model from being used in high-stake applications. In this paper, we propose a novel Language-guided Concept-Erasing (LanCE) framework. In particular, we empirically demonstrate that pre-trained vision-language models (VLMs) can approximate distinct visual domain shifts via domain descriptors while prompting large Language Models (LLMs) can easily simulate a wide range of descriptors of unseen visual domains. Then, we introduce a novel plug-in domain descriptor orthogonality (DDO) regularizer to mitigate the impact of these domain-specific concepts on the final predictions. Notably, the DDO regularizer is agnostic to the design of concept-based models and we integrate it into several prevailing models. Through evaluation of domain generalization on four standard benchmarks and three newly introduced benchmarks, we demonstrate that DDO can significantly improve the out-of-distribution (OOD) generalization over the previous state-of-the-art concept-based models.Our code is available at https://github.com/joeyz0z/LanCE.
comment: Accepted by CVPR2025
☆ Video-XL-Pro: Reconstructive Token Compression for Extremely Long Video Understanding
Despite advanced token compression techniques, existing multimodal large language models (MLLMs) still struggle with hour-long video understanding. In this work, we propose Video-XL-Pro, an efficient method for extremely long video understanding, built upon Reconstructive Compression of Tokens (ReCoT), a learnable module that leverages self-supervised learning to generate comprehensive and compact video tokens. ReCoT introduces two key components: (i) Dynamic Token Synthesizer (DTS): DTS generates pseudo-video tokens from static image tokens by learning intra-token relationships, which are then used in masked video modeling. (ii) Semantic-Guided Masking (SGM): SGM adaptively masks redundant visual tokens to facilitate more effective reconstructive learning. To improve training efficiency in MLLMs fine-tuning, we introduce a video-specific dataset pruning strategy and design a simple yet Query-aware Selector that enables the model to precisely locate query-relevant video tokens. With only 3B parameters, Video-XL-Pro outperforms most 7B models trained on larger datasets across multiple long video understanding benchmarks. Moreover, it can process over 8K frames on a single A100 GPU while maintaining high-quality performance.
☆ Global-Local Tree Search for Language Guided 3D Scene Generation CVPR 2025
Large Vision-Language Models (VLMs), such as GPT-4, have achieved remarkable success across various fields. However, there are few studies on 3D indoor scene generation with VLMs. This paper considers this task as a planning problem subject to spatial and layout common sense constraints. To solve the problem with a VLM, we propose a new global-local tree search algorithm. Globally, the method places each object sequentially and explores multiple placements during each placement process, where the problem space is represented as a tree. To reduce the depth of the tree, we decompose the scene structure hierarchically, i.e. room level, region level, floor object level, and supported object level. The algorithm independently generates the floor objects in different regions and supported objects placed on different floor objects. Locally, we also decompose the sub-task, the placement of each object, into multiple steps. The algorithm searches the tree of problem space. To leverage the VLM model to produce positions of objects, we discretize the top-down view space as a dense grid and fill each cell with diverse emojis to make to cells distinct. We prompt the VLM with the emoji grid and the VLM produces a reasonable location for the object by describing the position with the name of emojis. The quantitative and qualitative experimental results illustrate our approach generates more plausible 3D scenes than state-of-the-art approaches. Our source code is available at https://github.com/dw-dengwei/TreeSearchGen .
comment: Accepted by CVPR 2025
☆ MetaSpatial: Reinforcing 3D Spatial Reasoning in VLMs for the Metaverse
We present MetaSpatial, the first reinforcement learning (RL)-based framework designed to enhance 3D spatial reasoning in vision-language models (VLMs), enabling real-time 3D scene generation without the need for hard-coded optimizations. MetaSpatial addresses two core challenges: (i) the lack of internalized 3D spatial reasoning in VLMs, which limits their ability to generate realistic layouts, and (ii) the inefficiency of traditional supervised fine-tuning (SFT) for layout generation tasks, as perfect ground truth annotations are unavailable. Our key innovation is a multi-turn RL-based optimization mechanism that integrates physics-aware constraints and rendered image evaluations, ensuring generated 3D layouts are coherent, physically plausible, and aesthetically consistent. Methodologically, MetaSpatial introduces an adaptive, iterative reasoning process, where the VLM refines spatial arrangements over multiple turns by analyzing rendered outputs, improving scene coherence progressively. Empirical evaluations demonstrate that MetaSpatial significantly enhances the spatial consistency and formatting stability of various scale models. Post-training, object placements are more realistic, aligned, and functionally coherent, validating the effectiveness of RL for 3D spatial reasoning in metaverse, AR/VR, digital twins, and game development applications. Our code, data, and training pipeline are publicly available at https://github.com/PzySeere/MetaSpatial.
comment: Working Paper
☆ CFReID: Continual Few-shot Person Re-Identification
Real-world surveillance systems are dynamically evolving, requiring a person Re-identification model to continuously handle newly incoming data from various domains. To cope with these dynamics, Lifelong ReID (LReID) has been proposed to learn and accumulate knowledge across multiple domains incrementally. However, LReID models need to be trained on large-scale labeled data for each unseen domain, which are typically inaccessible due to privacy and cost concerns. In this paper, we propose a new paradigm called Continual Few-shot ReID (CFReID), which requires models to be incrementally trained using few-shot data and tested on all seen domains. Under few-shot conditions, CFREID faces two core challenges: 1) learning knowledge from few-shot data of unseen domain, and 2) avoiding catastrophic forgetting of seen domains. To tackle these two challenges, we propose a Stable Distribution Alignment (SDA) framework from feature distribution perspective. Specifically, our SDA is composed of two modules, i.e., Meta Distribution Alignment (MDA) and Prototype-based Few-shot Adaptation (PFA). To support the study of CFReID, we establish an evaluation benchmark for CFReID on five publicly available ReID datasets. Extensive experiments demonstrate that our SDA can enhance the few-shot learning and anti-forgetting capabilities under few-shot conditions. Notably, our approach, using only 5\% of the data, i.e., 32 IDs, significantly outperforms LReID's state-of-the-art performance, which requires 700 to 1,000 IDs.
comment: 16 pages, 8 figures
☆ SIT-FER: Integration of Semantic-, Instance-, Text-level Information for Semi-supervised Facial Expression Recognition
Semi-supervised deep facial expression recognition (SS-DFER) has gained increasingly research interest due to the difficulty in accessing sufficient labeled data in practical settings. However, existing SS-DFER methods mainly utilize generated semantic-level pseudo-labels for supervised learning, the unreliability of which compromises their performance and undermines the practical utility. In this paper, we propose a novel SS-DFER framework that simultaneously incorporates semantic, instance, and text-level information to generate high-quality pseudo-labels. Specifically, for the unlabeled data, considering the comprehensive knowledge within the textual descriptions and instance representations, we respectively calculate the similarities between the facial vision features and the corresponding textual and instance features to obtain the probabilities at the text- and instance-level. Combining with the semantic-level probability, these three-level probabilities are elaborately aggregated to gain the final pseudo-labels. Furthermore, to enhance the utilization of one-hot labels for the labeled data, we also incorporate text embeddings excavated from textual descriptions to co-supervise model training, enabling facial visual features to exhibit semantic correlations in the text space. Experiments on three datasets demonstrate that our method significantly outperforms current state-of-the-art SS-DFER methods and even exceeds fully supervised baselines. The code will be available at https://github.com/PatrickStarL/SIT-FER.
☆ PALATE: Peculiar Application of the Law of Total Expectation to Enhance the Evaluation of Deep Generative Models
Deep generative models (DGMs) have caused a paradigm shift in the field of machine learning, yielding noteworthy advancements in domains such as image synthesis, natural language processing, and other related areas. However, a comprehensive evaluation of these models that accounts for the trichotomy between fidelity, diversity, and novelty in generated samples remains a formidable challenge. A recently introduced solution that has emerged as a promising approach in this regard is the Feature Likelihood Divergence (FLD), a method that offers a theoretically motivated practical tool, yet also exhibits some computational challenges. In this paper, we propose PALATE, a novel enhancement to the evaluation of DGMs that addresses limitations of existing metrics. Our approach is based on a peculiar application of the law of total expectation to random variables representing accessible real data. When combined with the MMD baseline metric and DINOv2 feature extractor, PALATE offers a holistic evaluation framework that matches or surpasses state-of-the-art solutions while providing superior computational efficiency and scalability to large-scale datasets. Through a series of experiments, we demonstrate the effectiveness of the PALATE enhancement, contributing a computationally efficient, holistic evaluation approach that advances the field of DGMs assessment, especially in detecting sample memorization and evaluating generalization capabilities.
☆ MuMA: 3D PBR Texturing via Multi-Channel Multi-View Generation and Agentic Post-Processing
Current methods for 3D generation still fall short in physically based rendering (PBR) texturing, primarily due to limited data and challenges in modeling multi-channel materials. In this work, we propose MuMA, a method for 3D PBR texturing through Multi-channel Multi-view generation and Agentic post-processing. Our approach features two key innovations: 1) We opt to model shaded and albedo appearance channels, where the shaded channels enables the integration intrinsic decomposition modules for material properties. 2) Leveraging multimodal large language models, we emulate artists' techniques for material assessment and selection. Experiments demonstrate that MuMA achieves superior results in visual quality and material fidelity compared to existing methods.
comment: 17 pages, 14 figures
☆ Hiding Images in Diffusion Models by Editing Learned Score Functions
Hiding data using neural networks (i.e., neural steganography) has achieved remarkable success across both discriminative classifiers and generative adversarial networks. However, the potential of data hiding in diffusion models remains relatively unexplored. Current methods exhibit limitations in achieving high extraction accuracy, model fidelity, and hiding efficiency due primarily to the entanglement of the hiding and extraction processes with multiple denoising diffusion steps. To address these, we describe a simple yet effective approach that embeds images at specific timesteps in the reverse diffusion process by editing the learned score functions. Additionally, we introduce a parameter-efficient fine-tuning method that combines gradient-based parameter selection with low-rank adaptation to enhance model fidelity and hiding efficiency. Comprehensive experiments demonstrate that our method extracts high-quality images at human-indistinguishable levels, replicates the original model behaviors at both sample and population levels, and embeds images orders of magnitude faster than prior methods. Besides, our method naturally supports multi-recipient scenarios through independent extraction channels.
StableGS: A Floater-Free Framework for 3D Gaussian Splatting
Recent years have witnessed remarkable success of 3D Gaussian Splatting (3DGS) in novel view synthesis, surpassing prior differentiable rendering methods in both quality and efficiency. However, its training process suffers from coupled opacity-color optimization that frequently converges to local minima, producing floater artifacts that degrade visual fidelity. We present StableGS, a framework that eliminates floaters through cross-view depth consistency constraints while introducing a dual-opacity GS model to decouple geometry and material properties of translucent objects. To further enhance reconstruction quality in weakly-textured regions, we integrate DUSt3R depth estimation, significantly improving geometric stability. Our method fundamentally addresses 3DGS training instabilities, outperforming existing state-of-the-art methods across open-source datasets.
☆ InPO: Inversion Preference Optimization with Reparametrized DDIM for Efficient Diffusion Model Alignment CVPR2025
Without using explicit reward, direct preference optimization (DPO) employs paired human preference data to fine-tune generative models, a method that has garnered considerable attention in large language models (LLMs). However, exploration of aligning text-to-image (T2I) diffusion models with human preferences remains limited. In comparison to supervised fine-tuning, existing methods that align diffusion model suffer from low training efficiency and subpar generation quality due to the long Markov chain process and the intractability of the reverse process. To address these limitations, we introduce DDIM-InPO, an efficient method for direct preference alignment of diffusion models. Our approach conceptualizes diffusion model as a single-step generative model, allowing us to fine-tune the outputs of specific latent variables selectively. In order to accomplish this objective, we first assign implicit rewards to any latent variable directly via a reparameterization technique. Then we construct an Inversion technique to estimate appropriate latent variables for preference optimization. This modification process enables the diffusion model to only fine-tune the outputs of latent variables that have a strong correlation with the preference dataset. Experimental results indicate that our DDIM-InPO achieves state-of-the-art performance with just 400 steps of fine-tuning, surpassing all preference aligning baselines for T2I diffusion models in human preference evaluation tasks.
comment: Accepted by CVPR2025
☆ Latent Space Super-Resolution for Higher-Resolution Image Generation with Diffusion Models CVPR 2025
In this paper, we propose LSRNA, a novel framework for higher-resolution (exceeding 1K) image generation using diffusion models by leveraging super-resolution directly in the latent space. Existing diffusion models struggle with scaling beyond their training resolutions, often leading to structural distortions or content repetition. Reference-based methods address the issues by upsampling a low-resolution reference to guide higher-resolution generation. However, they face significant challenges: upsampling in latent space often causes manifold deviation, which degrades output quality. On the other hand, upsampling in RGB space tends to produce overly smoothed outputs. To overcome these limitations, LSRNA combines Latent space Super-Resolution (LSR) for manifold alignment and Region-wise Noise Addition (RNA) to enhance high-frequency details. Our extensive experiments demonstrate that integrating LSRNA outperforms state-of-the-art reference-based methods across various resolutions and metrics, while showing the critical role of latent space upsampling in preserving detail and sharpness. The code is available at https://github.com/3587jjh/LSRNA.
comment: Accepted by CVPR 2025
☆ Benchmarking Multi-modal Semantic Segmentation under Sensor Failures: Missing and Noisy Modality Robustness
Multi-modal semantic segmentation (MMSS) addresses the limitations of single-modality data by integrating complementary information across modalities. Despite notable progress, a significant gap persists between research and real-world deployment due to variability and uncertainty in multi-modal data quality. Robustness has thus become essential for practical MMSS applications. However, the absence of standardized benchmarks for evaluating robustness hinders further advancement. To address this, we first survey existing MMSS literature and categorize representative methods to provide a structured overview. We then introduce a robustness benchmark that evaluates MMSS models under three scenarios: Entire-Missing Modality (EMM), Random-Missing Modality (RMM), and Noisy Modality (NM). From a probabilistic standpoint, we model modality failure under two conditions: (1) all damaged combinations are equally probable; (2) each modality fails independently following a Bernoulli distribution. Based on these, we propose four metrics-$mIoU^{Avg}_{EMM}$, $mIoU^{E}_{EMM}$, $mIoU^{Avg}_{RMM}$, and $mIoU^{E}_{RMM}$-to assess model robustness under EMM and RMM. This work provides the first dedicated benchmark for MMSS robustness, offering new insights and tools to advance the field. Source code is available at https://github.com/Chenfei-Liao/Multi-Modal-Semantic-Segmentation-Robustness-Benchmark.
☆ ReconDreamer++: Harmonizing Generative and Reconstructive Models for Driving Scene Representation
Combining reconstruction models with generative models has emerged as a promising paradigm for closed-loop simulation in autonomous driving. For example, ReconDreamer has demonstrated remarkable success in rendering large-scale maneuvers. However, a significant gap remains between the generated data and real-world sensor observations, particularly in terms of fidelity for structured elements, such as the ground surface. To address these challenges, we propose ReconDreamer++, an enhanced framework that significantly improves the overall rendering quality by mitigating the domain gap and refining the representation of the ground surface. Specifically, ReconDreamer++ introduces the Novel Trajectory Deformable Network (NTDNet), which leverages learnable spatial deformation mechanisms to bridge the domain gap between synthesized novel views and original sensor observations. Moreover, for structured elements such as the ground surface, we preserve geometric prior knowledge in 3D Gaussians, and the optimization process focuses on refining appearance attributes while preserving the underlying geometric structure. Experimental evaluations conducted on multiple datasets (Waymo, nuScenes, PandaSet, and EUVS) confirm the superior performance of ReconDreamer++. Specifically, on Waymo, ReconDreamer++ achieves performance comparable to Street Gaussians for the original trajectory while significantly outperforming ReconDreamer on novel trajectories. In particular, it achieves substantial improvements, including a 6.1% increase in NTA-IoU, a 23. 0% improvement in FID, and a remarkable 4.5% gain in the ground surface metric NTL-IoU, highlighting its effectiveness in accurately reconstructing structured elements such as the road surface.
comment: Project Page: https://recondreamer-plus.github.io/
☆ On the Perception Bottleneck of VLMs for Chart Understanding
Chart understanding requires models to effectively analyze and reason about numerical data, textual elements, and complex visual components. Our observations reveal that the perception capabilities of existing large vision-language models (LVLMs) constitute a critical bottleneck in this process. In this study, we delve into this perception bottleneck by decomposing it into two components: the vision encoder bottleneck, where the visual representation may fail to encapsulate the correct information, and the extraction bottleneck, where the language model struggles to extract the necessary information from the provided visual representations. Through comprehensive experiments, we find that (1) the information embedded within visual representations is substantially richer than what is typically captured by linear extractors, such as the widely used retrieval accuracy metric; (2) While instruction tuning effectively enhances the extraction capability of LVLMs, the vision encoder remains a critical bottleneck, demanding focused attention and improvement. Therefore, we further enhance the visual encoder to mitigate the vision encoder bottleneck under a contrastive learning framework. Empirical results demonstrate that our approach significantly mitigates the perception bottleneck and improves the ability of LVLMs to comprehend charts. Code is publicly available at https://github.com/hkust-nlp/Vision4Chart.
☆ A Simple yet Effective Layout Token in Large Language Models for Document Understanding CVPR 2025
Recent methods that integrate spatial layouts with text for document understanding in large language models (LLMs) have shown promising results. A commonly used method is to represent layout information as text tokens and interleave them with text content as inputs to the LLMs. However, such a method still demonstrates limitations, as it requires additional position IDs for tokens that are used to represent layout information. Due to the constraint on max position IDs, assigning them to layout information reduces those available for text content, reducing the capacity for the model to learn from the text during training, while also introducing a large number of potentially untrained position IDs during long-context inference, which can hinder performance on document understanding tasks. To address these issues, we propose LayTokenLLM, a simple yet effective method for document understanding. LayTokenLLM represents layout information as a single token per text segment and uses a specialized positional encoding scheme. It shares position IDs between text and layout tokens, eliminating the need for additional position IDs. This design maintains the model's capacity to learn from text while mitigating long-context issues during inference. Furthermore, a novel pre-training objective called Next Interleaved Text and Layout Token Prediction (NTLP) is devised to enhance cross-modality learning between text and layout tokens. Extensive experiments show that LayTokenLLM outperforms existing layout-integrated LLMs and MLLMs of similar scales on multi-page document understanding tasks, as well as most single-page tasks.
comment: CVPR 2025
☆ CQ-DINO: Mitigating Gradient Dilution via Category Queries for Vast Vocabulary Object Detection
With the exponential growth of data, traditional object detection methods are increasingly struggling to handle vast vocabulary object detection tasks effectively. We analyze two key limitations of classification-based detectors: positive gradient dilution, where rare positive categories receive insufficient learning signals, and hard negative gradient dilution, where discriminative gradients are overwhelmed by numerous easy negatives. To address these challenges, we propose CQ-DINO, a category query-based object detection framework that reformulates classification as a contrastive task between object queries and learnable category queries. Our method introduces image-guided query selection, which reduces the negative space by adaptively retrieving top-K relevant categories per image via cross-attention, thereby rebalancing gradient distributions and facilitating implicit hard example mining. Furthermore, CQ-DINO flexibly integrates explicit hierarchical category relationships in structured datasets (e.g., V3Det) or learns implicit category correlations via self-attention in generic datasets (e.g., COCO). Experiments demonstrate that CQ-DINO achieves superior performance on the challenging V3Det benchmark (surpassing previous methods by 2.1% AP) while maintaining competitiveness in COCO. Our work provides a scalable solution for real-world detection systems requiring wide category coverage. The dataset and code will be publicly at https://github.com/RedAIGC/CQ-DINO.
☆ Teller: Real-Time Streaming Audio-Driven Portrait Animation with Autoregressive Motion Generation CVPR 2025
In this work, we introduce the first autoregressive framework for real-time, audio-driven portrait animation, a.k.a, talking head. Beyond the challenge of lengthy animation times, a critical challenge in realistic talking head generation lies in preserving the natural movement of diverse body parts. To this end, we propose Teller, the first streaming audio-driven protrait animation framework with autoregressive motion generation. Specifically, Teller first decomposes facial and body detail animation into two components: Facial Motion Latent Generation (FMLG) based on an autoregressive transfromer, and movement authenticity refinement using a Efficient Temporal Module (ETM).Concretely, FMLG employs a Residual VQ model to map the facial motion latent from the implicit keypoint-based model into discrete motion tokens, which are then temporally sliced with audio embeddings. This enables the AR tranformer to learn real-time, stream-based mappings from audio to motion. Furthermore, Teller incorporate ETM to capture finer motion details. This module ensures the physical consistency of body parts and accessories, such as neck muscles and earrings, improving the realism of these movements. Teller is designed to be efficient, surpassing the inference speed of diffusion-based models (Hallo 20.93s vs. Teller 0.92s for one second video generation), and achieves a real-time streaming performance of up to 25 FPS. Extensive experiments demonstrate that our method outperforms recent audio-driven portrait animation models, especially in small movements, as validated by human evaluations with a significant margin in quality and realism.
comment: Accept in CVPR 2025 Conference Submission
☆ Breaking the Encoder Barrier for Seamless Video-Language Understanding
Most Video-Large Language Models (Video-LLMs) adopt an encoder-decoder framework, where a vision encoder extracts frame-wise features for processing by a language model. However, this approach incurs high computational costs, introduces resolution biases, and struggles to capture fine-grained multimodal interactions. To overcome these limitations, we propose ELVA, an encoder-free Video-LLM that directly models nuanced video-language interactions without relying on a vision encoder. ELVA employs token merging to construct a bottom-up hierarchical representation and incorporates a video guidance supervisor for direct spatiotemporal representation learning. Additionally, a hybrid-resolution mechanism strategically integrates high- and low-resolution frames as inputs to achieve an optimal balance between performance and efficiency. With only 7M publicly available video-text pairs, ELVA achieves performance on par with encoder-based Video-LLMs while reducing FLOPs by up to 95\% and inference latency by 92\%, offering a scalable and efficient solution for real-time video understanding.
comment: 12 pages
☆ 4DGC: Rate-Aware 4D Gaussian Compression for Efficient Streamable Free-Viewpoint Video CVPR2025
3D Gaussian Splatting (3DGS) has substantial potential for enabling photorealistic Free-Viewpoint Video (FVV) experiences. However, the vast number of Gaussians and their associated attributes poses significant challenges for storage and transmission. Existing methods typically handle dynamic 3DGS representation and compression separately, neglecting motion information and the rate-distortion (RD) trade-off during training, leading to performance degradation and increased model redundancy. To address this gap, we propose 4DGC, a novel rate-aware 4D Gaussian compression framework that significantly reduces storage size while maintaining superior RD performance for FVV. Specifically, 4DGC introduces a motion-aware dynamic Gaussian representation that utilizes a compact motion grid combined with sparse compensated Gaussians to exploit inter-frame similarities. This representation effectively handles large motions, preserving quality and reducing temporal redundancy. Furthermore, we present an end-to-end compression scheme that employs differentiable quantization and a tiny implicit entropy model to compress the motion grid and compensated Gaussians efficiently. The entire framework is jointly optimized using a rate-distortion trade-off. Extensive experiments demonstrate that 4DGC supports variable bitrates and consistently outperforms existing methods in RD performance across multiple datasets.
comment: CVPR2025
☆ Panorama Generation From NFoV Image Done Right CVPR2025
Generating 360-degree panoramas from narrow field of view (NFoV) image is a promising computer vision task for Virtual Reality (VR) applications. Existing methods mostly assess the generated panoramas with InceptionNet or CLIP based metrics, which tend to perceive the image quality and is \textbf{not suitable for evaluating the distortion}. In this work, we first propose a distortion-specific CLIP, named Distort-CLIP to accurately evaluate the panorama distortion and discover the \textbf{``visual cheating''} phenomenon in previous works (\ie, tending to improve the visual results by sacrificing distortion accuracy). This phenomenon arises because prior methods employ a single network to learn the distinct panorama distortion and content completion at once, which leads the model to prioritize optimizing the latter. To address the phenomenon, we propose \textbf{PanoDecouple}, a decoupled diffusion model framework, which decouples the panorama generation into distortion guidance and content completion, aiming to generate panoramas with both accurate distortion and visual appeal. Specifically, we design a DistortNet for distortion guidance by imposing panorama-specific distortion prior and a modified condition registration mechanism; and a ContentNet for content completion by imposing perspective image information. Additionally, a distortion correction loss function with Distort-CLIP is introduced to constrain the distortion explicitly. The extensive experiments validate that PanoDecouple surpasses existing methods both in distortion and visual metrics.
comment: CVPR2025. Project page:https://isee-laboratory.github.io/PanoDecouple/ Code:https://github.com/iSEE-Laboratory/PanoDecouple/
☆ U-REPA: Aligning Diffusion U-Nets to ViTs
Representation Alignment (REPA) that aligns Diffusion Transformer (DiT) hidden-states with ViT visual encoders has proven highly effective in DiT training, demonstrating superior convergence properties, but it has not been validated on the canonical diffusion U-Net architecture that shows faster convergence compared to DiTs. However, adapting REPA to U-Net architectures presents unique challenges: (1) different block functionalities necessitate revised alignment strategies; (2) spatial-dimension inconsistencies emerge from U-Net's spatial downsampling operations; (3) space gaps between U-Net and ViT hinder the effectiveness of tokenwise alignment. To encounter these challenges, we propose U-REPA, a representation alignment paradigm that bridges U-Net hidden states and ViT features as follows: Firstly, we propose via observation that due to skip connection, the middle stage of U-Net is the best alignment option. Secondly, we propose upsampling of U-Net features after passing them through MLPs. Thirdly, we observe difficulty when performing tokenwise similarity alignment, and further introduces a manifold loss that regularizes the relative similarity between samples. Experiments indicate that the resulting U-REPA could achieve excellent generation quality and greatly accelerates the convergence speed. With CFG guidance interval, U-REPA could reach $FID<1.5$ in 200 epochs or 1M iterations on ImageNet 256 $\times$ 256, and needs only half the total epochs to perform better than REPA. Codes are available at https://github.com/YuchuanTian/U-REPA.
comment: 15 pages, 7 figures
☆ Fast and Physically-based Neural Explicit Surface for Relightable Human Avatars
Efficiently modeling relightable human avatars from sparse-view videos is crucial for AR/VR applications. Current methods use neural implicit representations to capture dynamic geometry and reflectance, which incur high costs due to the need for dense sampling in volume rendering. To overcome these challenges, we introduce Physically-based Neural Explicit Surface (PhyNES), which employs compact neural material maps based on the Neural Explicit Surface (NES) representation. PhyNES organizes human models in a compact 2D space, enhancing material disentanglement efficiency. By connecting Signed Distance Fields to explicit surfaces, PhyNES enables efficient geometry inference around a parameterized human shape model. This approach models dynamic geometry, texture, and material maps as 2D neural representations, enabling efficient rasterization. PhyNES effectively captures physical surface attributes under varying illumination, enabling real-time physically-based rendering. Experiments show that PhyNES achieves relighting quality comparable to SOTA methods while significantly improving rendering speed, memory efficiency, and reconstruction quality.
☆ VTD-CLIP: Video-to-Text Discretization via Prompting CLIP
Vision-language models bridge visual and linguistic understanding and have proven to be powerful for video recognition tasks. Existing approaches primarily rely on parameter-efficient fine-tuning of image-text pre-trained models, yet they often suffer from limited interpretability and poor generalization due to inadequate temporal modeling. To address these, we propose a simple yet effective video-to-text discretization framework. Our method repurposes the frozen text encoder to construct a visual codebook from video class labels due to the many-to-one contrastive alignment between visual and textual embeddings in multimodal pretraining. This codebook effectively transforms temporal visual data into textual tokens via feature lookups and offers interpretable video representations through explicit video modeling. Then, to enhance robustness against irrelevant or noisy frames, we introduce a confidence-aware fusion module that dynamically weights keyframes by assessing their semantic relevance via the codebook. Furthermore, our method incorporates learnable text prompts to conduct adaptive codebook updates. Extensive experiments on HMDB-51, UCF-101, SSv2, and Kinetics-400 have validated the superiority of our approach, achieving more competitive improvements over state-of-the-art methods. The code will be publicly available at https://github.com/isxinxin/VTD-CLIP.
☆ Instruct-CLIP: Improving Instruction-Guided Image Editing with Automated Data Refinement Using Contrastive Learning
Although natural language instructions offer an intuitive way to guide automated image editing, deep-learning models often struggle to achieve high-quality results, largely due to challenges in creating large, high-quality training datasets. Previous work has typically relied on text-toimage (T2I) generative models to produce pairs of original and edited images that simulate the input/output of an instruction-guided image-editing model. However, these image pairs often fail to align with the specified edit instructions due to the limitations of T2I models, which negatively impacts models trained on such datasets. To address this, we present Instruct-CLIP, a self-supervised method that learns the semantic changes between original and edited images to refine and better align the instructions in existing datasets. Furthermore, we adapt Instruct-CLIP to handle noisy latent images and diffusion timesteps so that it can be used to train latent diffusion models (LDMs) [19] and efficiently enforce alignment between the edit instruction and the image changes in latent space at any step of the diffusion pipeline. We use Instruct-CLIP to correct the InstructPix2Pix dataset and get over 120K refined samples we then use to fine-tune their model, guided by our novel Instruct-CLIP-based loss function. The resulting model can produce edits that are more aligned with the given instructions. Our code and dataset are available at https://github.com/SherryXTChen/Instruct-CLIP.git.
comment: Computer Vision and Pattern Recognition 2025
☆ Offline Meteorology-Pollution Coupling Global Air Pollution Forecasting Model with Bilinear Pooling
Air pollution has become a major threat to human health, making accurate forecasting crucial for pollution control. Traditional physics-based models forecast global air pollution by coupling meteorology and pollution processes, using either online or offline methods depending on whether fully integrated with meteorological models and run simultaneously. However, the high computational demands of both methods severely limit real-time prediction efficiency. Existing deep learning (DL) solutions employ online coupling strategies for global air pollution forecasting, which finetune pollution forecasting based on pretrained atmospheric models, requiring substantial training resources. This study pioneers a DL-based offline coupling framework that utilizes bilinear pooling to achieve offline coupling between meteorological fields and pollutants. The proposed model requires only 13% of the parameters of DL-based online coupling models while achieving competitive performance. Compared with the state-of-the-art global air pollution forecasting model CAMS, our approach demonstrates superiority in 63% variables across all forecast time steps and 85% variables in predictions exceeding 48 hours. This work pioneers experimental validation of the effectiveness of meteorological fields in DL-based global air pollution forecasting, demonstrating that offline coupling meteorological fields with pollutants can achieve a 15% relative reduction in RMSE across all pollution variables. The research establishes a new paradigm for real-time global air pollution warning systems and delivers critical technical support for developing more efficient and comprehensive AI-powered global atmospheric forecasting frameworks.
☆ Knowledge Graph Enhanced Generative Multi-modal Models for Class-Incremental Learning
Continual learning in computer vision faces the critical challenge of catastrophic forgetting, where models struggle to retain prior knowledge while adapting to new tasks. Although recent studies have attempted to leverage the generalization capabilities of pre-trained models to mitigate overfitting on current tasks, models still tend to forget details of previously learned categories as tasks progress, leading to misclassification. To address these limitations, we introduce a novel Knowledge Graph Enhanced Generative Multi-modal model (KG-GMM) that builds an evolving knowledge graph throughout the learning process. Our approach utilizes relationships within the knowledge graph to augment the class labels and assigns different relations to similar categories to enhance model differentiation. During testing, we propose a Knowledge Graph Augmented Inference method that locates specific categories by analyzing relationships within the generated text, thereby reducing the loss of detailed information about old classes when learning new knowledge and alleviating forgetting. Experiments demonstrate that our method effectively leverages relational information to help the model correct mispredictions, achieving state-of-the-art results in both conventional CIL and few-shot CIL settings, confirming the efficacy of knowledge graphs at preserving knowledge in the continual learning scenarios.
☆ DashGaussian: Optimizing 3D Gaussian Splatting in 200 Seconds CVPR2025
3D Gaussian Splatting (3DGS) renders pixels by rasterizing Gaussian primitives, where the rendering resolution and the primitive number, concluded as the optimization complexity, dominate the time cost in primitive optimization. In this paper, we propose DashGaussian, a scheduling scheme over the optimization complexity of 3DGS that strips redundant complexity to accelerate 3DGS optimization. Specifically, we formulate 3DGS optimization as progressively fitting 3DGS to higher levels of frequency components in the training views, and propose a dynamic rendering resolution scheme that largely reduces the optimization complexity based on this formulation. Besides, we argue that a specific rendering resolution should cooperate with a proper primitive number for a better balance between computing redundancy and fitting quality, where we schedule the growth of the primitives to synchronize with the rendering resolution. Extensive experiments show that our method accelerates the optimization of various 3DGS backbones by 45.7% on average while preserving the rendering quality.
comment: Accepted by CVPR2025. Project page: https://dashgaussian.github.io
☆ PDDM: Pseudo Depth Diffusion Model for RGB-PD Semantic Segmentation Based in Complex Indoor Scenes
The integration of RGB and depth modalities significantly enhances the accuracy of segmenting complex indoor scenes, with depth data from RGB-D cameras playing a crucial role in this improvement. However, collecting an RGB-D dataset is more expensive than an RGB dataset due to the need for specialized depth sensors. Aligning depth and RGB images also poses challenges due to sensor positioning and issues like missing data and noise. In contrast, Pseudo Depth (PD) from high-precision depth estimation algorithms can eliminate the dependence on RGB-D sensors and alignment processes, as well as provide effective depth information and show significant potential in semantic segmentation. Therefore, to explore the practicality of utilizing pseudo depth instead of real depth for semantic segmentation, we design an RGB-PD segmentation pipeline to integrate RGB and pseudo depth and propose a Pseudo Depth Aggregation Module (PDAM) for fully exploiting the informative clues provided by the diverse pseudo depth maps. The PDAM aggregates multiple pseudo depth maps into a single modality, making it easily adaptable to other RGB-D segmentation methods. In addition, the pre-trained diffusion model serves as a strong feature extractor for RGB segmentation tasks, but multi-modal diffusion-based segmentation methods remain unexplored. Therefore, we present a Pseudo Depth Diffusion Model (PDDM) that adopts a large-scale text-image diffusion model as a feature extractor and a simple yet effective fusion strategy to integrate pseudo depth. To verify the applicability of pseudo depth and our PDDM, we perform extensive experiments on the NYUv2 and SUNRGB-D datasets. The experimental results demonstrate that pseudo depth can effectively enhance segmentation performance, and our PDDM achieves state-of-the-art performance, outperforming other methods by +6.98 mIoU on NYUv2 and +2.11 mIoU on SUNRGB-D.
☆ Resource-Efficient Motion Control for Video Generation via Dynamic Mask Guidance
Recent advances in diffusion models bring new vitality to visual content creation. However, current text-to-video generation models still face significant challenges such as high training costs, substantial data requirements, and difficulties in maintaining consistency between given text and motion of the foreground object. To address these challenges, we propose mask-guided video generation, which can control video generation through mask motion sequences, while requiring limited training data. Our model enhances existing architectures by incorporating foreground masks for precise text-position matching and motion trajectory control. Through mask motion sequences, we guide the video generation process to maintain consistent foreground objects throughout the sequence. Additionally, through a first-frame sharing strategy and autoregressive extension approach, we achieve more stable and longer video generation. Extensive qualitative and quantitative experiments demonstrate that this approach excels in various video generation tasks, such as video editing and generating artistic videos, outperforming previous methods in terms of consistency and quality. Our generated results can be viewed in the supplementary materials.
☆ LiDAR Remote Sensing Meets Weak Supervision: Concepts, Methods, and Perspectives
LiDAR (Light Detection and Ranging) enables rapid and accurate acquisition of three-dimensional spatial data, widely applied in remote sensing areas such as surface mapping, environmental monitoring, urban modeling, and forestry inventory. LiDAR remote sensing primarily includes data interpretation and LiDAR-based inversion. However, LiDAR interpretation typically relies on dense and precise annotations, which are costly and time-consuming. Similarly, LiDAR inversion depends on scarce supervisory signals and expensive field surveys for annotations. To address this challenge, weakly supervised learning has gained significant attention in recent years, with many methods emerging to tackle LiDAR remote sensing tasks using incomplete, inaccurate, and inexact annotations, as well as annotations from other domains. Existing review articles treat LiDAR interpretation and inversion as separate tasks. This review, for the first time, adopts a unified weakly supervised learning perspective to systematically examine research on both LiDAR interpretation and inversion. We summarize the latest advancements, provide a comprehensive review of the development and application of weakly supervised techniques in LiDAR remote sensing, and discuss potential future research directions in this field.
☆ PP-FormulaNet: Bridging Accuracy and Efficiency in Advanced Formula Recognition
Formula recognition is an important task in document intelligence. It involves converting mathematical expressions from document images into structured symbolic formats that computers can easily work with. LaTeX is the most common format used for this purpose. In this work, we present PP-FormulaNet, a state-of-the-art formula recognition model that excels in both accuracy and efficiency. To meet the diverse needs of applications, we have developed two specialized models: PP-FormulaNet-L, tailored for high-accuracy scenarios, and PP-FormulaNet-S, optimized for high-efficiency contexts. Our extensive evaluations reveal that PP-FormulaNet-L attains accuracy levels that surpass those of prominent models such as UniMERNet by a significant 6%. Conversely, PP-FormulaNet-S operates at speeds that are over 16 times faster. These advancements facilitate seamless integration of PP-FormulaNet into a broad spectrum of document processing environments that involve intricate mathematical formulas. Furthermore, we introduce a Formula Mining System, which is capable of extracting a vast amount of high-quality formula data. This system further enhances the robustness and applicability of our formula recognition model. Code and models are publicly available at PaddleOCR(https://github.com/PaddlePaddle/PaddleOCR) and PaddleX(https://github.com/PaddlePaddle/PaddleX).
☆ Exploring State Space Model in Wavelet Domain: An Infrared and Visible Image Fusion Network via Wavelet Transform and State Space Model
Deep learning techniques have revolutionized the infrared and visible image fusion (IVIF), showing remarkable efficacy on complex scenarios. However, current methods do not fully combine frequency domain features with global semantic information, which will result in suboptimal extraction of global features across modalities and insufficient preservation of local texture details. To address these issues, we propose Wavelet-Mamba (W-Mamba), which integrates wavelet transform with the state-space model (SSM). Specifically, we introduce Wavelet-SSM module, which incorporates wavelet-based frequency domain feature extraction and global information extraction through SSM, thereby effectively capturing both global and local features. Additionally, we propose a cross-modal feature attention modulation, which facilitates efficient interaction and fusion between different modalities. The experimental results indicate that our method achieves both visually compelling results and superior performance compared to current state-of-the-art methods. Our code is available at https://github.com/Lmmh058/W-Mamba.
☆ Do Your Best and Get Enough Rest for Continual Learning CVPR
According to the forgetting curve theory, we can enhance memory retention by learning extensive data and taking adequate rest. This means that in order to effectively retain new knowledge, it is essential to learn it thoroughly and ensure sufficient rest so that our brain can memorize without forgetting. The main takeaway from this theory is that learning extensive data at once necessitates sufficient rest before learning the same data again. This aspect of human long-term memory retention can be effectively utilized to address the continual learning of neural networks. Retaining new knowledge for a long period of time without catastrophic forgetting is the critical problem of continual learning. Therefore, based on Ebbinghaus' theory, we introduce the view-batch model that adjusts the learning schedules to optimize the recall interval between retraining the same samples. The proposed view-batch model allows the network to get enough rest to learn extensive knowledge from the same samples with a recall interval of sufficient length. To this end, we specifically present two approaches: 1) a replay method that guarantees the optimal recall interval, and 2) a self-supervised learning that acquires extensive knowledge from a single training sample at a time. We empirically show that these approaches of our method are aligned with the forgetting curve theory, which can enhance long-term memory. In our experiments, we also demonstrate that our method significantly improves many state-of-the-art continual learning methods in various protocols and scenarios. We open-source this project at https://github.com/hankyul2/ViewBatchModel.
comment: Accepted by IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) 2025
☆ DiffusedWrinkles: A Diffusion-Based Model for Data-Driven Garment Animation BMVC 2024
We present a data-driven method for learning to generate animations of 3D garments using a 2D image diffusion model. In contrast to existing methods, typically based on fully connected networks, graph neural networks, or generative adversarial networks, which have difficulties to cope with parametric garments with fine wrinkle detail, our approach is able to synthesize high-quality 3D animations for a wide variety of garments and body shapes, while being agnostic to the garment mesh topology. Our key idea is to represent 3D garment deformations as a 2D layout-consistent texture that encodes 3D offsets with respect to a parametric garment template. Using this representation, we encode a large dataset of garments simulated in various motions and shapes and train a novel conditional diffusion model that is able to synthesize high-quality pose-shape-and-design dependent 3D garment deformations. Since our model is generative, we can synthesize various plausible deformations for a given target pose, shape, and design. Additionally, we show that we can further condition our model using an existing garment state, which enables the generation of temporally coherent sequences.
comment: BMVC 2024
☆ MoST: Efficient Monarch Sparse Tuning for 3D Representation Learning
We introduce Monarch Sparse Tuning (MoST), the first reparameterization-based parameter-efficient fine-tuning (PEFT) method tailored for 3D representation learning. Unlike existing adapter-based and prompt-tuning 3D PEFT methods, MoST introduces no additional inference overhead and is compatible with many 3D representation learning backbones. At its core, we present a new family of structured matrices for 3D point clouds, Point Monarch, which can capture local geometric features of irregular points while offering high expressiveness. MoST reparameterizes the dense update weight matrices as our sparse Point Monarch matrices, significantly reducing parameters while retaining strong performance. Experiments on various backbones show that MoST is simple, effective, and highly generalizable. It captures local features in point clouds, achieving state-of-the-art results on multiple benchmarks, e.g., 97.5% acc. on ScanObjectNN (PB_50_RS) and 96.2% on ModelNet40 classification, while it can also combine with other matrix decompositions (e.g., Low-rank, Kronecker) to further reduce parameters.
comment: 11 pages, 6 figures, 6 tables. Code and weights are available at https://github.com/xhanxu/MoST
☆ MaSS13K: A Matting-level Semantic Segmentation Benchmark
High-resolution semantic segmentation is essential for applications such as image editing, bokeh imaging, AR/VR, etc. Unfortunately, existing datasets often have limited resolution and lack precise mask details and boundaries. In this work, we build a large-scale, matting-level semantic segmentation dataset, named MaSS13K, which consists of 13,348 real-world images, all at 4K resolution. MaSS13K provides high-quality mask annotations of a number of objects, which are categorized into seven categories: human, vegetation, ground, sky, water, building, and others. MaSS13K features precise masks, with an average mask complexity 20-50 times higher than existing semantic segmentation datasets. We consequently present a method specifically designed for high-resolution semantic segmentation, namely MaSSFormer, which employs an efficient pixel decoder that aggregates high-level semantic features and low-level texture features across three stages, aiming to produce high-resolution masks with minimal computational cost. Finally, we propose a new learning paradigm, which integrates the high-quality masks of the seven given categories with pseudo labels from new classes, enabling MaSSFormer to transfer its accurate segmentation capability to other classes of objects. Our proposed MaSSFormer is comprehensively evaluated on the MaSS13K benchmark together with 14 representative segmentation models. We expect that our meticulously annotated MaSS13K dataset and the MaSSFormer model can facilitate the research of high-resolution and high-quality semantic segmentation. Datasets and codes can be found at https://github.com/xiechenxi99/MaSS13K.
☆ MonoInstance: Enhancing Monocular Priors via Multi-view Instance Alignment for Neural Rendering and Reconstruction CVPR 2025
Monocular depth priors have been widely adopted by neural rendering in multi-view based tasks such as 3D reconstruction and novel view synthesis. However, due to the inconsistent prediction on each view, how to more effectively leverage monocular cues in a multi-view context remains a challenge. Current methods treat the entire estimated depth map indiscriminately, and use it as ground truth supervision, while ignoring the inherent inaccuracy and cross-view inconsistency in monocular priors. To resolve these issues, we propose MonoInstance, a general approach that explores the uncertainty of monocular depths to provide enhanced geometric priors for neural rendering and reconstruction. Our key insight lies in aligning each segmented instance depths from multiple views within a common 3D space, thereby casting the uncertainty estimation of monocular depths into a density measure within noisy point clouds. For high-uncertainty areas where depth priors are unreliable, we further introduce a constraint term that encourages the projected instances to align with corresponding instance masks on nearby views. MonoInstance is a versatile strategy which can be seamlessly integrated into various multi-view neural rendering frameworks. Our experimental results demonstrate that MonoInstance significantly improves the performance in both reconstruction and novel view synthesis under various benchmarks.
comment: Accepted by CVPR 2025. Project page: https://wen-yuan-zhang.github.io/MonoInstance/
☆ NeRFPrior: Learning Neural Radiance Field as a Prior for Indoor Scene Reconstruction CVPR 2025
Recently, it has shown that priors are vital for neural implicit functions to reconstruct high-quality surfaces from multi-view RGB images. However, current priors require large-scale pre-training, and merely provide geometric clues without considering the importance of color. In this paper, we present NeRFPrior, which adopts a neural radiance field as a prior to learn signed distance fields using volume rendering for surface reconstruction. Our NeRF prior can provide both geometric and color clues, and also get trained fast under the same scene without additional data. Based on the NeRF prior, we are enabled to learn a signed distance function (SDF) by explicitly imposing a multi-view consistency constraint on each ray intersection for surface inference. Specifically, at each ray intersection, we use the density in the prior as a coarse geometry estimation, while using the color near the surface as a clue to check its visibility from another view angle. For the textureless areas where the multi-view consistency constraint does not work well, we further introduce a depth consistency loss with confidence weights to infer the SDF. Our experimental results outperform the state-of-the-art methods under the widely used benchmarks.
comment: Accepted by CVPR 2025. Project page: https://wen-yuan-zhang.github.io/NeRFPrior/
☆ Context-Enhanced Memory-Refined Transformer for Online Action Detection CVPR 2025
Online Action Detection (OAD) detects actions in streaming videos using past observations. State-of-the-art OAD approaches model past observations and their interactions with an anticipated future. The past is encoded using short- and long-term memories to capture immediate and long-range dependencies, while anticipation compensates for missing future context. We identify a training-inference discrepancy in existing OAD methods that hinders learning effectiveness. The training uses varying lengths of short-term memory, while inference relies on a full-length short-term memory. As a remedy, we propose a Context-enhanced Memory-Refined Transformer (CMeRT). CMeRT introduces a context-enhanced encoder to improve frame representations using additional near-past context. It also features a memory-refined decoder to leverage near-future generation to enhance performance. CMeRT achieves state-of-the-art in online detection and anticipation on THUMOS'14, CrossTask, and EPIC-Kitchens-100.
comment: Accepted by CVPR 2025
☆ Cost-Sensitive Learning for Long-Tailed Temporal Action Segmentation
Temporal action segmentation in untrimmed procedural videos aims to densely label frames into action classes. These videos inherently exhibit long-tailed distributions, where actions vary widely in frequency and duration. In temporal action segmentation approaches, we identified a bi-level learning bias. This bias encompasses (1) a class-level bias, stemming from class imbalance favoring head classes, and (2) a transition-level bias arising from variations in transitions, prioritizing commonly observed transitions. As a remedy, we introduce a constrained optimization problem to alleviate both biases. We define learning states for action classes and their associated transitions and integrate them into the optimization process. We propose a novel cost-sensitive loss function formulated as a weighted cross-entropy loss, with weights adaptively adjusted based on the learning state of actions and their transitions. Experiments on three challenging temporal segmentation benchmarks and various frameworks demonstrate the effectiveness of our approach, resulting in significant improvements in both per-class frame-wise and segment-wise performance.
comment: BMCV 2024
Diffusion-4K: Ultra-High-Resolution Image Synthesis with Latent Diffusion Models CVPR 2025
In this paper, we present Diffusion-4K, a novel framework for direct ultra-high-resolution image synthesis using text-to-image diffusion models. The core advancements include: (1) Aesthetic-4K Benchmark: addressing the absence of a publicly available 4K image synthesis dataset, we construct Aesthetic-4K, a comprehensive benchmark for ultra-high-resolution image generation. We curated a high-quality 4K dataset with carefully selected images and captions generated by GPT-4o. Additionally, we introduce GLCM Score and Compression Ratio metrics to evaluate fine details, combined with holistic measures such as FID, Aesthetics and CLIPScore for a comprehensive assessment of ultra-high-resolution images. (2) Wavelet-based Fine-tuning: we propose a wavelet-based fine-tuning approach for direct training with photorealistic 4K images, applicable to various latent diffusion models, demonstrating its effectiveness in synthesizing highly detailed 4K images. Consequently, Diffusion-4K achieves impressive performance in high-quality image synthesis and text prompt adherence, especially when powered by modern large-scale diffusion models (e.g., SD3-2B and Flux-12B). Extensive experimental results from our benchmark demonstrate the superiority of Diffusion-4K in ultra-high-resolution image synthesis.
comment: Accepted to CVPR 2025
☆ Human-Object Interaction with Vision-Language Model Guided Relative Movement Dynamics
Human-Object Interaction (HOI) is vital for advancing simulation, animation, and robotics, enabling the generation of long-term, physically plausible motions in 3D environments. However, existing methods often fall short of achieving physics realism and supporting diverse types of interactions. To address these challenges, this paper introduces a unified Human-Object Interaction framework that provides unified control over interactions with static scenes and dynamic objects using language commands. The interactions between human and object parts can always be described as the continuous stable Relative Movement Dynamics (RMD) between human and object parts. By leveraging the world knowledge and scene perception capabilities of Vision-Language Models (VLMs), we translate language commands into RMD diagrams, which are used to guide goal-conditioned reinforcement learning for sequential interaction with objects. Our framework supports long-horizon interactions among dynamic, articulated, and static objects. To support the training and evaluation of our framework, we present a new dataset named Interplay, which includes multi-round task plans generated by VLMs, covering both static and dynamic HOI tasks. Extensive experiments demonstrate that our proposed framework can effectively handle a wide range of HOI tasks, showcasing its ability to maintain long-term, multi-round transitions. For more details, please refer to our project webpage: https://rmd-hoi.github.io/.
☆ PS-EIP: Robust Photometric Stereo Based on Event Interval Profile CVPR2025
Recently, the energy-efficient photometric stereo method using an event camera has been proposed to recover surface normals from events triggered by changes in logarithmic Lambertian reflections under a moving directional light source. However, EventPS treats each event interval independently, making it sensitive to noise, shadows, and non-Lambertian reflections. This paper proposes Photometric Stereo based on Event Interval Profile (PS-EIP), a robust method that recovers pixelwise surface normals from a time-series profile of event intervals. By exploiting the continuity of the profile and introducing an outlier detection method based on profile shape, our approach enhances robustness against outliers from shadows and specular reflections. Experiments using real event data from 3D-printed objects demonstrate that PS-EIP significantly improves robustness to outliers compared to EventPS's deep-learning variant, EventPS-FCN, without relying on deep learning.
comment: CVPR2025
☆ GranQ: Granular Zero-Shot Quantization with Unified Layer-Channel Awareness
Zero-shot quantization (ZSQ) enables neural network compression without training data, which is crucial in restricted data access environments. However, existing ZSQ methods suffer from significant activation loss in low-bit environments owing to their coarse-grained scaling strategy. To address this issue, we propose GranQ, a novel ZSQ approach that leverages layer-channel awareness to minimize the quantization error. Unlike conventional layer- or channel-wise quantization, GranQ dynamically adjusts quantization granularity by considering both layer- and channel-level activation distributions. This enables fine-grained quantization while minimizing activation distortion. Additionally, we introduce vectorized activation quantization, which enables efficient parallel computation and reduces computational overhead while preserving accuracy. GranQ achieves superior performance compared with those of state-of-the-art ZSQ methods that employ quantization-aware training. With these findings, we anticipate that GranQ will inspire novel research directions beyond conventional ZSQ approaches focused on data generation and model training.
☆ SPMTrack: Spatio-Temporal Parameter-Efficient Fine-Tuning with Mixture of Experts for Scalable Visual Tracking CVPR2025
Most state-of-the-art trackers adopt one-stream paradigm, using a single Vision Transformer for joint feature extraction and relation modeling of template and search region images. However, relation modeling between different image patches exhibits significant variations. For instance, background regions dominated by target-irrelevant information require reduced attention allocation, while foreground, particularly boundary areas, need to be be emphasized. A single model may not effectively handle all kinds of relation modeling simultaneously. In this paper, we propose a novel tracker called SPMTrack based on mixture-of-experts tailored for visual tracking task (TMoE), combining the capability of multiple experts to handle diverse relation modeling more flexibly. Benefiting from TMoE, we extend relation modeling from image pairs to spatio-temporal context, further improving tracking accuracy with minimal increase in model parameters. Moreover, we employ TMoE as a parameter-efficient fine-tuning method, substantially reducing trainable parameters, which enables us to train SPMTrack of varying scales efficiently and preserve the generalization ability of pretrained models to achieve superior performance. We conduct experiments on seven datasets, and experimental results demonstrate that our method significantly outperforms current state-of-the-art trackers. The source code is available at https://github.com/WenRuiCai/SPMTrack.
comment: Accepted by CVPR2025
☆ Coeff-Tuning: A Graph Filter Subspace View for Tuning Attention-Based Large Models
Transformer-based large pre-trained models have shown remarkable generalization ability, and various parameter-efficient fine-tuning (PEFT) methods have been proposed to customize these models on downstream tasks with minimal computational and memory budgets. Previous PEFT methods are primarily designed from a tensor-decomposition perspective that tries to effectively tune the linear transformation by finding the smallest subset of parameters to train. Our study adopts an orthogonal view by representing the attention operation as a graph convolution and formulating the multi-head attention maps as a convolutional filter subspace, with each attention map as a subspace element. In this paper, we propose to tune the large pre-trained transformers by learning a small set of combination coefficients that construct a more expressive filter subspace from the original multi-head attention maps. We show analytically and experimentally that the tuned filter subspace can effectively expand the feature space of the multi-head attention and further enhance the capacity of transformers. We further stabilize the fine-tuning with a residual parameterization of the tunable subspace coefficients, and enhance the generalization with a regularization design by directly applying dropout on the tunable coefficient during training. The tunable coefficients take a tiny number of parameters and can be combined with previous PEFT methods in a plug-and-play manner. Extensive experiments show that our approach achieves superior performances than PEFT baselines with neglectable additional parameters.
☆ Mitigating Cache Noise in Test-Time Adaptation for Large Vision-Language Models ICME 2025
Test-time adaptation (TTA) of visual language models has recently attracted significant attention as a solution to the performance degradation caused by distribution shifts in downstream tasks. However, existing cache-based TTA methods have certain limitations. They mainly rely on the accuracy of cached feature labels, and the presence of noisy pseudo-labels can cause these features to deviate from their true distribution. This makes cache retrieval methods based on similarity matching highly sensitive to outliers or extreme samples. Moreover, current methods lack effective mechanisms to model class distributions, which limits their ability to fully exploit the potential of cached information. To address these challenges, we introduce a comprehensive and reliable caching mechanism and propose a novel zero-shot TTA method called ``Cache, Residual, Gaussian" (CRG). This method not only employs learnable residual parameters to better align positive and negative visual prototypes with text prototypes, thereby optimizing the quality of cached features, but also incorporates Gaussian Discriminant Analysis (GDA) to dynamically model intra-class feature distributions, further mitigating the impact of noisy features. Experimental results on 13 benchmarks demonstrate that CRG outperforms state-of-the-art TTA methods, showcasing exceptional robustness and adaptability.
comment: Accepted by ICME 2025 and ICLR 2025 Workshop on Foundation Models in the Wild
☆ TensoFlow: Tensorial Flow-based Sampler for Inverse Rendering CVPR 2025
Inverse rendering aims to recover scene geometry, material properties, and lighting from multi-view images. Given the complexity of light-surface interactions, importance sampling is essential for the evaluation of the rendering equation, as it reduces variance and enhances the efficiency of Monte Carlo sampling. Existing inverse rendering methods typically use pre-defined non-learnable importance samplers in prior manually, struggling to effectively match the spatially and directionally varied integrand and resulting in high variance and suboptimal performance. To address this limitation, we propose the concept of learning a spatially and directionally aware importance sampler for the rendering equation to accurately and flexibly capture the unconstrained complexity of a typical scene. We further formulate TensoFlow, a generic approach for sampler learning in inverse rendering, enabling to closely match the integrand of the rendering equation spatially and directionally. Concretely, our sampler is parameterized by normalizing flows, allowing both directional sampling of incident light and probability density function (PDF) inference. To capture the characteristics of the sampler spatially, we learn a tensorial representation of the scene space, which imposes spatial conditions, together with reflected direction, leading to spatially and directionally aware sampling distributions. Our model can be optimized by minimizing the difference between the integrand and our normalizing flow. Extensive experiments validate the superiority of TensoFlow over prior alternatives on both synthetic and real-world benchmarks.
comment: CVPR 2025. Code: https://github.com/fudan-zvg/tensoflow
☆ Towards Training-free Anomaly Detection with Vision and Language Foundation Models CVPR 2025
Anomaly detection is valuable for real-world applications, such as industrial quality inspection. However, most approaches focus on detecting local structural anomalies while neglecting compositional anomalies incorporating logical constraints. In this paper, we introduce LogSAD, a novel multi-modal framework that requires no training for both Logical and Structural Anomaly Detection. First, we propose a match-of-thought architecture that employs advanced large multi-modal models (i.e. GPT-4V) to generate matching proposals, formulating interests and compositional rules of thought for anomaly detection. Second, we elaborate on multi-granularity anomaly detection, consisting of patch tokens, sets of interests, and composition matching with vision and language foundation models. Subsequently, we present a calibration module to align anomaly scores from different detectors, followed by integration strategies for the final decision. Consequently, our approach addresses both logical and structural anomaly detection within a unified framework and achieves state-of-the-art results without the need for training, even when compared to supervised approaches, highlighting its robustness and effectiveness. Code is available at https://github.com/zhang0jhon/LogSAD.
comment: Accepted to CVPR 2025
☆ Plug-and-Play Interpretable Responsible Text-to-Image Generation via Dual-Space Multi-facet Concept Control
Ethical issues around text-to-image (T2I) models demand a comprehensive control over the generative content. Existing techniques addressing these issues for responsible T2I models aim for the generated content to be fair and safe (non-violent/explicit). However, these methods remain bounded to handling the facets of responsibility concepts individually, while also lacking in interpretability. Moreover, they often require alteration to the original model, which compromises the model performance. In this work, we propose a unique technique to enable responsible T2I generation by simultaneously accounting for an extensive range of concepts for fair and safe content generation in a scalable manner. The key idea is to distill the target T2I pipeline with an external plug-and-play mechanism that learns an interpretable composite responsible space for the desired concepts, conditioned on the target T2I pipeline. We use knowledge distillation and concept whitening to enable this. At inference, the learned space is utilized to modulate the generative content. A typical T2I pipeline presents two plug-in points for our approach, namely; the text embedding space and the diffusion model latent space. We develop modules for both points and show the effectiveness of our approach with a range of strong results.
☆ Diff-Palm: Realistic Palmprint Generation with Polynomial Creases and Intra-Class Variation Controllable Diffusion Models CVPR2025
Palmprint recognition is significantly limited by the lack of large-scale publicly available datasets. Previous methods have adopted B\'ezier curves to simulate the palm creases, which then serve as input for conditional GANs to generate realistic palmprints. However, without employing real data fine-tuning, the performance of the recognition model trained on these synthetic datasets would drastically decline, indicating a large gap between generated and real palmprints. This is primarily due to the utilization of an inaccurate palm crease representation and challenges in balancing intra-class variation with identity consistency. To address this, we introduce a polynomial-based palm crease representation that provides a new palm crease generation mechanism more closely aligned with the real distribution. We also propose the palm creases conditioned diffusion model with a novel intra-class variation control method. By applying our proposed $K$-step noise-sharing sampling, we are able to synthesize palmprint datasets with large intra-class variation and high identity consistency. Experimental results show that, for the first time, recognition models trained solely on our synthetic datasets, without any fine-tuning, outperform those trained on real datasets. Furthermore, our approach achieves superior recognition performance as the number of generated identities increases.
comment: Accepted by CVPR2025
☆ Image-to-Text for Medical Reports Using Adaptive Co-Attention and Triple-LSTM Module
Medical report generation requires specialized expertise that general large models often fail to accurately capture. Moreover, the inherent repetition and similarity in medical data make it difficult for models to extract meaningful features, resulting in a tendency to overfit. So in this paper, we propose a multimodal model, Co-Attention Triple-LSTM Network (CA-TriNet), a deep learning model that combines transformer architectures with a Multi-LSTM network. Its Co-Attention module synergistically links a vision transformer with a text transformer to better differentiate medical images with similarities, augmented by an adaptive weight operator to catch and amplify image labels with minor similarities. Furthermore, its Triple-LSTM module refines generated sentences using targeted image objects. Extensive evaluations over three public datasets have demonstrated that CA-TriNet outperforms state-of-the-art models in terms of comprehensive ability, even pre-trained large language models on some metrics.
☆ LGPS: A Lightweight GAN-Based Approach for Polyp Segmentation in Colonoscopy Images
Colorectal cancer (CRC) is a major global cause of cancer-related deaths, with early polyp detection and removal during colonoscopy being crucial for prevention. While deep learning methods have shown promise in polyp segmentation, challenges such as high computational costs, difficulty in segmenting small or low-contrast polyps, and limited generalizability across datasets persist. To address these issues, we propose LGPS, a lightweight GAN-based framework for polyp segmentation. LGPS incorporates three key innovations: (1) a MobileNetV2 backbone enhanced with modified residual blocks and Squeeze-and-Excitation (ResE) modules for efficient feature extraction; (2) Convolutional Conditional Random Fields (ConvCRF) for precise boundary refinement; and (3) a hybrid loss function combining Binary Cross-Entropy, Weighted IoU Loss, and Dice Loss to address class imbalance and enhance segmentation accuracy. LGPS is validated on five benchmark datasets and compared with state-of-the-art(SOTA) methods. On the largest and challenging PolypGen test dataset, LGPS achieves a Dice of 0.7299 and an IoU of 0.7867, outperformed all SOTA works and demonstrating robust generalization. With only 1.07 million parameters, LGPS is 17 times smaller than the smallest existing model, making it highly suitable for real-time clinical applications. Its lightweight design and strong performance underscore its potential for improving early CRC diagnosis. Code is available at https://github.com/Falmi/LGPS/.
comment: 10 pages, 56 Figures
☆ CO-SPY: Combining Semantic and Pixel Features to Detect Synthetic Images by AI
With the rapid advancement of generative AI, it is now possible to synthesize high-quality images in a few seconds. Despite the power of these technologies, they raise significant concerns regarding misuse. Current efforts to distinguish between real and AI-generated images may lack generalization, being effective for only certain types of generative models and susceptible to post-processing techniques like JPEG compression. To overcome these limitations, we propose a novel framework, Co-Spy, that first enhances existing semantic features (e.g., the number of fingers in a hand) and artifact features (e.g., pixel value differences), and then adaptively integrates them to achieve more general and robust synthetic image detection. Additionally, we create Co-Spy-Bench, a comprehensive dataset comprising 5 real image datasets and 22 state-of-the-art generative models, including the latest models like FLUX. We also collect 50k synthetic images in the wild from the Internet to enable evaluation in a more practical setting. Our extensive evaluations demonstrate that our detector outperforms existing methods under identical training conditions, achieving an average accuracy improvement of approximately 11% to 34%. The code is available at https://github.com/Megum1/Co-Spy.
☆ Voxel-based Point Cloud Geometry Compression with Space-to-Channel Context
Voxel-based methods are among the most efficient for point cloud geometry compression, particularly with dense point clouds. However, they face limitations due to a restricted receptive field, especially when handling high-bit depth point clouds. To overcome this issue, we introduce a stage-wise Space-to-Channel (S2C) context model for both dense point clouds and low-level sparse point clouds. This model utilizes a channel-wise autoregressive strategy to effectively integrate neighborhood information at a coarse resolution. For high-level sparse point clouds, we further propose a level-wise S2C context model that addresses resolution limitations by incorporating Geometry Residual Coding (GRC) for consistent-resolution cross-level prediction. Additionally, we use the spherical coordinate system for its compact representation and enhance our GRC approach with a Residual Probability Approximation (RPA) module, which features a large kernel size. Experimental results show that our S2C context model not only achieves bit savings while maintaining or improving reconstruction quality but also reduces computational complexity compared to state-of-the-art voxel-based compression methods.
☆ TrackID3x3: A Dataset and Algorithm for Multi-Player Tracking with Identification and Pose Estimation in 3x3 Basketball Full-court Videos
Multi-object tracking, player identification, and pose estimation are fundamental components of sports analytics, essential for analyzing player movements, performance, and tactical strategies. However, existing datasets and methodologies primarily target mainstream team sports such as soccer and conventional 5-on-5 basketball, often overlooking scenarios involving fixed-camera setups commonly used at amateur levels, less mainstream sports, or datasets that explicitly incorporate pose annotations. In this paper, we propose the TrackID3x3 dataset, the first publicly available comprehensive dataset specifically designed for multi-player tracking, player identification, and pose estimation in 3x3 basketball scenarios. The dataset comprises three distinct subsets (Indoor fixed-camera, Outdoor fixed-camera, and Drone camera footage), capturing diverse full-court camera perspectives and environments. We also introduce the Track-ID task, a simplified variant of the game state reconstruction task that excludes field detection and focuses exclusively on fixed-camera scenarios. To evaluate performance, we propose a baseline algorithm called Track-ID algorithm, tailored to assess tracking and identification quality. Furthermore, our benchmark experiments, utilizing recent multi-object tracking algorithms (e.g., BoT-SORT-ReID) and top-down pose estimation methods (HRNet, RTMPose, and SwinPose), demonstrate robust results and highlight remaining challenges. Our dataset and evaluation benchmarks provide a solid foundation for advancing automated analytics in 3x3 basketball. Dataset and code will be available at https://github.com/open-starlab/TrackID3x3.
☆ TopV: Compatible Token Pruning with Inference Time Optimization for Fast and Low-Memory Multimodal Vision Language Model CVPR 2025
Vision-Language Models (VLMs) demand substantial computational resources during inference, largely due to the extensive visual input tokens for representing visual information. Previous studies have noted that visual tokens tend to receive less attention than text tokens, suggesting their lower importance during inference and potential for pruning. However, their methods encounter several challenges: reliance on greedy heuristic criteria for token importance and incompatibility with FlashAttention and KV cache. To address these issues, we introduce \textbf{TopV}, a compatible \textbf{TO}ken \textbf{P}runing with inference Time Optimization for fast and low-memory \textbf{V}LM, achieving efficient pruning without additional training or fine-tuning. Instead of relying on attention scores, we formulate token pruning as an optimization problem, accurately identifying important visual tokens while remaining compatible with FlashAttention. Additionally, since we only perform this pruning once during the prefilling stage, it effectively reduces KV cache size. Our optimization framework incorporates a visual-aware cost function considering factors such as Feature Similarity, Relative Spatial Distance, and Absolute Central Distance, to measure the importance of each source visual token, enabling effective pruning of low-importance tokens. Extensive experiments demonstrate that our method outperforms previous token pruning methods, validating the effectiveness and efficiency of our approach.
comment: Accepted by CVPR 2025
☆ GI-SLAM: Gaussian-Inertial SLAM
3D Gaussian Splatting (3DGS) has recently emerged as a powerful representation of geometry and appearance for dense Simultaneous Localization and Mapping (SLAM). Through rapid, differentiable rasterization of 3D Gaussians, many 3DGS SLAM methods achieve near real-time rendering and accelerated training. However, these methods largely overlook inertial data, witch is a critical piece of information collected from the inertial measurement unit (IMU). In this paper, we present GI-SLAM, a novel gaussian-inertial SLAM system which consists of an IMU-enhanced camera tracking module and a realistic 3D Gaussian-based scene representation for mapping. Our method introduces an IMU loss that seamlessly integrates into the deep learning framework underpinning 3D Gaussian Splatting SLAM, effectively enhancing the accuracy, robustness and efficiency of camera tracking. Moreover, our SLAM system supports a wide range of sensor configurations, including monocular, stereo, and RGBD cameras, both with and without IMU integration. Our method achieves competitive performance compared with existing state-of-the-art real-time methods on the EuRoC and TUM-RGBD datasets.
comment: 10 pages, 2 figures, 5 tables
☆ Enhancing Dataset Distillation via Non-Critical Region Refinement CVPR 2025
Dataset distillation has become a popular method for compressing large datasets into smaller, more efficient representations while preserving critical information for model training. Data features are broadly categorized into two types: instance-specific features, which capture unique, fine-grained details of individual examples, and class-general features, which represent shared, broad patterns across a class. However, previous approaches often struggle to balance these features-some focus solely on class-general patterns, neglecting finer instance details, while others prioritize instance-specific features, overlooking the shared characteristics essential for class-level understanding. In this paper, we introduce the Non-Critical Region Refinement Dataset Distillation (NRR-DD) method, which preserves instance-specific details and fine-grained regions in synthetic data while enriching non-critical regions with class-general information. This approach enables models to leverage all pixel information, capturing both feature types and enhancing overall performance. Additionally, we present Distance-Based Representative (DBR) knowledge transfer, which eliminates the need for soft labels in training by relying on the distance between synthetic data predictions and one-hot encoded labels. Experimental results show that NRR-DD achieves state-of-the-art performance on both small- and large-scale datasets. Furthermore, by storing only two distances per instance, our method delivers comparable results across various settings. The code is available at https://github.com/tmtuan1307/NRR-DD.
comment: Accepted at CVPR 2025
☆ Surface-Aware Distilled 3D Semantic Features
Many 3D tasks such as pose alignment, animation, motion transfer, and 3D reconstruction rely on establishing correspondences between 3D shapes. This challenge has recently been approached by matching of semantic features from pre-trained vision models. However, despite their power, these features struggle to differentiate instances of the same semantic class such as "left hand" versus "right hand" which leads to substantial mapping errors. To solve this, we learn a surface-aware embedding space that is robust to these ambiguities. Importantly, our approach is self-supervised and requires only a small number of unpaired training meshes to infer features for new 3D shapes at test time. We achieve this by introducing a contrastive loss that preserves the semantic content of the features distilled from foundational models while disambiguating features located far apart on the shape's surface. We observe superior performance in correspondence matching benchmarks and enable downstream applications including in-part segmentation, pose alignment, and motion transfer. The project site is available at https://lukas.uzolas.com/SurfaceAware3DFeaturesSite.
☆ ZECO: ZeroFusion Guided 3D MRI Conditional Generation
Medical image segmentation is crucial for enhancing diagnostic accuracy and treatment planning in Magnetic Resonance Imaging (MRI). However, acquiring precise lesion masks for segmentation model training demands specialized expertise and significant time investment, leading to a small dataset scale in clinical practice. In this paper, we present ZECO, a ZeroFusion guided 3D MRI conditional generation framework that extracts, compresses, and generates high-fidelity MRI images with corresponding 3D segmentation masks to mitigate data scarcity. To effectively capture inter-slice relationships within volumes, we introduce a Spatial Transformation Module that encodes MRI images into a compact latent space for the diffusion process. Moving beyond unconditional generation, our novel ZeroFusion method progressively maps 3D masks to MRI images in latent space, enabling robust training on limited datasets while avoiding overfitting. ZECO outperforms state-of-the-art models in both quantitative and qualitative evaluations on Brain MRI datasets across various modalities, showcasing its exceptional capability in synthesizing high-quality MRI images conditioned on segmentation masks.
comment: Project page: \url{https://brack-wang.github.io/ZECO_web/}; Github Code: \url{https://github.com/Brack-Wang/ZECO}
☆ LGPS: A Lightweight GAN-Based Approach for Polyp Segmentation in Colonoscopy Images
Colorectal cancer (CRC) is a major global cause of cancer-related deaths, with early polyp detection and removal during colonoscopy being crucial for prevention. While deep learning methods have shown promise in polyp segmentation, challenges such as high computational costs, difficulty in segmenting small or low-contrast polyps, and limited generalizability across datasets persist. To address these issues, we propose LGPS, a lightweight GAN-based framework for polyp segmentation. LGPS incorporates three key innovations: (1) a MobileNetV2 backbone enhanced with modified residual blocks and Squeeze-and-Excitation (ResE) modules for efficient feature extraction; (2) Convolutional Conditional Random Fields (ConvCRF) for precise boundary refinement; and (3) a hybrid loss function combining Binary Cross-Entropy, Weighted IoU Loss, and Dice Loss to address class imbalance and enhance segmentation accuracy. LGPS is validated on five benchmark datasets and compared with state-of-the-art(SOTA) methods. On the largest and challenging PolypGen test dataset, LGPS achieves a Dice of 0.7299 and an IoU of 0.7867, outperformed all SOTA works and demonstrating robust generalization. With only 1.07 million parameters, LGPS is 17 times smaller than the smallest existing model, making it highly suitable for real-time clinical applications. Its lightweight design and strong performance underscore its potential for improving early CRC diagnosis. Code is available at https://github.com/Falmi/LGPS/.
comment: 10 pages, 6 Figures
♻ ☆ GroundCap: A Visually Grounded Image Captioning Dataset
Current image captioning systems lack the ability to link descriptive text to specific visual elements, making their outputs difficult to verify. While recent approaches offer some grounding capabilities, they cannot track object identities across multiple references or ground both actions and objects simultaneously. We propose a novel ID-based grounding system that enables consistent object reference tracking and action-object linking, and present GroundCap, a dataset containing 52,016 images from 77 movies, with 344 human-annotated and 52,016 automatically generated captions. Each caption is grounded on detected objects (132 classes) and actions (51 classes) using a tag system that maintains object identity while linking actions to the corresponding objects. Our approach features persistent object IDs for reference tracking, explicit action-object linking, and segmentation of background elements through K-means clustering. We propose gMETEOR, a metric combining caption quality with grounding accuracy, and establish baseline performance by fine-tuning Pixtral-12B. Human evaluation demonstrates our approach's effectiveness in producing verifiable descriptions with coherent object references.
comment: 37 pages
♻ ☆ Zero-Shot Styled Text Image Generation, but Make It Autoregressive CVPR2025
Styled Handwritten Text Generation (HTG) has recently received attention from the computer vision and document analysis communities, which have developed several solutions, either GAN- or diffusion-based, that achieved promising results. Nonetheless, these strategies fail to generalize to novel styles and have technical constraints, particularly in terms of maximum output length and training efficiency. To overcome these limitations, in this work, we propose a novel framework for text image generation, dubbed Emuru. Our approach leverages a powerful text image representation model (a variational autoencoder) combined with an autoregressive Transformer. Our approach enables the generation of styled text images conditioned on textual content and style examples, such as specific fonts or handwriting styles. We train our model solely on a diverse, synthetic dataset of English text rendered in over 100,000 typewritten and calligraphy fonts, which gives it the capability to reproduce unseen styles (both fonts and users' handwriting) in zero-shot. To the best of our knowledge, Emuru is the first autoregressive model for HTG, and the first designed specifically for generalization to novel styles. Moreover, our model generates images without background artifacts, which are easier to use for downstream applications. Extensive evaluation on both typewritten and handwritten, any-length text image generation scenarios demonstrates the effectiveness of our approach.
comment: Accepted at CVPR2025
♻ ☆ Relative Pose Estimation through Affine Corrections of Monocular Depth Priors CVPR 2025
Monocular depth estimation (MDE) models have undergone significant advancements over recent years. Many MDE models aim to predict affine-invariant relative depth from monocular images, while recent developments in large-scale training and vision foundation models enable reasonable estimation of metric (absolute) depth. However, effectively leveraging these predictions for geometric vision tasks, in particular relative pose estimation, remains relatively under explored. While depths provide rich constraints for cross-view image alignment, the intrinsic noise and ambiguity from the monocular depth priors present practical challenges to improving upon classic keypoint-based solutions. In this paper, we develop three solvers for relative pose estimation that explicitly account for independent affine (scale and shift) ambiguities, covering both calibrated and uncalibrated conditions. We further propose a hybrid estimation pipeline that combines our proposed solvers with classic point-based solvers and epipolar constraints. We find that the affine correction modeling is beneficial to not only the relative depth priors but also, surprisingly, the "metric" ones. Results across multiple datasets demonstrate large improvements of our approach over classic keypoint-based baselines and PnP-based solutions, under both calibrated and uncalibrated setups. We also show that our method improves consistently with different feature matchers and MDE models, and can further benefit from very recent advances on both modules. Code is available at https://github.com/MarkYu98/madpose.
comment: CVPR 2025
♻ ☆ Complementary Advantages: Exploiting Cross-Field Frequency Correlation for NIR-Assisted Image Denoising
Existing single-image denoising algorithms often struggle to restore details when dealing with complex noisy images. The introduction of near-infrared (NIR) images offers new possibilities for RGB image denoising. However, due to the inconsistency between NIR and RGB images, the existing works still struggle to balance the contributions of two fields in the process of image fusion. In response to this, in this paper, we develop a cross-field Frequency Correlation Exploiting Network (FCENet) for NIR-assisted image denoising. We first propose the frequency correlation prior based on an in-depth statistical frequency analysis of NIR-RGB image pairs. The prior reveals the complementary correlation of NIR and RGB images in the frequency domain. Leveraging frequency correlation prior, we then establish a frequency learning framework composed of Frequency Dynamic Selection Mechanism (FDSM) and Frequency Exhaustive Fusion Mechanism (FEFM). FDSM dynamically selects complementary information from NIR and RGB images in the frequency domain, and FEFM strengthens the control of common and differential features during the fusion process of NIR and RGB features. Extensive experiments on simulated and real data validate that the proposed method outperforms other state-of-the-art methods. The code will be released at https://github.com/yuchenwang815/FCENet.
♻ ☆ Merging synthetic and real embryo data for advanced AI predictions
Accurate embryo morphology assessment is essential in assisted reproductive technology for selecting the most viable embryo. Artificial intelligence has the potential to enhance this process. However, the limited availability of embryo data presents challenges for training deep learning models. To address this, we trained two generative models using two datasets-one we created and made publicly available, and one existing public dataset-to generate synthetic embryo images at various cell stages, including 2-cell, 4-cell, 8-cell, morula, and blastocyst. These were combined with real images to train classification models for embryo cell stage prediction. Our results demonstrate that incorporating synthetic images alongside real data improved classification performance, with the model achieving 97% accuracy compared to 94.5% when trained solely on real data. This trend remained consistent when tested on an external Blastocyst dataset from a different clinic. Notably, even when trained exclusively on synthetic data and tested on real data, the model achieved a high accuracy of 92%. Furthermore, combining synthetic data from both generative models yielded better classification results than using data from a single generative model. Four embryologists evaluated the fidelity of the synthetic images through a Turing test, during which they annotated inaccuracies and offered feedback. The analysis showed the diffusion model outperformed the generative adversarial network, deceiving embryologists 66.6% versus 25.3% and achieving lower Frechet inception distance scores.
♻ ☆ RaCFormer: Towards High-Quality 3D Object Detection via Query-based Radar-Camera Fusion CVPR 2025
We propose Radar-Camera fusion transformer (RaCFormer) to boost the accuracy of 3D object detection by the following insight. The Radar-Camera fusion in outdoor 3D scene perception is capped by the image-to-BEV transformation--if the depth of pixels is not accurately estimated, the naive combination of BEV features actually integrates unaligned visual content. To avoid this problem, we propose a query-based framework that enables adaptive sampling of instance-relevant features from both the bird's-eye view (BEV) and the original image view. Furthermore, we enhance system performance by two key designs: optimizing query initialization and strengthening the representational capacity of BEV. For the former, we introduce an adaptive circular distribution in polar coordinates to refine the initialization of object queries, allowing for a distance-based adjustment of query density. For the latter, we initially incorporate a radar-guided depth head to refine the transformation from image view to BEV. Subsequently, we focus on leveraging the Doppler effect of radar and introduce an implicit dynamic catcher to capture the temporal elements within the BEV. Extensive experiments on nuScenes and View-of-Delft (VoD) datasets validate the merits of our design. Remarkably, our method achieves superior results of 64.9% mAP and 70.2% NDS on nuScenes. RaCFormer also secures the state-of-the-art performance on the VoD dataset. Code is available at https://github.com/cxmomo/RaCFormer.
comment: Accepted to CVPR 2025
♻ ☆ MotionMap: Representing Multimodality in Human Pose Forecasting CVPR 2025
Human pose forecasting is inherently multimodal since multiple futures exist for an observed pose sequence. However, evaluating multimodality is challenging since the task is ill-posed. Therefore, we first propose an alternative paradigm to make the task well-posed. Next, while state-of-the-art methods predict multimodality, this requires oversampling a large volume of predictions. This raises key questions: (1) Can we capture multimodality by efficiently sampling a smaller number of predictions? (2) Subsequently, which of the predicted futures is more likely for an observed pose sequence? We address these questions with MotionMap, a simple yet effective heatmap based representation for multimodality. We extend heatmaps to represent a spatial distribution over the space of all possible motions, where different local maxima correspond to different forecasts for a given observation. MotionMap can capture a variable number of modes per observation and provide confidence measures for different modes. Further, MotionMap allows us to introduce the notion of uncertainty and controllability over the forecasted pose sequence. Finally, MotionMap captures rare modes that are non-trivial to evaluate yet critical for safety. We support our claims through multiple qualitative and quantitative experiments using popular 3D human pose datasets: Human3.6M and AMASS, highlighting the strengths and limitations of our proposed method. Project Page: https://vita-epfl.github.io/MotionMap
comment: CVPR 2025. We propose a new representation for learning multimodality in human pose forecasting which does not depend on generative models
♻ ☆ Dora: Sampling and Benchmarking for 3D Shape Variational Auto-Encoders CVPR 2025
Recent 3D content generation pipelines commonly employ Variational Autoencoders (VAEs) to encode shapes into compact latent representations for diffusion-based generation. However, the widely adopted uniform point sampling strategy in Shape VAE training often leads to a significant loss of geometric details, limiting the quality of shape reconstruction and downstream generation tasks. We present Dora-VAE, a novel approach that enhances VAE reconstruction through our proposed sharp edge sampling strategy and a dual cross-attention mechanism. By identifying and prioritizing regions with high geometric complexity during training, our method significantly improves the preservation of fine-grained shape features. Such sampling strategy and the dual attention mechanism enable the VAE to focus on crucial geometric details that are typically missed by uniform sampling approaches. To systematically evaluate VAE reconstruction quality, we additionally propose Dora-bench, a benchmark that quantifies shape complexity through the density of sharp edges, introducing a new metric focused on reconstruction accuracy at these salient geometric features. Extensive experiments on the Dora-bench demonstrate that Dora-VAE achieves comparable reconstruction quality to the state-of-the-art dense XCube-VAE while requiring a latent space at least 8$\times$ smaller (1,280 vs. > 10,000 codes).
comment: Accepted by CVPR 2025. Project page: https://aruichen.github.io/Dora/
♻ ☆ Visual Position Prompt for MLLM based Visual Grounding
Although Multimodal Large Language Models (MLLMs) excel at various image-related tasks, they encounter challenges in precisely aligning coordinates with spatial information within images, particularly in position-aware tasks such as visual grounding. This limitation arises from two key factors. First, MLLMs lack explicit spatial references, making it difficult to associate textual descriptions with precise image locations. Second, their feature extraction processes prioritize global context over fine-grained spatial details, leading to weak localization capability. To address this issue, we introduce VPP-LLaVA, an MLLM equipped with Visual Position Prompt (VPP) to improve its grounding capability. VPP-LLaVA integrates two complementary mechanisms. The global VPP overlays learnable, axis-like embeddings onto the input image to provide structured spatial cues. The local VPP focuses on fine-grained localization by incorporating position-aware queries, which suggests probable object locations. We also introduce a VPP-SFT dataset with 0.6M samples, consolidating high-quality visual grounding data into a compact format for efficient model training. Training on this dataset with VPP enhances the model's performance, achieving state-of-the-art results on standard grounding benchmarks despite using fewer training samples compared to other MLLMs like MiniGPT-v2, which rely on much larger datasets ($\sim$21M samples). The code and VPP-SFT dataset will be available at https://github.com/WayneTomas/VPP-LLaVA upon acceptance.
♻ ☆ PRISM: Privacy-Preserving Improved Stochastic Masking for Federated Generative Models
Despite recent advancements in federated learning (FL), the integration of generative models into FL has been limited due to challenges such as high communication costs and unstable training in heterogeneous data environments. To address these issues, we propose PRISM, a FL framework tailored for generative models that ensures (i) stable performance in heterogeneous data distributions and (ii) resource efficiency in terms of communication cost and final model size. The key of our method is to search for an optimal stochastic binary mask for a random network rather than updating the model weights, identifying a sparse subnetwork with high generative performance; i.e., a ``strong lottery ticket''. By communicating binary masks in a stochastic manner, PRISM minimizes communication overhead. This approach, combined with the utilization of maximum mean discrepancy (MMD) loss and a mask-aware dynamic moving average aggregation method (MADA) on the server side, facilitates stable and strong generative capabilities by mitigating local divergence in FL scenarios. Moreover, thanks to its sparsifying characteristic, PRISM yields a lightweight model without extra pruning or quantization, making it ideal for environments such as edge devices. Experiments on MNIST, FMNIST, CelebA, and CIFAR10 demonstrate that PRISM outperforms existing methods, while maintaining privacy with minimal communication costs. PRISM is the first to successfully generate images under challenging non-IID and privacy-preserving FL environments on complex datasets, where previous methods have struggled.
♻ ☆ STEVE: A Step Verification Pipeline for Computer-use Agent Training
Developing AI agents to autonomously manipulate graphical user interfaces is a long challenging task. Recent advances in data scaling law inspire us to train computer-use agents with a scaled instruction set, yet using behavior cloning to train agents still requires immense high-quality trajectories. To meet the scalability need, we designed STEVE, a step verification pipeline for computer-use agent training. First, we establish a large instruction set for computer-use agents and collect trajectory data with some suboptimal agents. GPT-4o is used to verify the correctness of each step in the trajectories based on the screens before and after the action execution, assigning each step with a binary label. Last, we adopt the Kahneman and Tversky Optimization to optimize the agent from the binary stepwise labels. Extensive experiments manifest that our agent outperforms supervised finetuning by leveraging both positive and negative actions within a trajectory. Also, STEVE enables us to train a 7B vision-language model as a computer-use agent, achieving leading performance in the challenging live desktop environment WinAgentArena with great efficiency at a reduced cost. Code and data: https://github.com/FanbinLu/STEVE.
♻ ☆ Positive2Negative: Breaking the Information-Lossy Barrier in Self-Supervised Single Image Denoising
Image denoising enhances image quality, serving as a foundational technique across various computational photography applications. The obstacle to clean image acquisition in real scenarios necessitates the development of self-supervised image denoising methods only depending on noisy images, especially a single noisy image. Existing self-supervised image denoising paradigms (Noise2Noise and Noise2Void) rely heavily on information-lossy operations, such as downsampling and masking, culminating in low quality denoising performance. In this paper, we propose a novel self-supervised single image denoising paradigm, Positive2Negative, to break the information-lossy barrier. Our paradigm involves two key steps: Renoised Data Construction (RDC) and Denoised Consistency Supervision (DCS). RDC renoises the predicted denoised image by the predicted noise to construct multiple noisy images, preserving all the information of the original image. DCS ensures consistency across the multiple denoised images, supervising the network to learn robust denoising. Our Positive2Negative paradigm achieves state-of-the-art performance in self-supervised single image denoising with significant speed improvements. The code is released to the public at https://github.com/Li-Tong-621/P2N.
comment: 8 figures, 5 tables, 11 pages
♻ ☆ Interleaved Scene Graphs for Interleaved Text-and-Image Generation Assessment ICLR 2025
Many real-world user queries (e.g. "How do to make egg fried rice?") could benefit from systems capable of generating responses with both textual steps with accompanying images, similar to a cookbook. Models designed to generate interleaved text and images face challenges in ensuring consistency within and across these modalities. To address these challenges, we present ISG, a comprehensive evaluation framework for interleaved text-and-image generation. ISG leverages a scene graph structure to capture relationships between text and image blocks, evaluating responses on four levels of granularity: holistic, structural, block-level, and image-specific. This multi-tiered evaluation allows for a nuanced assessment of consistency, coherence, and accuracy, and provides interpretable question-answer feedback. In conjunction with ISG, we introduce a benchmark, ISG-Bench, encompassing 1,150 samples across 8 categories and 21 subcategories. This benchmark dataset includes complex language-vision dependencies and golden answers to evaluate models effectively on vision-centric tasks such as style transfer, a challenging area for current models. Using ISG-Bench, we demonstrate that recent unified vision-language models perform poorly on generating interleaved content. While compositional approaches that combine separate language and image models show a 111% improvement over unified models at the holistic level, their performance remains suboptimal at both block and image levels. To facilitate future work, we develop ISG-Agent, a baseline agent employing a "plan-execute-refine" pipeline to invoke tools, achieving a 122% performance improvement.
comment: Accepted by ICLR 2025 as Spotlight. Project homepage: https://interleave-eval.github.io/
♻ ☆ Generative Omnimatte: Learning to Decompose Video into Layers CVPR 2025
Given a video and a set of input object masks, an omnimatte method aims to decompose the video into semantically meaningful layers containing individual objects along with their associated effects, such as shadows and reflections. Existing omnimatte methods assume a static background or accurate pose and depth estimation and produce poor decompositions when these assumptions are violated. Furthermore, due to the lack of generative prior on natural videos, existing methods cannot complete dynamic occluded regions. We present a novel generative layered video decomposition framework to address the omnimatte problem. Our method does not assume a stationary scene or require camera pose or depth information and produces clean, complete layers, including convincing completions of occluded dynamic regions. Our core idea is to train a video diffusion model to identify and remove scene effects caused by a specific object. We show that this model can be finetuned from an existing video inpainting model with a small, carefully curated dataset, and demonstrate high-quality decompositions and editing results for a wide range of casually captured videos containing soft shadows, glossy reflections, splashing water, and more.
comment: CVPR 2025. Project page: https://gen-omnimatte.github.io/
♻ ☆ Learned, uncertainty-driven adaptive acquisition for photon-efficient scanning microscopy
Scanning microscopy systems, such as confocal and multiphoton microscopy, are powerful imaging tools for probing deep into biological tissue. However, scanning systems have an inherent trade-off between acquisition time, field of view, phototoxicity, and image quality, often resulting in noisy measurements when fast, large field of view, and/or gentle imaging is needed. Deep learning could be used to denoise noisy microscopy measurements, but these algorithms can be prone to hallucination, which can be disastrous for medical and scientific applications. We propose a method to simultaneously denoise and predict pixel-wise uncertainty for scanning microscopy systems, improving algorithm trustworthiness and providing statistical guarantees for deep learning predictions. Furthermore, we propose to leverage this learned, pixel-wise uncertainty to drive an adaptive acquisition technique that rescans only the most uncertain regions of a sample, saving time and reducing the total light dose to the sample. We demonstrate our method on experimental confocal and multiphoton microscopy systems, showing that our uncertainty maps can pinpoint hallucinations in the deep learned predictions. Finally, with our adaptive acquisition technique, we demonstrate up to 16X reduction in acquisition time and total light dose while successfully recovering fine features in the sample and reducing hallucinations. We are the first to demonstrate distribution-free uncertainty quantification for a denoising task with real experimental data and the first to propose adaptive acquisition based on reconstruction uncertainty.
♻ ☆ Lightweight Embedded FPGA Deployment of Learned Image Compression with Knowledge Distillation and Hybrid Quantization
Learnable Image Compression (LIC) has shown the potential to outperform standardized video codecs in RD efficiency, prompting the research for hardware-friendly implementations. Most existing LIC hardware implementations prioritize latency to RD-efficiency and through an extensive exploration of the hardware design space. We present a novel design paradigm where the burden of tuning the design for a specific hardware platform is shifted towards model dimensioning and without compromising on RD-efficiency. First, we design a framework for distilling a leaner student LIC model from a reference teacher: by tuning a single model hyperparameters, we can meet the constraints of different hardware platforms without a complex hardware design exploration. Second, we propose a hardware-friendly implementation of the Generalized Divisive Normalization - GDN activation that preserves RD efficiency even post parameter quantization. Third, we design a pipelined FPGA configuration which takes full advantage of available FPGA resources by leveraging parallel processing and optimizing resource allocation. Our experiments with a state of the art LIC model show that we outperform all existing FPGA implementations while performing very close to the original model.
comment: Submitted to IEEE Transactions on Circuits and Systems for Video Technology in March 2025
♻ ☆ Mani-GS: Gaussian Splatting Manipulation with Triangular Mesh CVPR 2025
Neural 3D representations such as Neural Radiance Fields (NeRF), excel at producing photo-realistic rendering results but lack the flexibility for manipulation and editing which is crucial for content creation. Previous works have attempted to address this issue by deforming a NeRF in canonical space or manipulating the radiance field based on an explicit mesh. However, manipulating NeRF is not highly controllable and requires a long training and inference time. With the emergence of 3D Gaussian Splatting (3DGS), extremely high-fidelity novel view synthesis can be achieved using an explicit point-based 3D representation with much faster training and rendering speed. However, there is still a lack of effective means to manipulate 3DGS freely while maintaining rendering quality. In this work, we aim to tackle the challenge of achieving manipulable photo-realistic rendering. We propose to utilize a triangular mesh to manipulate 3DGS directly with self-adaptation. This approach reduces the need to design various algorithms for different types of Gaussian manipulation. By utilizing a triangle shape-aware Gaussian binding and adapting method, we can achieve 3DGS manipulation and preserve high-fidelity rendering after manipulation. Our approach is capable of handling large deformations, local manipulations, and soft body simulations while keeping high-quality rendering. Furthermore, we demonstrate that our method is also effective with inaccurate meshes extracted from 3DGS. Experiments conducted demonstrate the effectiveness of our method and its superiority over baseline approaches.
comment: CVPR 2025. Project page here: https://gaoxiangjun.github.io/mani_gs/
♻ ☆ MaRS: A Fast Sampler for Mean Reverting Diffusion based on ODE and SDE Solvers ICLR 2025
In applications of diffusion models, controllable generation is of practical significance, but is also challenging. Current methods for controllable generation primarily focus on modifying the score function of diffusion models, while Mean Reverting (MR) Diffusion directly modifies the structure of the stochastic differential equation (SDE), making the incorporation of image conditions simpler and more natural. However, current training-free fast samplers are not directly applicable to MR Diffusion. And thus MR Diffusion requires hundreds of NFEs (number of function evaluations) to obtain high-quality samples. In this paper, we propose a new algorithm named MaRS (MR Sampler) to reduce the sampling NFEs of MR Diffusion. We solve the reverse-time SDE and the probability flow ordinary differential equation (PF-ODE) associated with MR Diffusion, and derive semi-analytical solutions. The solutions consist of an analytical function and an integral parameterized by a neural network. Based on this solution, we can generate high-quality samples in fewer steps. Our approach does not require training and supports all mainstream parameterizations, including noise prediction, data prediction and velocity prediction. Extensive experiments demonstrate that MR Sampler maintains high sampling quality with a speedup of 10 to 20 times across ten different image restoration tasks. Our algorithm accelerates the sampling procedure of MR Diffusion, making it more practical in controllable generation.
comment: Accepted by ICLR 2025
♻ ☆ Beyond Training: Dynamic Token Merging for Zero-Shot Video Understanding
Recent advancements in multimodal large language models (MLLMs) have opened new avenues for video understanding. However, achieving high fidelity in zero-shot video tasks remains challenging. Traditional video processing methods rely heavily on fine-tuning to capture nuanced spatial-temporal details, which incurs significant data and computation costs. In contrast, training-free approaches, though efficient, often lack robustness in preserving context-rich features across complex video content. To this end, we propose DYTO, a novel dynamic token merging framework for zero-shot video understanding that adaptively optimizes token efficiency while preserving crucial scene details. DYTO integrates a hierarchical frame selection and a bipartite token merging strategy to dynamically cluster key frames and selectively compress token sequences, striking a balance between computational efficiency with semantic richness. Extensive experiments across multiple benchmarks demonstrate the effectiveness of DYTO, achieving superior performance compared to both fine-tuned and training-free methods and setting a new state-of-the-art for zero-shot video understanding.
comment: Code is available at https://github.com/Jam1ezhang/DYTO
♻ ☆ SAT-HMR: Real-Time Multi-Person 3D Mesh Estimation via Scale-Adaptive Tokens
We propose a one-stage framework for real-time multi-person 3D human mesh estimation from a single RGB image. While current one-stage methods, which follow a DETR-style pipeline, achieve state-of-the-art (SOTA) performance with high-resolution inputs, we observe that this particularly benefits the estimation of individuals in smaller scales of the image (e.g., those far from the camera), but at the cost of significantly increased computation overhead. To address this, we introduce scale-adaptive tokens that are dynamically adjusted based on the relative scale of each individual in the image within the DETR framework. Specifically, individuals in smaller scales are processed at higher resolutions, larger ones at lower resolutions, and background regions are further distilled. These scale-adaptive tokens more efficiently encode the image features, facilitating subsequent decoding to regress the human mesh, while allowing the model to allocate computational resources more effectively and focus on more challenging cases. Experiments show that our method preserves the accuracy benefits of high-resolution processing while substantially reducing computational cost, achieving real-time inference with performance comparable to SOTA methods.
comment: 18 pages, 12 figures
♻ ☆ RankCLIP: Ranking-Consistent Language-Image Pretraining
Self-supervised contrastive learning models, such as CLIP, have set new benchmarks for vision-language models in many downstream tasks. However, their dependency on rigid one-to-one mappings overlooks the complex and often multifaceted relationships between and within texts and images. To this end, we introduce RankCLIP, a novel pre-training method that extends beyond the rigid one-to-one matching framework of CLIP and its variants. By extending the traditional pair-wise loss to list-wise, and leveraging both in-modal and cross-modal ranking consistency, RankCLIP improves the alignment process, enabling it to capture the nuanced many-to-many relationships between and within each modality. Through comprehensive experiments, we demonstrate the effectiveness of RankCLIP in various downstream tasks, notably achieving significant gains in zero-shot classifications over state-of-the-art methods, underscoring the importance of this enhanced learning process.
comment: Code and model checkpoints are available at https://github.com/Jam1ezhang/RankCLIP
♻ ☆ DUNE: Distilling a Universal Encoder from Heterogeneous 2D and 3D Teachers CVPR-2025
Recent multi-teacher distillation methods have unified the encoders of multiple foundation models into a single encoder, achieving competitive performance on core vision tasks like classification, segmentation, and depth estimation. This led us to ask: Could similar success be achieved when the pool of teachers also includes vision models specialized in diverse tasks across both 2D and 3D perception? In this paper, we define and investigate the problem of heterogeneous teacher distillation, or co-distillation, a challenging multi-teacher distillation scenario where teacher models vary significantly in both (a) their design objectives and (b) the data they were trained on. We explore data-sharing strategies and teacher-specific encoding, and introduce DUNE, a single encoder excelling in 2D vision, 3D understanding, and 3D human perception. Our model achieves performance comparable to that of its larger teachers, sometimes even outperforming them, on their respective tasks. Notably, DUNE surpasses MASt3R in Map-free Visual Relocalization with a much smaller encoder.
comment: Accepted to CVPR-2025. Project page: https://europe.naverlabs.com/dune
♻ ☆ Understanding Model Calibration -- A gentle introduction and visual exploration of calibration and the expected calibration error (ECE)
To be considered reliable, a model must be calibrated so that its confidence in each decision closely reflects its true outcome. In this blogpost we'll take a look at the most commonly used definition for calibration and then dive into a frequently used evaluation measure for model calibration. We'll then cover some of the drawbacks of this measure and how these surfaced the need for additional notions of calibration, which require their own new evaluation measures. This post is not intended to be an in-depth dissection of all works on calibration, nor does it focus on how to calibrate models. Instead, it is meant to provide a gentle introduction to the different notions and their evaluation measures as well as to re-highlight some issues with a measure that is still widely used to evaluate calibration.
♻ ☆ GAEA: A Geolocation Aware Conversational Model
Image geolocalization, in which, traditionally, an AI model predicts the precise GPS coordinates of an image is a challenging task with many downstream applications. However, the user cannot utilize the model to further their knowledge other than the GPS coordinate; the model lacks an understanding of the location and the conversational ability to communicate with the user. In recent days, with tremendous progress of large multimodal models (LMMs) -- proprietary and open-source -- researchers have attempted to geolocalize images via LMMs. However, the issues remain unaddressed; beyond general tasks, for more specialized downstream tasks, one of which is geolocalization, LMMs struggle. In this work, we propose to solve this problem by introducing a conversational model GAEA that can provide information regarding the location of an image, as required by a user. No large-scale dataset enabling the training of such a model exists. Thus we propose GAEA-1.6M, a comprehensive dataset with 800K images and around 1.6M question-answer pairs constructed by leveraging OpenStreetMap (OSM) attributes and geographical context clues. For quantitative evaluation, we propose a diverse benchmark, GAEA-Bench, comprising 4K image-text pairs to evaluate conversational capabilities equipped with diverse question types. We consider 11 state-of-the-art open-source and proprietary LMMs and demonstrate that GAEA significantly outperforms the best open-source model, LLaVA-OneVision by 25.69% and the best proprietary model, GPT-4o by 8.28%. Our dataset, model and codes are available.
comment: The dataset and code used in this submission is available at: https://ucf-crcv.github.io/GAEA/
♻ ☆ Multiple Object Tracking as ID Prediction CVPR 2025
Multi-Object Tracking (MOT) has been a long-standing challenge in video understanding. A natural and intuitive approach is to split this task into two parts: object detection and association. Most mainstream methods employ meticulously crafted heuristic techniques to maintain trajectory information and compute cost matrices for object matching. Although these methods can achieve notable tracking performance, they often require a series of elaborate handcrafted modifications while facing complicated scenarios. We believe that manually assumed priors limit the method's adaptability and flexibility in learning optimal tracking capabilities from domain-specific data. Therefore, we introduce a new perspective that treats Multiple Object Tracking as an in-context ID Prediction task, transforming the aforementioned object association into an end-to-end trainable task. Based on this, we propose a simple yet effective method termed MOTIP. Given a set of trajectories carried with ID information, MOTIP directly decodes the ID labels for current detections to accomplish the association process. Without using tailored or sophisticated architectures, our method achieves state-of-the-art results across multiple benchmarks by solely leveraging object-level features as tracking cues. The simplicity and impressive results of MOTIP leave substantial room for future advancements, thereby making it a promising baseline for subsequent research. Our code and checkpoints are released at https://github.com/MCG-NJU/MOTIP.
comment: Accepted by CVPR 2025
♻ ☆ Kalib: Easy Hand-Eye Calibration with Reference Point Tracking
Hand-eye calibration aims to estimate the transformation between a camera and a robot. Traditional methods rely on fiducial markers, which require considerable manual effort and precise setup. Recent advances in deep learning have introduced markerless techniques but come with more prerequisites, such as retraining networks for each robot, and accessing accurate mesh models for data generation. In this paper, we propose Kalib, an automatic and easy-to-setup hand-eye calibration method that leverages the generalizability of visual foundation models to overcome these challenges. It features only two basic prerequisites, the robot's kinematic chain and a predefined reference point on the robot. During calibration, the reference point is tracked in the camera space. Its corresponding 3D coordinates in the robot coordinate can be inferred by forward kinematics. Then, a PnP solver directly estimates the transformation between the camera and the robot without training new networks or accessing mesh models. Evaluations in simulated and real-world benchmarks show that Kalib achieves good accuracy with a lower manual workload compared with recent baseline methods. We also demonstrate its application in multiple real-world settings with various robot arms and grippers. Kalib's user-friendly design and minimal setup requirements make it a possible solution for continuous operation in unstructured environments.
comment: The code, data, and supplementary materials are available at https://sites.google.com/view/hand-eye-kalib
♻ ☆ Towards Universal Soccer Video Understanding CVPR 2025
As a globally celebrated sport, soccer has attracted widespread interest from fans all over the world. This paper aims to develop a comprehensive multi-modal framework for soccer video understanding. Specifically, we make the following contributions in this paper: (i) we introduce SoccerReplay-1988, the largest multi-modal soccer dataset to date, featuring videos and detailed annotations from 1,988 complete matches, with an automated annotation pipeline; (ii) we present an advanced soccer-specific visual encoder, MatchVision, which leverages spatiotemporal information across soccer videos and excels in various downstream tasks; (iii) we conduct extensive experiments and ablation studies on event classification, commentary generation, and multi-view foul recognition. MatchVision demonstrates state-of-the-art performance on all of them, substantially outperforming existing models, which highlights the superiority of our proposed data and model. We believe that this work will offer a standard paradigm for sports understanding research.
comment: CVPR 2025; Project Page: https://jyrao.github.io/UniSoccer/
♻ ☆ CAGE: Unsupervised Visual Composition and Animation for Controllable Video Generation AAAI2025
The field of video generation has expanded significantly in recent years, with controllable and compositional video generation garnering considerable interest. Most methods rely on leveraging annotations such as text, objects' bounding boxes, and motion cues, which require substantial human effort and thus limit their scalability. In contrast, we address the challenge of controllable and compositional video generation without any annotations by introducing a novel unsupervised approach. Our model is trained from scratch on a dataset of unannotated videos. At inference time, it can compose plausible novel scenes and animate objects by placing object parts at the desired locations in space and time. The core innovation of our method lies in the unified control format and the training process, where video generation is conditioned on a randomly selected subset of pre-trained self-supervised local features. This conditioning compels the model to learn how to inpaint the missing information in the video both spatially and temporally, thereby learning the inherent compositionality of a scene and the dynamics of moving objects. The abstraction level and the imposed invariance of the conditioning input to minor visual perturbations enable control over object motion by simply using the same features at all the desired future locations. We call our model CAGE, which stands for visual Composition and Animation for video GEneration. We conduct extensive experiments to validate the effectiveness of CAGE across various scenarios, demonstrating its capability to accurately follow the control and to generate high-quality videos that exhibit coherent scene composition and realistic animation.
comment: Published at AAAI2025; Project website: https://araachie.github.io/cage
♻ ☆ Re-HOLD: Video Hand Object Interaction Reenactment via adaptive Layout-instructed Diffusion Model CVPR 2025
Current digital human studies focusing on lip-syncing and body movement are no longer sufficient to meet the growing industrial demand, while human video generation techniques that support interacting with real-world environments (e.g., objects) have not been well investigated. Despite human hand synthesis already being an intricate problem, generating objects in contact with hands and their interactions presents an even more challenging task, especially when the objects exhibit obvious variations in size and shape. To cope with these issues, we present a novel video Reenactment framework focusing on Human-Object Interaction (HOI) via an adaptive Layout-instructed Diffusion model (Re-HOLD). Our key insight is to employ specialized layout representation for hands and objects, respectively. Such representations enable effective disentanglement of hand modeling and object adaptation to diverse motion sequences. To further improve the generation quality of HOI, we have designed an interactive textural enhancement module for both hands and objects by introducing two independent memory banks. We also propose a layout-adjusting strategy for the cross-object reenactment scenario to adaptively adjust unreasonable layouts caused by diverse object sizes during inference. Comprehensive qualitative and quantitative evaluations demonstrate that our proposed framework significantly outperforms existing methods. Project page: https://fyycs.github.io/Re-HOLD.
comment: Accepted to CVPR 2025
♻ ☆ IncEventGS: Pose-Free Gaussian Splatting from a Single Event Camera
Implicit neural representation and explicit 3D Gaussian Splatting (3D-GS) for novel view synthesis have achieved remarkable progress with frame-based camera (e.g. RGB and RGB-D cameras) recently. Compared to frame-based camera, a novel type of bio-inspired visual sensor, i.e. event camera, has demonstrated advantages in high temporal resolution, high dynamic range, low power consumption and low latency. Due to its unique asynchronous and irregular data capturing process, limited work has been proposed to apply neural representation or 3D Gaussian splatting for an event camera. In this work, we present IncEventGS, an incremental 3D Gaussian Splatting reconstruction algorithm with a single event camera. To recover the 3D scene representation incrementally, we exploit the tracking and mapping paradigm of conventional SLAM pipelines for IncEventGS. Given the incoming event stream, the tracker firstly estimates an initial camera motion based on prior reconstructed 3D-GS scene representation. The mapper then jointly refines both the 3D scene representation and camera motion based on the previously estimated motion trajectory from the tracker. The experimental results demonstrate that IncEventGS delivers superior performance compared to prior NeRF-based methods and other related baselines, even we do not have the ground-truth camera poses. Furthermore, our method can also deliver better performance compared to state-of-the-art event visual odometry methods in terms of camera motion estimation. Code is publicly available at: https://github.com/wu-cvgl/IncEventGS.
comment: Code Page: https://github.com/wu-cvgl/IncEventGS
♻ ☆ CholecTrack20: A Multi-Perspective Tracking Dataset for Surgical Tools CVPR 2025
Tool tracking in surgical videos is essential for advancing computer-assisted interventions, such as skill assessment, safety zone estimation, and human-machine collaboration. However, the lack of context-rich datasets limits AI applications in this field. Existing datasets rely on overly generic tracking formalizations that fail to capture surgical-specific dynamics, such as tools moving out of the camera's view or exiting the body. This results in less clinically relevant trajectories and a lack of flexibility for real-world surgical applications. Methods trained on these datasets often struggle with visual challenges such as smoke, reflection, and bleeding, further exposing the limitations of current approaches. We introduce CholecTrack20, a specialized dataset for multi-class, multi-tool tracking in surgical procedures. It redefines tracking formalization with three perspectives: (i) intraoperative, (ii) intracorporeal, and (iii) visibility, enabling adaptable and clinically meaningful tool trajectories. The dataset comprises 20 full-length surgical videos, annotated at 1 fps, yielding over 35K frames and 65K labeled tool instances. Annotations include spatial location, category, identity, operator, phase, and scene visual challenge. Benchmarking state-of-the-art methods on CholecTrack20 reveals significant performance gaps, with current approaches (< 45\% HOTA) failing to meet the accuracy required for clinical translation. These findings motivate the need for advanced and intuitive tracking algorithms and establish CholecTrack20 as a foundation for developing robust AI-driven surgical assistance systems.
comment: Surgical tool tracking dataset paper, 11 pages, 10 figures, 3 tables, CVPR 2025
♻ ☆ ETAP: Event-based Tracking of Any Point
Tracking any point (TAP) recently shifted the motion estimation paradigm from focusing on individual salient points with local templates to tracking arbitrary points with global image contexts. However, while research has mostly focused on driving the accuracy of models in nominal settings, addressing scenarios with difficult lighting conditions and high-speed motions remains out of reach due to the limitations of the sensor. This work addresses this challenge with the first event camera-based TAP method. It leverages the high temporal resolution and high dynamic range of event cameras for robust high-speed tracking, and the global contexts in TAP methods to handle asynchronous and sparse event measurements. We further extend the TAP framework to handle event feature variations induced by motion -- thereby addressing an open challenge in purely event-based tracking -- with a novel feature-alignment loss which ensures the learning of motion-robust features. Our method is trained with data from a new data generation pipeline and systematically ablated across all design decisions. Our method shows strong cross-dataset generalization and performs 136% better on the average Jaccard metric than the baselines. Moreover, on an established feature tracking benchmark, it achieves a 20% improvement over the previous best event-only method and even surpasses the previous best events-and-frames method by 4.1%. Our code is available at https://github.com/tub-rip/ETAP
comment: 17 pages, 15 figures, 8 tables. Project page: https://github.com/tub-rip/ETAP
♻ ☆ NeSS-ST: Detecting Good and Stable Keypoints with a Neural Stability Score and the Shi-Tomasi Detector ICCV 2023
Learning a feature point detector presents a challenge both due to the ambiguity of the definition of a keypoint and, correspondingly, the need for specially prepared ground truth labels for such points. In our work, we address both of these issues by utilizing a combination of a hand-crafted Shi-Tomasi detector, a specially designed metric that assesses the quality of keypoints, the stability score (SS), and a neural network. We build on the principled and localized keypoints provided by the Shi-Tomasi detector and learn the neural network to select good feature points via the stability score. The neural network incorporates the knowledge from the training targets in the form of the neural stability score (NeSS). Therefore, our method is named NeSS-ST since it combines the Shi-Tomasi detector and the properties of the neural stability score. It only requires sets of images for training without dataset pre-labeling or the need for reconstructed correspondence labels. We evaluate NeSS-ST on HPatches, ScanNet, MegaDepth and IMC-PT demonstrating state-of-the-art performance and good generalization on downstream tasks.
comment: Camera-ready version of ICCV 2023 paper
♻ ☆ CarPlanner: Consistent Auto-regressive Trajectory Planning for Large-scale Reinforcement Learning in Autonomous Driving CVPR 2025
Trajectory planning is vital for autonomous driving, ensuring safe and efficient navigation in complex environments. While recent learning-based methods, particularly reinforcement learning (RL), have shown promise in specific scenarios, RL planners struggle with training inefficiencies and managing large-scale, real-world driving scenarios. In this paper, we introduce \textbf{CarPlanner}, a \textbf{C}onsistent \textbf{a}uto-\textbf{r}egressive \textbf{Planner} that uses RL to generate multi-modal trajectories. The auto-regressive structure enables efficient large-scale RL training, while the incorporation of consistency ensures stable policy learning by maintaining coherent temporal consistency across time steps. Moreover, CarPlanner employs a generation-selection framework with an expert-guided reward function and an invariant-view module, simplifying RL training and enhancing policy performance. Extensive analysis demonstrates that our proposed RL framework effectively addresses the challenges of training efficiency and performance enhancement, positioning CarPlanner as a promising solution for trajectory planning in autonomous driving. To the best of our knowledge, we are the first to demonstrate that the RL-based planner can surpass both IL- and rule-based state-of-the-arts (SOTAs) on the challenging large-scale real-world dataset nuPlan. Our proposed CarPlanner surpasses RL-, IL-, and rule-based SOTA approaches within this demanding dataset.
comment: CVPR 2025
♻ ☆ Revisiting Monocular 3D Object Detection with Depth Thickness Field
Monocular 3D object detection is challenging due to the lack of accurate depth. However, existing depth-assisted solutions still exhibit inferior performance, whose reason is universally acknowledged as the unsatisfactory accuracy of monocular depth estimation models. In this paper, we revisit monocular 3D object detection from the depth perspective and formulate an additional issue as the limited 3D structure-aware capability of existing depth representations (e.g., depth one-hot encoding or depth distribution). To address this issue, we introduce a novel Depth Thickness Field approach to embed clear 3D structures of the scenes. Specifically, we present MonoDTF, a scene-to-instance depth-adapted network for monocular 3D object detection. The framework mainly comprises a Scene-Level Depth Retargeting (SDR) module and an Instance-Level Spatial Refinement (ISR) module. The former retargets traditional depth representations to the proposed depth thickness field, incorporating the scene-level perception of 3D structures. The latter refines the voxel space with the guidance of instances, enhancing the 3D instance-aware capability of the depth thickness field and thus improving detection accuracy. Extensive experiments on the KITTI and Waymo datasets demonstrate our superiority to existing state-of-the-art (SoTA) methods and the universality when equipped with different depth estimation models. The code will be available.
♻ ☆ OnlineAnySeg: Online Zero-Shot 3D Segmentation by Visual Foundation Model Guided 2D Mask Merging
Online 3D open-vocabulary segmentation of a progressively reconstructed scene is both a critical and challenging task for embodied applications. With the success of visual foundation models (VFMs) in the image domain, leveraging 2D priors to address 3D online segmentation has become a prominent research focus. Since segmentation results provided by 2D priors often require spatial consistency to be lifted into final 3D segmentation, an efficient method for identifying spatial overlap among 2D masks is essential - yet existing methods rarely achieve this in real time, mainly limiting its use to offline approaches. To address this, we propose an efficient method that lifts 2D masks generated by VFMs into a unified 3D instance using a hashing technique. By employing voxel hashing for efficient 3D scene querying, our approach reduces the time complexity of costly spatial overlap queries from $O(n^2)$ to $O(n)$. Accurate spatial associations further enable 3D merging of 2D masks through simple similarity-based filtering in a zero-shot manner, making our approach more robust to incomplete and noisy data. Evaluated on the ScanNet and SceneNN benchmarks, our approach achieves state-of-the-art performance in online, open-vocabulary 3D instance segmentation with leading efficiency.
♻ ☆ MoCha-Stereo: Motif Channel Attention Network for Stereo Matching CVPR 2024
Learning-based stereo matching techniques have made significant progress. However, existing methods inevitably lose geometrical structure information during the feature channel generation process, resulting in edge detail mismatches. In this paper, the Motif Cha}nnel Attention Stereo Matching Network (MoCha-Stereo) is designed to address this problem. We provide the Motif Channel Correlation Volume (MCCV) to determine more accurate edge matching costs. MCCV is achieved by projecting motif channels, which capture common geometric structures in feature channels, onto feature maps and cost volumes. In addition, edge variations in %potential feature channels of the reconstruction error map also affect details matching, we propose the Reconstruction Error Motif Penalty (REMP) module to further refine the full-resolution disparity estimation. REMP integrates the frequency information of typical channel features from the reconstruction error. MoCha-Stereo ranks 1st on the KITTI-2015 and KITTI-2012 Reflective leaderboards. Our structure also shows excellent performance in Multi-View Stereo. Code is avaliable at https://github.com/ZYangChen/MoCha-Stereo.
comment: Accepted to CVPR 2024
♻ ☆ Believing is Seeing: Unobserved Object Detection using Generative Models
Can objects that are not visible in an image -- but are in the vicinity of the camera -- be detected? This study introduces the novel tasks of 2D, 2.5D and 3D unobserved object detection for predicting the location of nearby objects that are occluded or lie outside the image frame. We adapt several state-of-the-art pre-trained generative models to address this task, including 2D and 3D diffusion models and vision-language models, and show that they can be used to infer the presence of objects that are not directly observed. To benchmark this task, we propose a suite of metrics that capture different aspects of performance. Our empirical evaluation on indoor scenes from the RealEstate10k and NYU Depth v2 datasets demonstrate results that motivate the use of generative models for the unobserved object detection task.
comment: IEEE/CVF Computer Vision and Pattern Recognition 2025; 22 pages
♻ ☆ Image Over Text: Transforming Formula Recognition Evaluation with Character Detection Matching CVPR2025
Formula recognition presents significant challenges due to the complicated structure and varied notation of mathematical expressions. Despite continuous advancements in formula recognition models, the evaluation metrics employed by these models, such as BLEU and Edit Distance, still exhibit notable limitations. They overlook the fact that the same formula has diverse representations and is highly sensitive to the distribution of training data, thereby causing unfairness in formula recognition evaluation. To this end, we propose a Character Detection Matching (CDM) metric, ensuring the evaluation objectivity by designing an image-level rather than a LaTeX-level metric score. Specifically, CDM renders both the model-predicted LaTeX and the ground-truth LaTeX formulas into image-formatted formulas, then employs visual feature extraction and localization techniques for precise character-level matching, incorporating spatial position information. Such a spatially-aware and character-matching method offers a more accurate and equitable evaluation compared with previous BLEU and Edit Distance metrics that rely solely on text-based character matching. Experimentally, we evaluated various formula recognition models using CDM, BLEU, and ExpRate metrics. Their results demonstrate that the CDM aligns more closely with human evaluation standards and provides a fairer comparison across different models by eliminating discrepancies caused by diverse formula representations. Code is available at https://github.com/opendatalab/UniMERNet/tree/main/cdm.
comment: Accepted by CVPR2025
♻ ☆ Deep Polycuboid Fitting for Compact 3D Representation of Indoor Scenes 3DV 2025
This paper presents a novel framework for compactly representing a 3D indoor scene using a set of polycuboids through a deep learning-based fitting method. Indoor scenes mainly consist of man-made objects, such as furniture, which often exhibit rectilinear geometry. This property allows indoor scenes to be represented using combinations of polycuboids, providing a compact representation that benefits downstream applications like furniture rearrangement. Our framework takes a noisy point cloud as input and first detects six types of cuboid faces using a transformer network. Then, a graph neural network is used to validate the spatial relationships of the detected faces to form potential polycuboids. Finally, each polycuboid instance is reconstructed by forming a set of boxes based on the aggregated face labels. To train our networks, we introduce a synthetic dataset encompassing a diverse range of cuboid and polycuboid shapes that reflect the characteristics of indoor scenes. Our framework generalizes well to real-world indoor scene datasets, including Replica, ScanNet, and scenes captured with an iPhone. The versatility of our method is demonstrated through practical applications, such as virtual room tours and scene editing.
comment: Accepted to 3DV 2025. For project page, see this https://waldstein94.github.io/deep-polycuboid-fitting/
♻ ☆ Learning Physics From Video: Unsupervised Physical Parameter Estimation for Continuous Dynamical Systems
Extracting physical dynamical system parameters from recorded observations is key in natural science. Current methods for automatic parameter estimation from video train supervised deep networks on large datasets. Such datasets require labels, which are difficult to acquire. While some unsupervised techniques--which depend on frame prediction--exist, they suffer from long training times, initialization instabilities, only consider motion-based dynamical systems, and are evaluated mainly on synthetic data. In this work, we propose an unsupervised method to estimate the physical parameters of known, continuous governing equations from single videos suitable for different dynamical systems beyond motion and robust to initialization. Moreover, we remove the need for frame prediction by implementing a KL-divergence-based loss function in the latent space, which avoids convergence to trivial solutions and reduces model size and compute. We first evaluate our model on synthetic data, as commonly done. After which, we take the field closer to reality by recording Delfys75: our own real-world dataset of 75 videos for five different types of dynamical systems to evaluate our method and others. Our method compares favorably to others. %, yet, and real-world video datasets and demonstrate improved parameter estimation accuracy compared to existing methods. Code and data are available online:https://github.com/Alejandro-neuro/Learning_physics_from_video.
♻ ☆ Learnable Infinite Taylor Gaussian for Dynamic View Rendering
Capturing the temporal evolution of Gaussian properties such as position, rotation, and scale is a challenging task due to the vast number of time-varying parameters and the limited photometric data available, which generally results in convergence issues, making it difficult to find an optimal solution. While feeding all inputs into an end-to-end neural network can effectively model complex temporal dynamics, this approach lacks explicit supervision and struggles to generate high-quality transformation fields. On the other hand, using time-conditioned polynomial functions to model Gaussian trajectories and orientations provides a more explicit and interpretable solution, but requires significant handcrafted effort and lacks generalizability across diverse scenes. To overcome these limitations, this paper introduces a novel approach based on a learnable infinite Taylor Formula to model the temporal evolution of Gaussians. This method offers both the flexibility of an implicit network-based approach and the interpretability of explicit polynomial functions, allowing for more robust and generalizable modeling of Gaussian dynamics across various dynamic scenes. Extensive experiments on dynamic novel view rendering tasks are conducted on public datasets, demonstrating that the proposed method achieves state-of-the-art performance in this domain. More information is available on our project page(https://ellisonking.github.io/TaylorGaussian).
♻ ☆ SAMKD: Spatial-aware Adaptive Masking Knowledge Distillation for Object Detection
Most of recent attention-guided feature masking distillation methods perform knowledge transfer via global teacher attention maps without delving into fine-grained clues. Instead, performing distillation at finer granularity is conducive to uncovering local details supplementary to global knowledge transfer and reconstructing comprehensive student features. In this study, we propose a Spatial-aware Adaptive Masking Knowledge Distillation (SAMKD) framework for accurate object detection. Different from previous feature distillation methods which mainly perform single-scale feature masking, we develop spatially hierarchical feature masking distillation scheme, such that the object-aware locality is encoded during coarse-to-fine distillation process for improved feature reconstruction. In addition, our spatial-aware feature distillation strategy is combined with a masking logit distillation scheme in which region-specific feature difference between teacher and student networks is utilized to adaptively guide the distillation process. Thus, it can help the student model to better learn from the teacher counterpart with improved knowledge transfer and reduced gap. Extensive experiments for detection task demonstrate the superiority of our method. For example, when FCOS is used as teacher detector with ResNet101 backbone, our method improves the student network from 35.3\% to 38.8\% mAP, outperforming state-of-the-art distillation methods including MGD, FreeKD and DMKD.
♻ ☆ GaussTR: Foundation Model-Aligned Gaussian Transformer for Self-Supervised 3D Spatial Understanding CVPR 2025
3D Semantic Occupancy Prediction is fundamental for spatial understanding, yet existing approaches face challenges in scalability and generalization due to their reliance on extensive labeled data and computationally intensive voxel-wise representations. In this paper, we introduce GaussTR, a novel Gaussian-based Transformer framework that unifies sparse 3D modeling with foundation model alignment through Gaussian representations to advance 3D spatial understanding. GaussTR predicts sparse sets of Gaussians in a feed-forward manner to represent 3D scenes. By splatting the Gaussians into 2D views and aligning the rendered features with foundation models, GaussTR facilitates self-supervised 3D representation learning and enables open-vocabulary semantic occupancy prediction without requiring explicit annotations. Empirical experiments on the Occ3D-nuScenes dataset demonstrate GaussTR's state-of-the-art zero-shot performance of 12.27 mIoU, along with a 40% reduction in training time. These results highlight the efficacy of GaussTR for scalable and holistic 3D spatial understanding, with promising implications in autonomous driving and embodied agents. The code is available at https://github.com/hustvl/GaussTR.
comment: CVPR 2025
♻ ☆ Personalization Toolkit: Training Free Personalization of Large Vision Language Models
Large Vision Language Models (LVLMs) have significant potential to provide personalized assistance by adapting to the unique needs and preferences of individual users. The personalization of LVLMs has emerged as a field that focuses on customizing models to recognize specific object instances and provide tailored responses. However, current methodologies depend on time-consuming test-time training for each user and object, which proves to be impractical. This paper introduces a novel, training-free approach to LVLM personalization by leveraging pre-trained vision foundation models to extract distinct features, retrieval-augmented generation (RAG) techniques to recognize instances in the visual input, and visual prompting methods. Our model-agnostic vision toolkit enables flexible and efficient personalization without the need for extensive retraining. We demonstrate state-of-the-art results, surpassing conventional training-based approaches, and set a new benchmark for LVLM personalization.
♻ ☆ BI-RADS prediction of mammographic masses using uncertainty information extracted from a Bayesian Deep Learning model
The BI_RADS score is a probabilistic reporting tool used by radiologists to express the level of uncertainty in predicting breast cancer based on some morphological features in mammography images. There is a significant variability in describing masses which sometimes leads to BI_RADS misclassification. Using a BI_RADS prediction system is required to support the final radiologist decisions. In this study, the uncertainty information extracted by a Bayesian deep learning model is utilized to predict the BI_RADS score. The investigation results based on the pathology information demonstrate that the f1-scores of the predictions of the radiologist are 42.86%, 48.33% and 48.28%, meanwhile, the f1-scores of the model performance are 73.33%, 59.60% and 59.26% in the BI_RADS 2, 3 and 5 dataset samples, respectively. Also, the model can distinguish malignant from benign samples in the BI_RADS 0 category of the used dataset with an accuracy of 75.86% and correctly identify all malignant samples as BI_RADS 5. The Grad-CAM visualization shows the model pays attention to the morphological features of the lesions. Therefore, this study shows the uncertainty-aware Bayesian Deep Learning model can report his uncertainty about the malignancy of a lesion based on morphological features, like a radiologist.
♻ ☆ From Objects to Events: Unlocking Complex Visual Understanding in Object Detectors via LLM-guided Symbolic Reasoning
Our key innovation lies in bridging the semantic gap between object detection and event understanding without requiring expensive task-specific training. The proposed plug-and-play framework interfaces with any open-vocabulary detector while extending their inherent capabilities across architectures. At its core, our approach combines (i) a symbolic regression mechanism exploring relationship patterns among detected entities and (ii) a LLM-guided strategically guiding the search toward meaningful expressions. These discovered symbolic rules transform low-level visual perception into interpretable event understanding, providing a transparent reasoning path from objects to events with strong transferability across domains.We compared our training-free framework against specialized event recognition systems across diverse application domains. Experiments demonstrate that our framework enhances multiple object detector architectures to recognize complex events such as illegal fishing activities (75% AUROC, +8.36% improvement), construction safety violations (+15.77%), and abnormal crowd behaviors (+23.16%). The code will be released soon.
comment: 13 pages, 5 figures
♻ ☆ A Vision Centric Remote Sensing Benchmark
Multimodal Large Language Models (MLLMs) have achieved remarkable success in vision-language tasks but their remote sensing (RS) counterpart are relatively under explored. Unlike natural images, RS imagery presents unique challenges that current MLLMs struggle to handle, particularly in visual grounding and spatial reasoning. This study investigates the limitations of CLIP-based MLLMs in RS, highlighting their failure to differentiate visually distinct yet semantically similar RS images. To address this, we introduce a remote sensing multimodal visual patterns (RSMMVP) benchmark. It is designed to evaluate MLLMs in RS tasks by identifying the CLIP-blind pairs, where CLIP-based models incorrectly assign high similarity scores to visually distinct RS images. Through a visual question answering (VQA) evaluation, we analyze the performance of state-of-the-art MLLMs, revealing significant limitations in RS specific representation learning. The results provide valuable insights into the weaknesses of CLIP-based visual encoding and offer a foundation for future research to develop more effective MLLMs tailored for remote sensing applications.
comment: Under Review
♻ ☆ No Thing, Nothing: Highlighting Safety-Critical Classes for Robust LiDAR Semantic Segmentation in Adverse Weather CVPR 2025
Existing domain generalization methods for LiDAR semantic segmentation under adverse weather struggle to accurately predict "things" categories compared to "stuff" categories. In typical driving scenes, "things" categories can be dynamic and associated with higher collision risks, making them crucial for safe navigation and planning. Recognizing the importance of "things" categories, we identify their performance drop as a serious bottleneck in existing approaches. We observed that adverse weather induces degradation of semantic-level features and both corruption of local features, leading to a misprediction of "things" as "stuff". To mitigate these corruptions, we suggest our method, NTN - segmeNt Things for No-accident. To address semantic-level feature corruption, we bind each point feature to its superclass, preventing the misprediction of things classes into visually dissimilar categories. Additionally, to enhance robustness against local corruption caused by adverse weather, we define each LiDAR beam as a local region and propose a regularization term that aligns the clean data with its corrupted counterpart in feature space. NTN achieves state-of-the-art performance with a +2.6 mIoU gain on the SemanticKITTI-to-SemanticSTF benchmark and +7.9 mIoU on the SemanticPOSS-to-SemanticSTF benchmark. Notably, NTN achieves a +4.8 and +7.9 mIoU improvement on "things" classes, respectively, highlighting its effectiveness.
comment: 18 pages, accepted in CVPR 2025
♻ ☆ Attention Overlap Is Responsible for The Entity Missing Problem in Text-to-image Diffusion Models!
Text-to-image diffusion models, such as Stable Diffusion and DALL-E, are capable of generating high-quality, diverse, and realistic images from textual prompts. However, they sometimes struggle to accurately depict specific entities described in prompts, a limitation known as the entity missing problem in compositional generation. While prior studies suggested that adjusting cross-attention maps during the denoising process could alleviate this problem, they did not systematically investigate which objective functions could best address it. This study examines three potential causes of the entity-missing problem, focusing on cross-attention dynamics: (1) insufficient attention intensity for certain entities, (2) overly broad attention spread, and (3) excessive overlap between attention maps of different entities. We found that reducing overlap in attention maps between entities can effectively minimize the rate of entity missing. Specifically, we hypothesize that tokens related to specific entities compete for attention on certain image regions during the denoising process, which can lead to divided attention across tokens and prevent accurate representation of each entity. To address this issue, we introduced four loss functions, Intersection over Union (IoU), center-of-mass (CoM) distance, Kullback-Leibler (KL) divergence, and clustering compactness (CC) to regulate attention overlap during denoising steps without the need for retraining. Experimental results across a wide variety of benchmarks reveal that these proposed training-free methods significantly improve compositional accuracy, outperforming previous approaches in visual question answering (VQA), captioning scores, CLIP similarity, and human evaluations. Notably, these methods improved human evaluation scores by 9% over the best baseline, demonstrating substantial improvements in compositional alignment.
comment: TMLR - 2025
♻ ☆ Superpixel Tokenization for Vision Transformers: Preserving Semantic Integrity in Visual Tokens
Transformers, a groundbreaking architecture proposed for Natural Language Processing (NLP), have also achieved remarkable success in Computer Vision. A cornerstone of their success lies in the attention mechanism, which models relationships among tokens. While the tokenization process in NLP inherently ensures that a single token does not contain multiple semantics, the tokenization of Vision Transformer (ViT) utilizes tokens from uniformly partitioned square image patches, which may result in an arbitrary mixing of visual concepts in a token. In this work, we propose to substitute the grid-based tokenization in ViT with superpixel tokenization, which employs superpixels to generate a token that encapsulates a sole visual concept. Unfortunately, the diverse shapes, sizes, and locations of superpixels make integrating superpixels into ViT tokenization rather challenging. Our tokenization pipeline, comprised of pre-aggregate extraction and superpixel-aware aggregation, overcomes the challenges that arise in superpixel tokenization. Extensive experiments demonstrate that our approach, which exhibits strong compatibility with existing frameworks, enhances the accuracy and robustness of ViT on various downstream tasks.
♻ ☆ GUI-World: A Video Benchmark and Dataset for Multimodal GUI-oriented Understanding ICLR 2025
Recently, Multimodal Large Language Models (MLLMs) have been used as agents to control keyboard and mouse inputs by directly perceiving the Graphical User Interface (GUI) and generating corresponding commands. However, current agents primarily demonstrate strong understanding capabilities in static environments and are mainly applied to relatively simple domains, such as Web or mobile interfaces. We argue that a robust GUI agent should be capable of perceiving temporal information on the GUI, including dynamic Web content and multi-step tasks. Additionally, it should possess a comprehensive understanding of various GUI scenarios, including desktop software and multi-window interactions. To this end, this paper introduces a new dataset, termed GUI-World, which features meticulously crafted Human-MLLM annotations, extensively covering six GUI scenarios and eight types of GUI-oriented questions in three formats. We evaluate the capabilities of current state-of-the-art MLLMs, including Image LLMs and Video LLMs, in understanding various types of GUI content, especially dynamic and sequential content. Our findings reveal that current models struggle with dynamic GUI content without manually annotated keyframes or operation history. On the other hand, Video LLMs fall short in all GUI-oriented tasks given the sparse GUI video dataset. Therefore, we take the initial step of leveraging a fine-tuned Video LLM, GUI-Vid, as a GUI-oriented assistant, demonstrating an improved understanding of various GUI tasks. However, due to the limitations in the performance of base LLMs, we conclude that using video LLMs as GUI agents remains a significant challenge. We believe our work provides valuable insights for future research in dynamic GUI content understanding. All the dataset and code are publicly available at: https://gui-world.github.io.
comment: Accepted by ICLR 2025
♻ ☆ CompMarkGS: Robust Watermarking for Compression 3D Gaussian Splatting
3D Gaussian Splatting (3DGS) enables rapid differentiable rendering for 3D reconstruction and novel view synthesis, leading to its widespread commercial use. Consequently, copyright protection via watermarking has become critical. However, because 3DGS relies on millions of Gaussians, which require gigabytes of storage, efficient transfer and storage require compression. Existing 3DGS watermarking methods are vulnerable to quantization-based compression, often resulting in the loss of the embedded watermark. To address this challenge, we propose a novel watermarking method that ensures watermark robustness after model compression while maintaining high rendering quality. In detail, we incorporate a quantization distortion layer that simulates compression during training, preserving the watermark under quantization-based compression. Also, we propose a learnable watermark embedding feature that embeds the watermark into the anchor feature, ensuring structural consistency and seamless integration into the 3D scene. Furthermore, we present a frequency-aware anchor growing mechanism to enhance image quality in high-frequency regions by effectively identifying Guassians within these regions. Experimental results confirm that our method preserves the watermark and maintains superior image quality under high compression, validating it as a promising approach for a secure 3DGS model.
comment: 23 pages, 17 figures
♻ ☆ Grounded Chain-of-Thought for Multimodal Large Language Models
Despite great progress, existing multimodal large language models (MLLMs) are prone to visual hallucination, greatly impeding their trustworthy applications. In this paper, we study this problem from the perspective of visual-spatial reasoning, and propose a new learning task for MLLMs, termed Grounded Chain-of-Thought (GCoT). Different from recent visual CoT studies, which focus more on visual knowledge reasoning, GCoT is keen to helping MLLMs to recognize and ground the relevant visual cues step by step, thereby predicting the correct answer with grounding coordinates as the intuitive basis. To facilitate this task, we also carefully design and construct a dataset called multimodal grounded chain-of-thought (MM-GCoT) consisting of 24,022 GCoT examples for 5,033 images. Besides, a comprehensive consistency evaluation system is also introduced, including the metrics of answer accuracy, grounding accuracy and answer-grounding consistency. We further design and conduct a bunch of experiments on 12 advanced MLLMs, and reveal some notable findings: i. most MLLMs performs poorly on the consistency evaluation, indicating obvious visual hallucination; ii. visual hallucination is not directly related to the parameter size and general multimodal performance, i.e., a larger and stronger MLLM is not less affected by this issue. Lastly, we also demonstrate that the proposed dataset can help existing MLLMs to well cultivate their GCoT capability and reduce the inconsistent answering significantly. Moreover, their GCoT can be also generalized to exiting multimodal tasks, such as open-world QA and REC.
♻ ☆ FLARE: Feed-forward Geometry, Appearance and Camera Estimation from Uncalibrated Sparse Views CVPR 2025
We present FLARE, a feed-forward model designed to infer high-quality camera poses and 3D geometry from uncalibrated sparse-view images (i.e., as few as 2-8 inputs), which is a challenging yet practical setting in real-world applications. Our solution features a cascaded learning paradigm with camera pose serving as the critical bridge, recognizing its essential role in mapping 3D structures onto 2D image planes. Concretely, FLARE starts with camera pose estimation, whose results condition the subsequent learning of geometric structure and appearance, optimized through the objectives of geometry reconstruction and novel-view synthesis. Utilizing large-scale public datasets for training, our method delivers state-of-the-art performance in the tasks of pose estimation, geometry reconstruction, and novel view synthesis, while maintaining the inference efficiency (i.e., less than 0.5 seconds). The project page and code can be found at: https://zhanghe3z.github.io/FLARE/
comment: CVPR 2025. Website: https://zhanghe3z.github.io/FLARE/
♻ ☆ Task-driven Image Fusion with Learnable Fusion Loss CVPR 2025
Multi-modal image fusion aggregates information from multiple sensor sources, achieving superior visual quality and perceptual features compared to single-source images, often improving downstream tasks. However, current fusion methods for downstream tasks still use predefined fusion objectives that potentially mismatch the downstream tasks, limiting adaptive guidance and reducing model flexibility. To address this, we propose Task-driven Image Fusion (TDFusion), a fusion framework incorporating a learnable fusion loss guided by task loss. Specifically, our fusion loss includes learnable parameters modeled by a neural network called the loss generation module. This module is supervised by the downstream task loss in a meta-learning manner. The learning objective is to minimize the task loss of fused images after optimizing the fusion module with the fusion loss. Iterative updates between the fusion module and the loss module ensure that the fusion network evolves toward minimizing task loss, guiding the fusion process toward the task objectives. TDFusion's training relies entirely on the downstream task loss, making it adaptable to any specific task. It can be applied to any architecture of fusion and task networks. Experiments demonstrate TDFusion's performance through fusion experiments conducted on four different datasets, in addition to evaluations on semantic segmentation and object detection tasks.
comment: Accepted to CVPR 2025
♻ ☆ ComicsPAP: understanding comic strips by picking the correct panel
Large multimodal models (LMMs) have made impressive strides in image captioning, VQA, and video comprehension, yet they still struggle with the intricate temporal and spatial cues found in comics. To address this gap, we introduce ComicsPAP, a large-scale benchmark designed for comic strip understanding. Comprising over 100k samples and organized into 5 subtasks under a Pick-a-Panel framework, ComicsPAP demands models to identify the missing panel in a sequence. Our evaluations, conducted under both multi-image and single-image protocols, reveal that current state-of-the-art LMMs perform near chance on these tasks, underscoring significant limitations in capturing sequential and contextual dependencies. To close the gap, we adapted LMMs for comic strip understanding, obtaining better results on ComicsPAP than 10x bigger models, demonstrating that ComicsPAP offers a robust resource to drive future research in multimodal comic comprehension.
♻ ☆ Continual Low-Rank Scaled Dot-product Attention
Transformers are widely used for their ability to capture data relations in sequence processing, with great success for a wide range of static tasks. However, the computational and memory footprint of their main component, i.e., the Scaled Dot-product Attention, is commonly overlooked. This makes their adoption in applications involving stream data processing with constraints in response latency, computational and memory resources infeasible. Some works have proposed methods to lower the computational cost of Transformers, i.e. low-rank approximations, sparsity in attention, and efficient formulations for Continual Inference. In this paper, we introduce a new formulation of the Scaled Dot-product Attention based on the Nystr\"om approximation that is suitable for Continual Inference. In experiments on Online Audio Classification and Online Action Detection tasks, the proposed Continual Scaled Dot-product Attention can lower the number of operations by up to three orders of magnitude compared to the original Transformers while retaining the predictive performance of competing models.
comment: 14 pages, 7 figures
♻ ☆ Fractal Calibration for long-tailed object detection CVPR2025
Real-world datasets follow an imbalanced distribution, which poses significant challenges in rare-category object detection. Recent studies tackle this problem by developing re-weighting and re-sampling methods, that utilise the class frequencies of the dataset. However, these techniques focus solely on the frequency statistics and ignore the distribution of the classes in image space, missing important information. In contrast to them, we propose FRActal CALibration (FRACAL): a novel post-calibration method for long-tailed object detection. FRACAL devises a logit adjustment method that utilises the fractal dimension to estimate how uniformly classes are distributed in image space. During inference, it uses the fractal dimension to inversely downweight the probabilities of uniformly spaced class predictions achieving balance in two axes: between frequent and rare categories, and between uniformly spaced and sparsely spaced classes. FRACAL is a post-processing method and it does not require any training, also it can be combined with many off-the-shelf models such as one-stage sigmoid detectors and two-stage instance segmentation models. FRACAL boosts the rare class performance by up to 8.6% and surpasses all previous methods on LVIS dataset, while showing good generalisation to other datasets such as COCO, V3Det and OpenImages. We provide the code at https://github.com/kostas1515/FRACAL.
comment: CVPR2025 (camera-ready)
♻ ☆ Constraint-Aware Feature Learning for Parametric Point Cloud
Parametric point clouds are sampled from CAD shapes and are becoming increasingly common in industrial manufacturing. Most existing CAD-specific deep learning methods only focus on geometric features, while overlooking constraints which are inherent and important in CAD shapes. This limits their ability to discern CAD shapes with similar appearance but different constraints. To tackle this challenge, we first analyze the constraint importance via a simple validation experiment. Then, we introduce a deep learning-friendly constraints representation with three vectorized components, and design a constraint-aware feature learning network (CstNet), which includes two stages. Stage 1 extracts constraint feature from B-Rep data or point cloud based on shape local information. It enables better generalization ability to unseen dataset after model pre-training. Stage 2 employs attention layers to adaptively adjust the weights of three constraints' components. It facilitates the effective utilization of constraints. In addition, we built the first multi-modal parametric-purpose dataset, i.e. Param20K, comprising about 20K shape instances of 75 classes. On this dataset, we performed the classification and rotation robustness experiments, and CstNet achieved 3.52\% and 26.17\% absolute improvements in instance accuracy over the state-of-the-art methods, respectively. To the best of our knowledge, CstNet is the first constraint-aware deep learning method tailored for parametric point cloud analysis in CAD domain.
♻ ☆ APLA: A Simple Adaptation Method for Vision Transformers
Existing adaptation techniques typically require architectural modifications or added parameters, leading to high computational costs and complexity. We introduce Attention Projection Layer Adaptation (APLA), a simple approach to adapt vision transformers (ViTs) without altering the architecture or adding parameters. Through a systematic analysis, we find that the layer immediately after the attention mechanism is crucial for adaptation. By updating only this projection layer, or even just a random subset of this layer's weights, APLA achieves state-of-the-art performance while reducing GPU memory usage by up to 52.63% and training time by up to 43.0%, with no extra cost at inference. Across 46 datasets covering a variety of tasks including scene classification, medical imaging, satellite imaging, and fine-grained classification, APLA consistently outperforms 17 other leading adaptation methods, including full fine-tuning, on classification, segmentation, and detection tasks. The code is available at https://github.com/MoeinSorkhei/APLA.
♻ ☆ Auto Cherry-Picker: Learning from High-quality Generative Data Driven by Language CVPR2025
Diffusion models can generate realistic and diverse images, potentially facilitating data availability for data-intensive perception tasks. However, leveraging these models to boost performance on downstream tasks with synthetic data poses several challenges, including aligning with real data distribution, scaling synthetic sample volumes, and ensuring their quality. To bridge these gaps, we present \textbf{A}uto \textbf{C}herry-\textbf{P}icker (ACP), a novel framework that generates high-quality cross-modality training samples at scale to augment perception and multi-modal training. ACP first uses LLMs to sample descriptions and layouts based on object combinations from real data priors, eliminating the need for ground truth image captions or annotations. Next, we use an off-the-shelf controllable diffusion model to generate multiple images. Then, the generated data are refined using a comprehensively designed metric, Composite Layout and Image Score (CLIS), to ensure quality. Our customized synthetic high-quality samples boost performance in various scenarios, especially in addressing challenges associated with long-tailed distribution and imbalanced datasets. Experiment results on downstream tasks demonstrate that ACP can significantly improve the performance of existing models. In addition, we find a positive correlation between CLIS and performance gains in downstream tasks. This finding shows the potential for evaluation metrics as the role for various visual perception and MLLM tasks.
comment: Accepted to CVPR2025
♻ ☆ Forensics Adapter: Adapting CLIP for Generalizable Face Forgery Detection CVPR 2025
We describe the Forensics Adapter, an adapter network designed to transform CLIP into an effective and generalizable face forgery detector. Although CLIP is highly versatile, adapting it for face forgery detection is non-trivial as forgery-related knowledge is entangled with a wide range of unrelated knowledge. Existing methods treat CLIP merely as a feature extractor, lacking task-specific adaptation, which limits their effectiveness. To address this, we introduce an adapter to learn face forgery traces -- the blending boundaries unique to forged faces, guided by task-specific objectives. Then we enhance the CLIP visual tokens with a dedicated interaction strategy that communicates knowledge across CLIP and the adapter. Since the adapter is alongside CLIP, its versatility is highly retained, naturally ensuring strong generalizability in face forgery detection. With only 5.7M trainable parameters, our method achieves a significant performance boost, improving by approximately 7% on average across five standard datasets. We believe the proposed method can serve as a baseline for future CLIP-based face forgery detection methods. The code is available at https://github.com/OUC-VAS/ForensicsAdapter.
comment: CVPR 2025
♻ ☆ LLM Post-Training: A Deep Dive into Reasoning Large Language Models
Large Language Models (LLMs) have transformed the natural language processing landscape and brought to life diverse applications. Pretraining on vast web-scale data has laid the foundation for these models, yet the research community is now increasingly shifting focus toward post-training techniques to achieve further breakthroughs. While pretraining provides a broad linguistic foundation, post-training methods enable LLMs to refine their knowledge, improve reasoning, enhance factual accuracy, and align more effectively with user intents and ethical considerations. Fine-tuning, reinforcement learning, and test-time scaling have emerged as critical strategies for optimizing LLMs performance, ensuring robustness, and improving adaptability across various real-world tasks. This survey provides a systematic exploration of post-training methodologies, analyzing their role in refining LLMs beyond pretraining, addressing key challenges such as catastrophic forgetting, reward hacking, and inference-time trade-offs. We highlight emerging directions in model alignment, scalable adaptation, and inference-time reasoning, and outline future research directions. We also provide a public repository to continually track developments in this fast-evolving field: https://github.com/mbzuai-oryx/Awesome-LLM-Post-training.
comment: 32 pages, 7 figures, 3 tables, 377 references. Github Repo: https://github.com/mbzuai-oryx/Awesome-LLM-Post-training
♻ ☆ Bokeh Diffusion: Defocus Blur Control in Text-to-Image Diffusion Models
Recent advances in large-scale text-to-image models have revolutionized creative fields by generating visually captivating outputs from textual prompts; however, while traditional photography offers precise control over camera settings to shape visual aesthetics -- such as depth-of-field -- current diffusion models typically rely on prompt engineering to mimic such effects. This approach often results in crude approximations and inadvertently altering the scene content. In this work, we propose Bokeh Diffusion, a scene-consistent bokeh control framework that explicitly conditions a diffusion model on a physical defocus blur parameter. By grounding depth-of-field adjustments, our method preserves the underlying scene structure as the level of blur is varied. To overcome the scarcity of paired real-world images captured under different camera settings, we introduce a hybrid training pipeline that aligns in-the-wild images with synthetic blur augmentations. Extensive experiments demonstrate that our approach not only achieves flexible, lens-like blur control but also supports applications such as real image editing via inversion.
comment: Project page: https://atfortes.github.io/projects/bokeh-diffusion/
♻ ☆ HRGR: Enhancing Image Manipulation Detection via Hierarchical Region-aware Graph Reasoning ICME 2025
Image manipulation detection is to identify the authenticity of each pixel in images. One typical approach to uncover manipulation traces is to model image correlations. The previous methods commonly adopt the grids, which are fixed-size squares, as graph nodes to model correlations. However, these grids, being independent of image content, struggle to retain local content coherence, resulting in imprecise detection.To address this issue, we describe a new method named Hierarchical Region-aware Graph Reasoning (HRGR) to enhance image manipulation detection. Unlike existing grid-based methods, we model image correlations based on content-coherence feature regions with irregular shapes, generated by a novel Differentiable Feature Partition strategy. Then we construct a Hierarchical Region-aware Graph based on these regions within and across different feature layers. Subsequently, we describe a structural-agnostic graph reasoning strategy tailored for our graph to enhance the representation of nodes. Our method is fully differentiable and can seamlessly integrate into mainstream networks in an end-to-end manner, without requiring additional supervision. Extensive experiments demonstrate the effectiveness of our method in image manipulation detection, exhibiting its great potential as a plug-and-play component for existing architectures. Codes and models are available at https://github.com/OUC-VAS/HRGR-IMD.
comment: ICME 2025
♻ ☆ Distilled Prompt Learning for Incomplete Multimodal Survival Prediction CVPR2025
The integration of multimodal data including pathology images and gene profiles is widely applied in precise survival prediction. Despite recent advances in multimodal survival models, collecting complete modalities for multimodal fusion still poses a significant challenge, hindering their application in clinical settings. Current approaches tackling incomplete modalities often fall short, as they typically compensate for only a limited part of the knowledge of missing modalities. To address this issue, we propose a Distilled Prompt Learning framework (DisPro) to utilize the strong robustness of Large Language Models (LLMs) to missing modalities, which employs two-stage prompting for compensation of comprehensive information for missing modalities. In the first stage, Unimodal Prompting (UniPro) distills the knowledge distribution of each modality, preparing for supplementing modality-specific knowledge of the missing modality in the subsequent stage. In the second stage, Multimodal Prompting (MultiPro) leverages available modalities as prompts for LLMs to infer the missing modality, which provides modality-common information. Simultaneously, the unimodal knowledge acquired in the first stage is injected into multimodal inference to compensate for the modality-specific knowledge of the missing modality. Extensive experiments covering various missing scenarios demonstrated the superiority of the proposed method. The code is available at https://github.com/Innse/DisPro.
comment: Accepted by CVPR2025
♻ ☆ S2CFormer: Revisiting the RD-Latency Trade-off in Transformer-based Learned Image Compression
Transformer-based Learned Image Compression (LIC) suffers from a suboptimal trade-off between decoding latency and rate-distortion (R-D) performance. Moreover, the critical role of the FeedForward Network (FFN)-based channel aggregation module has been largely overlooked. Our research reveals that efficient channel aggregation-rather than complex and time-consuming spatial operations-is the key to achieving competitive LIC models. Based on this insight, we initiate the ``S2CFormer'' paradigm, a general architecture that simplifies spatial operations and enhances channel operations to overcome the previous trade-off. We present two instances of the S2CFormer: S2C-Conv, and S2C-Attention. Both models demonstrate state-of-the-art (SOTA) R-D performance and significantly faster decoding speed. Furthermore, we introduce S2C-Hybrid, an enhanced variant that maximizes the strengths of different S2CFormer instances to achieve a better performance-latency trade-off. This model outperforms all the existing methods on the Kodak, Tecnick, and CLIC Professional Validation datasets, setting a new benchmark for efficient and high-performance LIC. The code is at \href{https://github.com/YunuoChen/S2CFormer}{https://github.com/YunuoChen/S2CFormer}.
♻ ☆ JanusFlow: Harmonizing Autoregression and Rectified Flow for Unified Multimodal Understanding and Generation CVPR 2025
We present JanusFlow, a powerful framework that unifies image understanding and generation in a single model. JanusFlow introduces a minimalist architecture that integrates autoregressive language models with rectified flow, a state-of-the-art method in generative modeling. Our key finding demonstrates that rectified flow can be straightforwardly trained within the large language model framework, eliminating the need for complex architectural modifications. To further improve the performance of our unified model, we adopt two key strategies: (i) decoupling the understanding and generation encoders, and (ii) aligning their representations during unified training. Extensive experiments show that JanusFlow achieves comparable or superior performance to specialized models in their respective domains, while significantly outperforming existing unified approaches across standard benchmarks. This work represents a step toward more efficient and versatile vision-language models.
comment: Accepted by CVPR 2025
♻ ☆ Humanoid Policy ~ Human Policy
Training manipulation policies for humanoid robots with diverse data enhances their robustness and generalization across tasks and platforms. However, learning solely from robot demonstrations is labor-intensive, requiring expensive tele-operated data collection which is difficult to scale. This paper investigates a more scalable data source, egocentric human demonstrations, to serve as cross-embodiment training data for robot learning. We mitigate the embodiment gap between humanoids and humans from both the data and modeling perspectives. We collect an egocentric task-oriented dataset (PH2D) that is directly aligned with humanoid manipulation demonstrations. We then train a human-humanoid behavior policy, which we term Human Action Transformer (HAT). The state-action space of HAT is unified for both humans and humanoid robots and can be differentiably retargeted to robot actions. Co-trained with smaller-scale robot data, HAT directly models humanoid robots and humans as different embodiments without additional supervision. We show that human data improves both generalization and robustness of HAT with significantly better data collection efficiency. Code and data: https://human-as-robot.github.io/
comment: Code and data: https://human-as-robot.github.io/
♻ ☆ Adapter Merging with Centroid Prototype Mapping for Scalable Class-Incremental Learning CVPR 2025
We propose Adapter Merging with Centroid Prototype Mapping (ACMap), an exemplar-free framework for class-incremental learning (CIL) that addresses both catastrophic forgetting and scalability. While existing methods involve a trade-off between inference time and accuracy, ACMap consolidates task-specific adapters into a single adapter, thus achieving constant inference time across tasks without sacrificing accuracy. The framework employs adapter merging to build a shared subspace that aligns task representations and mitigates forgetting, while centroid prototype mapping maintains high accuracy by consistently adapting representations within the shared subspace. To further improve scalability, an early stopping strategy limits adapter merging as tasks increase. Extensive experiments on five benchmark datasets demonstrate that ACMap matches state-of-the-art accuracy while maintaining inference time comparable to the fastest existing methods. The code is available at https://github.com/tf63/ACMap.
comment: Accepted to CVPR 2025
♻ ☆ Using deep neural networks to detect non-analytically defined expert event labels in canoe sprint force sensor signals
Assessing an athlete's performance in canoe sprint is often established by measuring a variety of kinematic parameters during training sessions. Many of these parameters are related to single or multiple paddle stroke cycles. Determining on- and offset of these cycles in force sensor signals is usually not straightforward and requires human interaction. This paper explores convolutional neural networks (CNNs) and recurrent neural networks (RNNs) in terms of their ability to automatically predict these events. In addition, our work proposes an extension to the recently published SoftED metric for event detection in order to properly assess the model performance on time windows. In our results, an RNN based on bidirectional gated recurrent units (BGRUs) turned out to be the most suitable model for paddle stroke detection.
♻ ☆ MoDec-GS: Global-to-Local Motion Decomposition and Temporal Interval Adjustment for Compact Dynamic 3D Gaussian Splatting CVPR2025
3D Gaussian Splatting (3DGS) has made significant strides in scene representation and neural rendering, with intense efforts focused on adapting it for dynamic scenes. Despite delivering remarkable rendering quality and speed, existing methods struggle with storage demands and representing complex real-world motions. To tackle these issues, we propose MoDecGS, a memory-efficient Gaussian splatting framework designed for reconstructing novel views in challenging scenarios with complex motions. We introduce GlobaltoLocal Motion Decomposition (GLMD) to effectively capture dynamic motions in a coarsetofine manner. This approach leverages Global Canonical Scaffolds (Global CS) and Local Canonical Scaffolds (Local CS), extending static Scaffold representation to dynamic video reconstruction. For Global CS, we propose Global Anchor Deformation (GAD) to efficiently represent global dynamics along complex motions, by directly deforming the implicit Scaffold attributes which are anchor position, offset, and local context features. Next, we finely adjust local motions via the Local Gaussian Deformation (LGD) of Local CS explicitly. Additionally, we introduce Temporal Interval Adjustment (TIA) to automatically control the temporal coverage of each Local CS during training, allowing MoDecGS to find optimal interval assignments based on the specified number of temporal segments. Extensive evaluations demonstrate that MoDecGS achieves an average 70% reduction in model size over stateoftheart methods for dynamic 3D Gaussians from realworld dynamic videos while maintaining or even improving rendering quality.
comment: CVPR2025 (camera ready ver.). The last two authors are co-corresponding authors. Please visit our project page at https://kaist-viclab.github.io/MoDecGS-site/
♻ ☆ GEOcc: Geometrically Enhanced 3D Occupancy Network with Implicit-Explicit Depth Fusion and Contextual Self-Supervision
3D occupancy perception holds a pivotal role in recent vision-centric autonomous driving systems by converting surround-view images into integrated geometric and semantic representations within dense 3D grids. Nevertheless, current models still encounter two main challenges: modeling depth accurately in the 2D-3D view transformation stage, and overcoming the lack of generalizability issues due to sparse LiDAR supervision. To address these issues, this paper presents GEOcc, a Geometric-Enhanced Occupancy network tailored for vision-only surround-view perception. Our approach is three-fold: 1) Integration of explicit lift-based depth prediction and implicit projection-based transformers for depth modeling, enhancing the density and robustness of view transformation. 2) Utilization of mask-based encoder-decoder architecture for fine-grained semantic predictions; 3) Adoption of context-aware self-training loss functions in the pertaining stage to complement LiDAR supervision, involving the re-rendering of 2D depth maps from 3D occupancy features and leveraging image reconstruction loss to obtain denser depth supervision besides sparse LiDAR ground-truths. Our approach achieves State-Of-The-Art performance on the Occ3D-nuScenes dataset with the least image resolution needed and the most weightless image backbone compared with current models, marking an improvement of 3.3% due to our proposed contributions. Comprehensive experimentation also demonstrates the consistent superiority of our method over baselines and alternative approaches.
comment: This work has been accepted for publication in IEEE Transactions on Intelligent Transportation Systems
♻ ☆ Visual Persona: Foundation Model for Full-Body Human Customization CVPR 2025
We introduce Visual Persona, a foundation model for text-to-image full-body human customization that, given a single in-the-wild human image, generates diverse images of the individual guided by text descriptions. Unlike prior methods that focus solely on preserving facial identity, our approach captures detailed full-body appearance, aligning with text descriptions for body structure and scene variations. Training this model requires large-scale paired human data, consisting of multiple images per individual with consistent full-body identities, which is notoriously difficult to obtain. To address this, we propose a data curation pipeline leveraging vision-language models to evaluate full-body appearance consistency, resulting in Visual Persona-500K, a dataset of 580k paired human images across 100k unique identities. For precise appearance transfer, we introduce a transformer encoder-decoder architecture adapted to a pre-trained text-to-image diffusion model, which augments the input image into distinct body regions, encodes these regions as local appearance features, and projects them into dense identity embeddings independently to condition the diffusion model for synthesizing customized images. Visual Persona consistently surpasses existing approaches, generating high-quality, customized images from in-the-wild inputs. Extensive ablation studies validate design choices, and we demonstrate the versatility of Visual Persona across various downstream tasks.
comment: CVPR 2025, Project page is available at https://cvlab-kaist.github.io/Visual-Persona
♻ ☆ A Deep Learning Model for Coronary Artery Segmentation and Quantitative Stenosis Detection in Angiographic Images
Coronary artery disease (CAD) is a leading cause of cardiovascular-related mortality, and accurate stenosis detection is crucial for effective clinical decision-making. Coronary angiography remains the gold standard for diagnosing CAD, but manual analysis of angiograms is prone to errors and subjectivity. This study aims to develop a deep learning-based approach for the automatic segmentation of coronary arteries from angiographic images and the quantitative detection of stenosis, thereby improving the accuracy and efficiency of CAD diagnosis. We propose a novel deep learning-based method for the automatic segmentation of coronary arteries in angiographic images, coupled with a dynamic cohort method for stenosis detection. The segmentation model combines the MedSAM and VM-UNet architectures to achieve high-performance results. After segmentation, the vascular centerline is extracted, vessel diameter is computed, and the degree of stenosis is measured with high precision, enabling accurate identification of arterial stenosis. On the mixed dataset (including the ARCADE, DCA1, and GH datasets), the model achieved an average IoU of 0.6308, with sensitivity and specificity of 0.9772 and 0.9903, respectively. On the ARCADE dataset, the average IoU was 0.6303, with sensitivity of 0.9832 and specificity of 0.9933. Additionally, the stenosis detection algorithm achieved a true positive rate (TPR) of 0.5867 and a positive predictive value (PPV) of 0.5911, demonstrating the effectiveness of our model in analyzing coronary angiography images. SAM-VMNet offers a promising tool for the automated segmentation and detection of coronary artery stenosis. The model's high accuracy and robustness provide significant clinical value for the early diagnosis and treatment planning of CAD. The code and examples are available at https://github.com/qimingfan10/SAM-VMNet.
♻ ☆ Divide and Conquer: Heterogeneous Noise Integration for Diffusion-based Adversarial Purification
Existing diffusion-based purification methods aim to disrupt adversarial perturbations by introducing a certain amount of noise through a forward diffusion process, followed by a reverse process to recover clean examples. However, this approach is fundamentally flawed: the uniform operation of the forward process across all pixels compromises normal pixels while attempting to combat adversarial perturbations, resulting in the target model producing incorrect predictions. Simply relying on low-intensity noise is insufficient for effective defense. To address this critical issue, we implement a heterogeneous purification strategy grounded in the interpretability of neural networks. Our method decisively applies higher-intensity noise to specific pixels that the target model focuses on while the remaining pixels are subjected to only low-intensity noise. This requirement motivates us to redesign the sampling process of the diffusion model, allowing for the effective removal of varying noise levels. Furthermore, to evaluate our method against strong adaptative attack, our proposed method sharply reduces time cost and memory usage through a single-step resampling. The empirical evidence from extensive experiments across three datasets demonstrates that our method outperforms most current adversarial training and purification techniques by a substantial margin.
♻ ☆ Lessons and Insights from a Unifying Study of Parameter-Efficient Fine-Tuning (PEFT) in Visual Recognition CVPR 2025
Parameter-efficient fine-tuning (PEFT) has attracted significant attention due to the growth of pre-trained model sizes and the need to fine-tune (FT) them for superior downstream performance. Despite a surge in new PEFT methods, a systematic study to understand their performance and suitable application scenarios is lacking, leaving questions like "when to apply PEFT" and "which method to use" largely unanswered, especially in visual recognition. In this paper, we conduct a unifying empirical study of representative PEFT methods with Vision Transformers. We systematically tune their hyper-parameters to fairly compare their accuracy on downstream tasks. Our study offers a practical user guide and unveils several new insights. First, if tuned carefully, different PEFT methods achieve similar accuracy in the low-shot benchmark VTAB-1K. This includes simple approaches like FT the bias terms that were reported inferior. Second, despite similar accuracy, we find that PEFT methods make different mistakes and high-confidence predictions, likely due to their different inductive biases. Such an inconsistency (or complementarity) opens up the opportunity for ensemble methods, and we make preliminary attempts at this. Third, going beyond the commonly used low-shot tasks, we find that PEFT is also useful in many-shot regimes, achieving comparable or better accuracy than full FT while using significantly fewer parameters. Lastly, we investigate PEFT's ability to preserve a pre-trained model's robustness to distribution shifts (e.g., CLIP). Perhaps not surprisingly, PEFT approaches outperform full FT alone. However, with weight-space ensembles, full FT can better balance target distribution and distribution shift performance, suggesting a future research direction for robust PEFT.
comment: CVPR 2025. The code is available at https://github.com/OSU-MLB/ViT_PEFT_Vision
♻ ☆ HiLM-D: Enhancing MLLMs with Multi-Scale High-Resolution Details for Autonomous Driving
Recent efforts to use natural language for interpretable driving focus mainly on planning, neglecting perception tasks. In this paper, we address this gap by introducing ROLISP (Risk Object Localization and Intention and Suggestion Prediction), which towards interpretable risk object detection and suggestion for ego car motions. Accurate ROLISP implementation requires extensive reasoning to identify critical traffic objects and infer their intentions, prompting us to explore the capabilities of multimodal large language models (MLLMs). However, the limited perception performance of CLIP-ViT vision encoders in existing MLLMs struggles with capturing essential visual perception information, e.g., high-resolution, multi-scale and visual-related inductive biases, which are important for autonomous driving. Addressing these challenges, we introduce HiLM-D, a resource-efficient framework that enhances visual information processing in MLLMs for ROLISP. Our method is motivated by the fact that the primary variations in autonomous driving scenarios are the motion trajectories rather than the semantic or appearance information (e.g., the shapes and colors) of objects. Hence, the visual process of HiLM-D is a two-stream framework: (i) a temporal reasoning stream, receiving low-resolution dynamic video content, to capture temporal semantics, and (ii) a spatial perception stream, receiving a single high-resolution frame, to capture holistic visual perception-related information. The spatial perception stream can be made very lightweight by a well-designed P-Adapter, which is lightweight, training-efficient, and easily integrated into existing MLLMs. Experiments on the DRAMA-ROLISP dataset show HiLM-D's significant improvements over current MLLMs, with a 3.7% in BLEU-4 for captioning and 8.7% in mIoU for detection.
comment: Accepted by IJCV
♻ ☆ Prediction and Reference Quality Adaptation for Learned Video Compression
Temporal prediction is one of the most important technologies for video compression. Various prediction coding modes are designed in traditional video codecs. Traditional video codecs will adaptively to decide the optimal coding mode according to the prediction quality and reference quality. Recently, learned video codecs have made great progress. However, they did not effectively address the problem of prediction and reference quality adaptation, which limits the effective utilization of temporal prediction and reduction of reconstruction error propagation. Therefore, in this paper, we first propose a confidence-based prediction quality adaptation (PQA) module to provide explicit discrimination for the spatial and channel-wise prediction quality difference. With this module, the prediction with low quality will be suppressed and that with high quality will be enhanced. The codec can adaptively decide which spatial or channel location of predictions to use. Then, we further propose a reference quality adaptation (RQA) module and an associated repeat-long training strategy to provide dynamic spatially variant filters for diverse reference qualities. With these filters, our codec can adapt to different reference qualities, making it easier to achieve the target reconstruction quality and reduce the reconstruction error propagation. Experimental results verify that our proposed modules can effectively help our codec achieve a higher compression performance.
♻ ☆ DEEPTalk: Dynamic Emotion Embedding for Probabilistic Speech-Driven 3D Face Animation AAAI 2025
Speech-driven 3D facial animation has garnered lots of attention thanks to its broad range of applications. Despite recent advancements in achieving realistic lip motion, current methods fail to capture the nuanced emotional undertones conveyed through speech and produce monotonous facial motion. These limitations result in blunt and repetitive facial animations, reducing user engagement and hindering their applicability. To address these challenges, we introduce DEEPTalk, a novel approach that generates diverse and emotionally rich 3D facial expressions directly from speech inputs. To achieve this, we first train DEE (Dynamic Emotion Embedding), which employs probabilistic contrastive learning to forge a joint emotion embedding space for both speech and facial motion. This probabilistic framework captures the uncertainty in interpreting emotions from speech and facial motion, enabling the derivation of emotion vectors from its multifaceted space. Moreover, to generate dynamic facial motion, we design TH-VQVAE (Temporally Hierarchical VQ-VAE) as an expressive and robust motion prior overcoming limitations of VAEs and VQ-VAEs. Utilizing these strong priors, we develop DEEPTalk, a talking head generator that non-autoregressively predicts codebook indices to create dynamic facial motion, incorporating a novel emotion consistency loss. Extensive experiments on various datasets demonstrate the effectiveness of our approach in creating diverse, emotionally expressive talking faces that maintain accurate lip-sync. Our project page is available at https://whwjdqls.github.io/deeptalk\_website/
comment: First two authors contributed equally. This is a revised version of the original submission, which has been accepted for publication at AAAI 2025
♻ ☆ Attention-guided Feature Distillation for Semantic Segmentation
Deep learning models have achieved significant results across various computer vision tasks. However, due to the large number of parameters in these models, deploying them in real-time scenarios is a critical challenge, specifically in dense prediction tasks such as semantic segmentation. Knowledge distillation has emerged as a successful technique for addressing this problem by transferring knowledge from a cumbersome model (teacher) to a lighter model (student). In contrast to existing complex methodologies commonly employed for distilling knowledge from a teacher to a student, this paper showcases the efficacy of a simple yet powerful method for utilizing refined feature maps to transfer attention. The proposed method has proven to be effective in distilling rich information, outperforming existing methods in semantic segmentation as a dense prediction task. The proposed Attention-guided Feature Distillation (AttnFD) method, employs the Convolutional Block Attention Module (CBAM), which refines feature maps by taking into account both channel-specific and spatial information content. Simply using the Mean Squared Error (MSE) loss function between the refined feature maps of the teacher and the student, AttnFD demonstrates outstanding performance in semantic segmentation, achieving state-of-the-art results in terms of improving the mean Intersection over Union (mIoU) of the student network on the PascalVoc 2012, Cityscapes, COCO, and CamVid datasets.
comment: 26 pages, 10 figures, and 6 tables
♻ ☆ DreamText: High Fidelity Scene Text Synthesis
Scene text synthesis involves rendering specified texts onto arbitrary images. Current methods typically formulate this task in an end-to-end manner but lack effective character-level guidance during training. Besides, their text encoders, pre-trained on a single font type, struggle to adapt to the diverse font styles encountered in practical applications. Consequently, these methods suffer from character distortion, repetition, and absence, particularly in polystylistic scenarios. To this end, this paper proposes DreamText for high-fidelity scene text synthesis. Our key idea is to reconstruct the diffusion training process, introducing more refined guidance tailored to this task, to expose and rectify the model's attention at the character level and strengthen its learning of text regions. This transformation poses a hybrid optimization challenge, involving both discrete and continuous variables. To effectively tackle this challenge, we employ a heuristic alternate optimization strategy. Meanwhile, we jointly train the text encoder and generator to comprehensively learn and utilize the diverse font present in the training dataset. This joint training is seamlessly integrated into the alternate optimization process, fostering a synergistic relationship between learning character embedding and re-estimating character attention. Specifically, in each step, we first encode potential character-generated position information from cross-attention maps into latent character masks. These masks are then utilized to update the representation of specific characters in the current step, which, in turn, enables the generator to correct the character's attention in the subsequent steps. Both qualitative and quantitative results demonstrate the superiority of our method to the state of the art.
comment: Code: https://github.com/CodeGoat24/DreamText, Project page: https://codegoat24.github.io/DreamText/
♻ ☆ MMO-IG: Multi-Class and Multi-Scale Object Image Generation for Remote Sensing
The rapid advancement of deep generative models (DGMs) has significantly advanced research in computer vision, providing a cost-effective alternative to acquiring vast quantities of expensive imagery. However, existing methods predominantly focus on synthesizing remote sensing (RS) images aligned with real images in a global layout view, which limits their applicability in RS image object detection (RSIOD) research. To address these challenges, we propose a multi-class and multi-scale object image generator based on DGMs, termed MMO-IG, designed to generate RS images with supervised object labels from global and local aspects simultaneously. Specifically, from the local view, MMO-IG encodes various RS instances using an iso-spacing instance map (ISIM). During the generation process, it decodes each instance region with iso-spacing value in ISIM-corresponding to both background and foreground instances-to produce RS images through the denoising process of diffusion models. Considering the complex interdependencies among MMOs, we construct a spatial-cross dependency knowledge graph (SCDKG). This ensures a realistic and reliable multidirectional distribution among MMOs for region embedding, thereby reducing the discrepancy between source and target domains. Besides, we propose a structured object distribution instruction (SODI) to guide the generation of synthesized RS image content from a global aspect with SCDKG-based ISIM together. Extensive experimental results demonstrate that our MMO-IG exhibits superior generation capabilities for RS images with dense MMO-supervised labels, and RS detectors pre-trained with MMO-IG show excellent performance on real-world datasets.
♻ ☆ Zero-Shot Head Swapping in Real-World Scenarios CVPR'25
With growing demand in media and social networks for personalized images, the need for advanced head-swapping techniques, integrating an entire head from the head image with the body from the body image, has increased. However, traditional head swapping methods heavily rely on face-centered cropped data with primarily frontal facing views, which limits their effectiveness in real world applications. Additionally, their masking methods, designed to indicate regions requiring editing, are optimized for these types of dataset but struggle to achieve seamless blending in complex situations, such as when the original data includes features like long hair extending beyond the masked area. To overcome these limitations and enhance adaptability in diverse and complex scenarios, we propose a novel head swapping method, HID, that is robust to images including the full head and the upper body, and handles from frontal to side views, while automatically generating context aware masks. For automatic mask generation, we introduce the IOMask, which enables seamless blending of the head and body, effectively addressing integration challenges. We further introduce the hair injection module to capture hair details with greater precision. Our experiments demonstrate that the proposed approach achieves state-of-the-art performance in head swapping, providing visually consistent and realistic results across a wide range of challenging conditions.
comment: CVPR'25
♻ ☆ SplatFlow: Multi-View Rectified Flow Model for 3D Gaussian Splatting Synthesis
Text-based generation and editing of 3D scenes hold significant potential for streamlining content creation through intuitive user interactions. While recent advances leverage 3D Gaussian Splatting (3DGS) for high-fidelity and real-time rendering, existing methods are often specialized and task-focused, lacking a unified framework for both generation and editing. In this paper, we introduce SplatFlow, a comprehensive framework that addresses this gap by enabling direct 3DGS generation and editing. SplatFlow comprises two main components: a multi-view rectified flow (RF) model and a Gaussian Splatting Decoder (GSDecoder). The multi-view RF model operates in latent space, generating multi-view images, depths, and camera poses simultaneously, conditioned on text prompts, thus addressing challenges like diverse scene scales and complex camera trajectories in real-world settings. Then, the GSDecoder efficiently translates these latent outputs into 3DGS representations through a feed-forward 3DGS method. Leveraging training-free inversion and inpainting techniques, SplatFlow enables seamless 3DGS editing and supports a broad range of 3D tasks-including object editing, novel view synthesis, and camera pose estimation-within a unified framework without requiring additional complex pipelines. We validate SplatFlow's capabilities on the MVImgNet and DL3DV-7K datasets, demonstrating its versatility and effectiveness in various 3D generation, editing, and inpainting-based tasks.
comment: Project Page: https://gohyojun15.github.io/SplatFlow/
♻ ☆ Large-scale Pre-training for Grounded Video Caption Generation
We propose a novel approach for captioning and object grounding in video, where the objects in the caption are grounded in the video via temporally dense bounding boxes. We introduce the following contributions. First, we present a large-scale automatic annotation method that aggregates captions grounded with bounding boxes across individual frames into temporally dense and consistent bounding box annotations. We apply this approach on the HowTo100M dataset to construct a large-scale pre-training dataset, named HowToGround1M. We also introduce a Grounded Video Caption Generation model, dubbed GROVE, and pre-train the model on HowToGround1M. Second, we introduce a new dataset, called iGround, of 3500 videos with manually annotated captions and dense spatio-temporally grounded bounding boxes. This allows us to measure progress on this challenging problem, as well as to fine-tune our model on this small-scale but high-quality data. Third, we demonstrate that our approach achieves state-of-the-art results on the proposed iGround dataset compared to a number of baselines, as well as on the VidSTG and ActivityNet-Entities datasets. We perform extensive ablations that demonstrate the importance of pre-training using our automatically annotated HowToGround1M dataset followed by fine-tuning on the manually annotated iGround dataset and validate the key technical contributions of our model.
♻ ☆ Task-Augmented Cross-View Imputation Network for Partial Multi-View Incomplete Multi-Label Classification
In real-world scenarios, multi-view multi-label learning often encounters the challenge of incomplete training data due to limitations in data collection and unreliable annotation processes. The absence of multi-view features impairs the comprehensive understanding of samples, omitting crucial details essential for classification. To address this issue, we present a task-augmented cross-view imputation network (TACVI-Net) for the purpose of handling partial multi-view incomplete multi-label classification. Specifically, we employ a two-stage network to derive highly task-relevant features to recover the missing views. In the first stage, we leverage the information bottleneck theory to obtain a discriminative representation of each view by extracting task-relevant information through a view-specific encoder-classifier architecture. In the second stage, an autoencoder based multi-view reconstruction network is utilized to extract high-level semantic representation of the augmented features and recover the missing data, thereby aiding the final classification task. Extensive experiments on five datasets demonstrate that our TACVI-Net outperforms other state-of-the-art methods.
♻ ☆ Flowing from Words to Pixels: A Noise-Free Framework for Cross-Modality Evolution CVPR 2025
Diffusion models, and their generalization, flow matching, have had a remarkable impact on the field of media generation. Here, the conventional approach is to learn the complex mapping from a simple source distribution of Gaussian noise to the target media distribution. For cross-modal tasks such as text-to-image generation, this same mapping from noise to image is learnt whilst including a conditioning mechanism in the model. One key and thus far relatively unexplored feature of flow matching is that, unlike Diffusion models, they are not constrained for the source distribution to be noise. Hence, in this paper, we propose a paradigm shift, and ask the question of whether we can instead train flow matching models to learn a direct mapping from the distribution of one modality to the distribution of another, thus obviating the need for both the noise distribution and conditioning mechanism. We present a general and simple framework, CrossFlow, for cross-modal flow matching. We show the importance of applying Variational Encoders to the input data, and introduce a method to enable Classifier-free guidance. Surprisingly, for text-to-image, CrossFlow with a vanilla transformer without cross attention slightly outperforms standard flow matching, and we show that it scales better with training steps and model size, while also allowing for interesting latent arithmetic which results in semantically meaningful edits in the output space. To demonstrate the generalizability of our approach, we also show that CrossFlow is on par with or outperforms the state-of-the-art for various cross-modal / intra-modal mapping tasks, viz. image captioning, depth estimation, and image super-resolution. We hope this paper contributes to accelerating progress in cross-modal media generation.
comment: CVPR 2025 camera-ready version. Project page: https://cross-flow.github.io/
♻ ☆ MOSAIC: Generating Consistent, Privacy-Preserving Scenes from Multiple Depth Views in Multi-Room Environments
We introduce a novel diffusion-based approach for generating privacy-preserving digital twins of multi-room indoor environments from depth images only. Central to our approach is a novel Multi-view Overlapped Scene Alignment with Implicit Consistency (MOSAIC) model that explicitly considers cross-view dependencies within the same scene in the probabilistic sense. MOSAIC operates through a novel inference-time optimization that avoids error accumulation common in sequential or single-room constraint in panorama-based approaches. MOSAIC scales to complex scenes with zero extra training and provably reduces the variance during denoising processes when more overlapping views are added, leading to improved generation quality. Experiments show that MOSAIC outperforms state-of-the-art baselines on image fidelity metrics in reconstructing complex multi-room environments. Project page is available at: https://mosaic-cmubig.github.io
♻ ☆ Multi-modal Multi-platform Person Re-Identification: Benchmark and Method
Conventional person re-identification (ReID) research is often limited to single-modality sensor data from static cameras, which fails to address the complexities of real-world scenarios where multi-modal signals are increasingly prevalent. For instance, consider an urban ReID system integrating stationary RGB cameras, nighttime infrared sensors, and UAVs equipped with dynamic tracking capabilities. Such systems face significant challenges due to variations in camera perspectives, lighting conditions, and sensor modalities, hindering effective person ReID. To address these challenges, we introduce the MP-ReID benchmark, a novel dataset designed specifically for multi-modality and multi-platform ReID. This benchmark uniquely compiles data from 1,930 identities across diverse modalities, including RGB, infrared, and thermal imaging, captured by both UAVs and ground-based cameras in indoor and outdoor environments. Building on this benchmark, we introduce Uni-Prompt ReID, a framework with specific-designed prompts, tailored for cross-modality and cross-platform scenarios. Our method consistently outperforms state-of-the-art approaches, establishing a robust foundation for future research in complex and dynamic ReID environments. Our dataset are available at:https://mp-reid.github.io/.
♻ ☆ IRGS: Inter-Reflective Gaussian Splatting with 2D Gaussian Ray Tracing CVPR 2025
In inverse rendering, accurately modeling visibility and indirect radiance for incident light is essential for capturing secondary effects. Due to the absence of a powerful Gaussian ray tracer, previous 3DGS-based methods have either adopted a simplified rendering equation or used learnable parameters to approximate incident light, resulting in inaccurate material and lighting estimations. To this end, we introduce inter-reflective Gaussian splatting (IRGS) for inverse rendering. To capture inter-reflection, we apply the full rendering equation without simplification and compute incident radiance on the fly using the proposed differentiable 2D Gaussian ray tracing. Additionally, we present an efficient optimization scheme to handle the computational demands of Monte Carlo sampling for rendering equation evaluation. Furthermore, we introduce a novel strategy for querying the indirect radiance of incident light when relighting the optimized scenes. Extensive experiments on multiple standard benchmarks validate the effectiveness of IRGS, demonstrating its capability to accurately model complex inter-reflection effects.
comment: CVPR 2025. Project page: https://fudan-zvg.github.io/IRGS
♻ ☆ iFlame: Interleaving Full and Linear Attention for Efficient Mesh Generation
This paper propose iFlame, a novel transformer-based network architecture for mesh generation. While attention-based models have demonstrated remarkable performance in mesh generation, their quadratic computational complexity limits scalability, particularly for high-resolution 3D data. Conversely, linear attention mechanisms offer lower computational costs but often struggle to capture long-range dependencies, resulting in suboptimal outcomes. To address this trade-off, we propose an interleaving autoregressive mesh generation framework that combines the efficiency of linear attention with the expressive power of full attention mechanisms. To further enhance efficiency and leverage the inherent structure of mesh representations, we integrate this interleaving approach into an hourglass architecture, which significantly boosts efficiency. Our approach reduces training time while achieving performance comparable to pure attention-based models. To improve inference efficiency, we implemented a caching algorithm that almost doubles the speed and reduces the KV cache size by seven-eighths compared to the original Transformer. We evaluate our framework on ShapeNet and Objaverse, demonstrating its ability to generate high-quality 3D meshes efficiently. Our results indicate that the proposed interleaving framework effectively balances computational efficiency and generative performance, making it a practical solution for mesh generation. The training takes only 2 days with 4 GPUs on 39k data with a maximum of 4k faces on Objaverse.
comment: Project website: https://wanghanxiao123.github.io/iFa/
♻ ☆ DiffusionDrive: Truncated Diffusion Model for End-to-End Autonomous Driving CVPR 2025
Recently, the diffusion model has emerged as a powerful generative technique for robotic policy learning, capable of modeling multi-mode action distributions. Leveraging its capability for end-to-end autonomous driving is a promising direction. However, the numerous denoising steps in the robotic diffusion policy and the more dynamic, open-world nature of traffic scenes pose substantial challenges for generating diverse driving actions at a real-time speed. To address these challenges, we propose a novel truncated diffusion policy that incorporates prior multi-mode anchors and truncates the diffusion schedule, enabling the model to learn denoising from anchored Gaussian distribution to the multi-mode driving action distribution. Additionally, we design an efficient cascade diffusion decoder for enhanced interaction with conditional scene context. The proposed model, DiffusionDrive, demonstrates 10$\times$ reduction in denoising steps compared to vanilla diffusion policy, delivering superior diversity and quality in just 2 steps. On the planning-oriented NAVSIM dataset, with the aligned ResNet-34 backbone, DiffusionDrive achieves 88.1 PDMS without bells and whistles, setting a new record, while running at a real-time speed of 45 FPS on an NVIDIA 4090. Qualitative results on challenging scenarios further confirm that DiffusionDrive can robustly generate diverse plausible driving actions. Code and model will be available at https://github.com/hustvl/DiffusionDrive.
comment: Accepted to CVPR 2025. Code & demo & model are available at https://github.com/hustvl/DiffusionDrive
♻ ☆ Applications of Spiking Neural Networks in Visual Place Recognition
In robotics, Spiking Neural Networks (SNNs) are increasingly recognized for their largely-unrealized potential energy efficiency and low latency particularly when implemented on neuromorphic hardware. Our paper highlights three advancements for SNNs in Visual Place Recognition (VPR). Firstly, we propose Modular SNNs, where each SNN represents a set of non-overlapping geographically distinct places, enabling scalable networks for large environments. Secondly, we present Ensembles of Modular SNNs, where multiple networks represent the same place, significantly enhancing accuracy compared to single-network models. Each of our Modular SNN modules is compact, comprising only 1500 neurons and 474k synapses, making them ideally suited for ensembling due to their small size. Lastly, we investigate the role of sequence matching in SNN-based VPR, a technique where consecutive images are used to refine place recognition. We demonstrate competitive performance of our method on a range of datasets, including higher responsiveness to ensembling compared to conventional VPR techniques and higher R@1 improvements with sequence matching than VPR techniques with comparable baseline performance. Our contributions highlight the viability of SNNs for VPR, offering scalable and robust solutions, and paving the way for their application in various energy-sensitive robotic tasks.
comment: 20 pages, 10 figures, IEEE Transactions on Robotics (TRO)
♻ ☆ Similarity-Dissimilarity Loss for Multi-label Supervised Contrastive Learning
Supervised contrastive learning has been explored in making use of label information for multi-label classification, but determining positive samples in multi-label scenario remains challenging. Previous studies have examined strategies for identifying positive samples, considering label overlap proportion between anchors and samples. However, they ignore various relations between given anchors and samples, as well as how to dynamically adjust the weights in contrastive loss functions based on different relations, leading to great ambiguity. In this paper, we introduce five distinct relations between multi-label samples and propose a Similarity-Dissimilarity Loss with contrastive learning for multi-label classification. Our loss function re-weights the loss by computing the similarity and dissimilarity between positive samples and a given anchor based on the introduced relations. We mainly conduct experiments for multi-label text classification on MIMIC datasets, then further extend the evaluation on MS-COCO. The Experimental results show that our proposed loss effectively improves the performance on all encoders under supervised contrastive learning paradigm, demonstrating its effectiveness and robustness.
♻ ☆ CoRLD: Contrastive Representation Learning Of Deformable Shapes In Images
Deformable shape representations, parameterized by deformations relative to a given template, have proven effective for improved image analysis tasks. However, their broader applicability is hindered by two major challenges. First, existing methods mainly rely on a known template during testing, which is impractical and limits flexibility. Second, they often struggle to capture fine-grained, voxel-level distinctions between similar shapes (e.g., anatomical variations among healthy individuals, those with mild cognitive impairment, and diseased states). To address these limitations, we propose a novel framework - Contrastive Representation Learning of Deformable shapes (CoRLD) in learned deformation spaces and demonstrate its effectiveness in the context of image classification. Our CoRLD leverages a class-aware contrastive supervised learning objective in latent deformation spaces, promoting proximity among representations of similar classes while ensuring separation of dissimilar groups. In contrast to previous deep learning networks that require a reference image as input to predict deformation changes, our approach eliminates this dependency. Instead, template images are utilized solely as ground truth in the loss function during the training process, making our model more flexible and generalizable to a wide range of medical applications. We validate CoRLD on diverse datasets, including real brain magnetic resonance imaging (MRIs) and adrenal shapes derived from computed tomography (CT) scans. Experimental results show that our model effectively extracts deformable shape features, which can be easily integrated with existing classifiers to substantially boost the classification accuracy. Our code is available at GitHub.
♻ ☆ BiM-VFI: Bidirectional Motion Field-Guided Frame Interpolation for Video with Non-uniform Motions
Existing Video Frame interpolation (VFI) models tend to suffer from time-to-location ambiguity when trained with video of non-uniform motions, such as accelerating, decelerating, and changing directions, which often yield blurred interpolated frames. In this paper, we propose (i) a novel motion description map, Bidirectional Motion field (BiM), to effectively describe non-uniform motions; (ii) a BiM-guided Flow Net (BiMFN) with Content-Aware Upsampling Network (CAUN) for precise optical flow estimation; and (iii) Knowledge Distillation for VFI-centric Flow supervision (KDVCF) to supervise the motion estimation of VFI model with VFI-centric teacher flows. The proposed VFI is called a Bidirectional Motion field-guided VFI (BiM-VFI) model. Extensive experiments show that our BiM-VFI model significantly surpasses the recent state-of-the-art VFI methods by 26% and 45% improvements in LPIPS and STLPIPS respectively, yielding interpolated frames with much fewer blurs at arbitrary time instances.
comment: The last two authors are co-corresponding authors
♻ ☆ ReTaKe: Reducing Temporal and Knowledge Redundancy for Long Video Understanding
Video Large Language Models (VideoLLMs) have made significant strides in video understanding but struggle with long videos due to the limitations of their backbone LLMs. Existing solutions rely on length extrapolation, which is memory-constrained, or visual token compression, which primarily leverages low-level temporal redundancy while overlooking the more effective high-level knowledge redundancy. To address this, we propose $\textbf{ReTaKe}$, a training-free method with two novel modules DPSelect and PivotKV, to jointly reduce both temporal visual redundancy and knowledge redundancy for video compression. To align with the way of human temporal perception, DPSelect identifies keyframes based on inter-frame distance peaks. To leverage LLMs' learned prior knowledge, PivotKV marks the keyframes as pivots and compress non-pivot frames by pruning low-attention tokens in their KV cache. ReTaKe enables VideoLLMs to process 8 times longer frames (up to 2048), outperforming similar-sized models by 3-5% and even rivaling much larger ones on VideoMME, MLVU, LongVideoBench, and LVBench. Moreover, by overlapping compression operations with prefilling, ReTaKe introduces only ~10% prefilling latency overhead while reducing decoding latency by ~20%. Our code is available at https://github.com/SCZwangxiao/video-ReTaKe.
comment: Rewrite the methods section. Add more ablation studies and results in LongVideoBench. Update metadata
♻ ☆ Spiking Neural Network as Adaptive Event Stream Slicer NeurIPS 2024
Event-based cameras are attracting significant interest as they provide rich edge information, high dynamic range, and high temporal resolution. Many state-of-the-art event-based algorithms rely on splitting the events into fixed groups, resulting in the omission of crucial temporal information, particularly when dealing with diverse motion scenarios (\eg, high/low speed).In this work, we propose SpikeSlicer, a novel-designed plug-and-play event processing method capable of splitting events stream adaptively.SpikeSlicer utilizes a low-energy spiking neural network (SNN) to trigger event slicing. To guide the SNN to fire spikes at optimal time steps, we propose the Spiking Position-aware Loss (SPA-Loss) to modulate the neuron's state. Additionally, we develop a Feedback-Update training strategy that refines the slicing decisions using feedback from the downstream artificial neural network (ANN). Extensive experiments demonstrate that our method yields significant performance improvements in event-based object tracking and recognition. Notably, SpikeSlicer provides a brand-new SNN-ANN cooperation paradigm, where the SNN acts as an efficient, low-energy data processor to assist the ANN in improving downstream performance, injecting new perspectives and potential avenues of exploration. Our code is available at https://github.com/AndyCao1125/SpikeSlicer.
comment: Accepted to NeurIPS 2024
♻ ☆ Good, Cheap, and Fast: Overfitted Image Compression with Wasserstein Distortion CVPR 2025
Inspired by the success of generative image models, recent work on learned image compression increasingly focuses on better probabilistic models of the natural image distribution, leading to excellent image quality. This, however, comes at the expense of a computational complexity that is several orders of magnitude higher than today's commercial codecs, and thus prohibitive for most practical applications. With this paper, we demonstrate that by focusing on modeling visual perception rather than the data distribution, we can achieve a very good trade-off between visual quality and bit rate similar to "generative" compression models such as HiFiC, while requiring less than 1% of the multiply-accumulate operations (MACs) for decompression. We do this by optimizing C3, an overfitted image codec, for Wasserstein Distortion (WD), and evaluating the image reconstructions with a human rater study, showing that WD clearly outperforms LPIPS as an optimization objective. The study also reveals that WD outperforms other perceptual metrics such as LPIPS, DISTS, and MS-SSIM as a predictor of human ratings, remarkably achieving over 94% Pearson correlation with Elo scores.
comment: 16 pages, 12 figures. Accepted for presentation at CVPR 2025
♻ ☆ Bayesian Prompt Flow Learning for Zero-Shot Anomaly Detection
Recently, vision-language models (e.g. CLIP) have demonstrated remarkable performance in zero-shot anomaly detection (ZSAD). By leveraging auxiliary data during training, these models can directly perform cross-category anomaly detection on target datasets, such as detecting defects on industrial product surfaces or identifying tumors in organ tissues. Existing approaches typically construct text prompts through either manual design or the optimization of learnable prompt vectors. However, these methods face several challenges: 1) handcrafted prompts require extensive expert knowledge and trial-and-error; 2) single-form learnable prompts struggle to capture complex anomaly semantics; and 3) an unconstrained prompt space limits generalization to unseen categories. To address these issues, we propose Bayesian Prompt Flow Learning (Bayes-PFL), which models the prompt space as a learnable probability distribution from a Bayesian perspective. Specifically, a prompt flow module is designed to learn both image-specific and image-agnostic distributions, which are jointly utilized to regularize the text prompt space and improve the model's generalization on unseen categories. These learned distributions are then sampled to generate diverse text prompts, effectively covering the prompt space. Additionally, a residual cross-model attention (RCA) module is introduced to better align dynamic text embeddings with fine-grained image features. Extensive experiments on 15 industrial and medical datasets demonstrate our method's superior performance. The code is available at https://github.com/xiaozhen228/Bayes-PFL.
♻ ☆ 3D-MVP: 3D Multiview Pretraining for Robotic Manipulation CVPR 2025
Recent works have shown that visual pretraining on egocentric datasets using masked autoencoders (MAE) can improve generalization for downstream robotics tasks. However, these approaches pretrain only on 2D images, while many robotics applications require 3D scene understanding. In this work, we propose 3D-MVP, a novel approach for 3D Multi-View Pretraining using masked autoencoders. We leverage Robotic View Transformer (RVT), which uses a multi-view transformer to understand the 3D scene and predict gripper pose actions. We split RVT's multi-view transformer into visual encoder and action decoder, and pretrain its visual encoder using masked autoencoding on large-scale 3D datasets such as Objaverse. We evaluate 3D-MVP on a suite of virtual robot manipulation tasks and demonstrate improved performance over baselines. Our results suggest that 3D-aware pretraining is a promising approach to improve generalization of vision-based robotic manipulation policies. Project site: https://jasonqsy.github.io/3DMVP
comment: CVPR 2025
♻ ☆ CADDreamer: CAD Object Generation from Single-view Images CVPR 2025
Diffusion-based 3D generation has made remarkable progress in recent years. However, existing 3D generative models often produce overly dense and unstructured meshes, which stand in stark contrast to the compact, structured, and sharply-edged Computer-Aided Design (CAD) models crafted by human designers. To address this gap, we introduce CADDreamer, a novel approach for generating boundary representations (B-rep) of CAD objects from a single image. CADDreamer employs a primitive-aware multi-view diffusion model that captures both local geometric details and high-level structural semantics during the generation process. By encoding primitive semantics into the color domain, the method leverages the strong priors of pre-trained diffusion models to align with well-defined primitives. This enables the inference of multi-view normal maps and semantic maps from a single image, facilitating the reconstruction of a mesh with primitive labels. Furthermore, we introduce geometric optimization techniques and topology-preserving extraction methods to mitigate noise and distortion in the generated primitives. These enhancements result in a complete and seamless B-rep of the CAD model. Experimental results demonstrate that our method effectively recovers high-quality CAD objects from single-view images. Compared to existing 3D generation techniques, the B-rep models produced by CADDreamer are compact in representation, clear in structure, sharp in edges, and watertight in topology.
comment: Accepted to CVPR 2025
♻ ☆ Zero-Shot Low Light Image Enhancement with Diffusion Prior
In this paper, we present a simple yet highly effective "free lunch" solution for low-light image enhancement (LLIE), which aims to restore low-light images as if acquired in well-illuminated environments. Our method necessitates no optimization, training, fine-tuning, text conditioning, or hyperparameter adjustments, yet it consistently reconstructs low-light images with superior fidelity. Specifically, we leverage a pre-trained text-to-image diffusion prior, learned from training on a large collection of natural images, and the features present in the model itself to guide the inference, in contrast to existing methods that depend on customized constraints. Comprehensive quantitative evaluations demonstrate that our approach outperforms SOTA methods on established datasets, while qualitative analyses indicate enhanced color accuracy and the rectification of subtle chromatic deviations. Furthermore, additional experiments reveal that our method, without any modifications, achieves SOTA-comparable performance in the auto white balance (AWB) task.
Information Retrieval 14
☆ Exploring Training and Inference Scaling Laws in Generative Retrieval
Generative retrieval has emerged as a novel paradigm that leverages large language models (LLMs) to autoregressively generate document identifiers. Although promising, the mechanisms that underpin its performance and scalability remain largely unclear. We conduct a systematic investigation of training and inference scaling laws in generative retrieval, exploring how model size, training data scale, and inference-time compute jointly influence retrieval performance. To address the lack of suitable metrics, we propose a novel evaluation measure inspired by contrastive entropy and generation loss, providing a continuous performance signal that enables robust comparisons across diverse generative retrieval methods. Our experiments show that n-gram-based methods demonstrate strong alignment with both training and inference scaling laws, especially when paired with larger LLMs. Furthermore, increasing inference computation yields substantial performance gains, revealing that generative retrieval can significantly benefit from higher compute budgets at inference. Across these settings, LLaMA models consistently outperform T5 models, suggesting a particular advantage for larger decoder-only models in generative retrieval. Taken together, our findings underscore that model sizes, data availability, and inference computation interact to unlock the full potential of generative retrieval, offering new insights for designing and optimizing future systems.
☆ Toward building next-generation Geocoding systems: a systematic review
Geocoding systems are widely used in both scientific research for spatial analysis and everyday life through location-based services. The quality of geocoded data significantly impacts subsequent processes and applications, underscoring the need for next-generation systems. In response to this demand, this review first examines the evolving requirements for geocoding inputs and outputs across various scenarios these systems must address. It then provides a detailed analysis of how to construct such systems by breaking them down into key functional components and reviewing a broad spectrum of existing approaches, from traditional rule-based methods to advanced techniques in information retrieval, natural language processing, and large language models. Finally, we identify opportunities to improve next-generation geocoding systems in light of recent technological advances.
☆ CCMusic: An Open and Diverse Database for Chinese Music Information Retrieval Research
Data are crucial in various computer-related fields, including music information retrieval (MIR), an interdisciplinary area bridging computer science and music. This paper introduces CCMusic, an open and diverse database comprising multiple datasets specifically designed for tasks related to Chinese music, highlighting our focus on this culturally rich domain. The database integrates both published and unpublished datasets, with steps taken such as data cleaning, label refinement, and data structure unification to ensure data consistency and create ready-to-use versions. We conduct benchmark evaluations for all datasets using a unified evaluation framework developed specifically for this purpose. This publicly available framework supports both classification and detection tasks, ensuring standardized and reproducible results across all datasets. The database is hosted on HuggingFace and ModelScope, two open and multifunctional data and model hosting platforms, ensuring ease of accessibility and usability.
comment: 17 pages, 18 figures
☆ ArchSeek: Retrieving Architectural Case Studies Using Vision-Language Models
Efficiently searching for relevant case studies is critical in architectural design, as designers rely on precedent examples to guide or inspire their ongoing projects. However, traditional text-based search tools struggle to capture the inherently visual and complex nature of architectural knowledge, often leading to time-consuming and imprecise exploration. This paper introduces ArchSeek, an innovative case study search system with recommendation capability, tailored for architecture design professionals. Powered by the visual understanding capabilities from vision-language models and cross-modal embeddings, it enables text and image queries with fine-grained control, and interaction-based design case recommendations. It offers architects a more efficient, personalized way to discover design inspirations, with potential applications across other visually driven design fields. The source code is available at https://github.com/danruili/ArchSeek.
comment: 15 pages, 8 figures, 3 tables. Accepted by CAAD Futures 2025
☆ A Comprehensive Review on Hashtag Recommendation: From Traditional to Deep Learning and Beyond
The exponential growth of user-generated content on social media platforms has precipitated significant challenges in information management, particularly in content organization, retrieval, and discovery. Hashtags, as a fundamental categorization mechanism, play a pivotal role in enhancing content visibility and user engagement. However, the development of accurate and robust hashtag recommendation systems remains a complex and evolving research challenge. Existing surveys in this domain are limited in scope and recency, focusing narrowly on specific platforms, methodologies, or timeframes. To address this gap, this review article conducts a systematic analysis of hashtag recommendation systems, comprehensively examining recent advancements across several dimensions. We investigate unimodal versus multimodal methodologies, diverse problem formulations, filtering strategies, methodological evolution from traditional frequency-based models to advanced deep learning architectures. Furthermore, we critically evaluate performance assessment paradigms, including quantitative metrics, qualitative analyses, and hybrid evaluation frameworks. Our analysis underscores a paradigm shift toward transformer-based deep learning models, which harness contextual and semantic features to achieve superior recommendation accuracy. Key challenges such as data sparsity, cold-start scenarios, polysemy, and model explainability are rigorously discussed, alongside practical applications in tweet classification, sentiment analysis, and content popularity prediction. By synthesizing insights from diverse methodological and platform-specific perspectives, this survey provides a structured taxonomy of current research, identifies unresolved gaps, and proposes future directions for developing adaptive, user-centric recommendation systems.
☆ Dense Retrieval for Low Resource Languages -- the Case of Amharic Language
This paper reports some difficulties and some results when using dense retrievers on Amharic, one of the low-resource languages spoken by 120 millions populations. The efforts put and difficulties faced by University Addis Ababa toward Amharic Information Retrieval will be developed during the presentation.
comment: 4 pages, 2 figures
☆ Robust-IR @ SIGIR 2025: The First Workshop on Robust Information Retrieval SIGIR 2025
With the advancement of information retrieval (IR) technologies, robustness is increasingly attracting attention. When deploying technology into practice, we consider not only its average performance under normal conditions but, more importantly, its ability to maintain functionality across a variety of exceptional situations. In recent years, the research on IR robustness covers theory, evaluation, methodology, and application, and all of them show a growing trend. The purpose of this workshop is to systematize the latest results of each research aspect, to foster comprehensive communication within this niche domain while also bridging robust IR research with the broader community, and to promote further future development of robust IR. To avoid the one-sided talk of mini-conferences, this workshop adopts a highly interactive format, including round-table and panel discussion sessions, to encourage active participation and meaningful exchange among attendees.
comment: Accept by SIGIR 2025
☆ PRECTR: A Synergistic Framework for Integrating Personalized Search Relevance Matching and CTR Prediction
The two primary tasks in the search recommendation system are search relevance matching and click-through rate (CTR) prediction -- the former focuses on seeking relevant items for user queries whereas the latter forecasts which item may better match user interest. Prior research typically develops two models to predict the CTR and search relevance separately, then ranking candidate items based on the fusion of the two outputs. However, such a divide-and-conquer paradigm creates the inconsistency between different models. Meanwhile, the search relevance model mainly concentrates on the degree of objective text matching while neglecting personalized differences among different users, leading to restricted model performance. To tackle these issues, we propose a unified \textbf{P}ersonalized Search RElevance Matching and CTR Prediction Fusion Model(PRECTR). Specifically, based on the conditional probability fusion mechanism, PRECTR integrates the CTR prediction and search relevance matching into one framework to enhance the interaction and consistency of the two modules. However, directly optimizing CTR binary classification loss may bring challenges to the fusion model's convergence and indefinitely promote the exposure of items with high CTR, regardless of their search relevance. Hence, we further introduce two-stage training and semantic consistency regularization to accelerate the model's convergence and restrain the recommendation of irrelevant items. Finally, acknowledging that different users may have varied relevance preferences, we assessed current users' relevance preferences by analyzing past users' preferences for similar queries and tailored incentives for different candidate items accordingly. Extensive experimental results on our production dataset and online A/B testing demonstrate the effectiveness and superiority of our proposed PRECTR method.
☆ Food Recommendation With Balancing Comfort and Curiosity
Food is a key pleasure of traveling, but travelers face a trade-off between exploring curious new local food and choosing comfortable, familiar options. This creates demand for personalized recommendation systems that balance these competing factors. To the best of our knowledge, conventional recommendation methods cannot provide recommendations that offer both curiosity and comfort for food unknown to the user at a travel destination. In this study, we propose new quantitative methods for estimating comfort and curiosity: Kernel Density Scoring (KDS) and Mahalanobis Distance Scoring (MDS). KDS probabilistically estimates food history distribution using kernel density estimation, while MDS uses Mahalanobis distances between foods. These methods score food based on how their representation vectors fit the estimated distributions. We also propose a ranking method measuring the balance between comfort and curiosity based on taste and ingredients. This balance is defined as curiosity (return) gained per unit of comfort (risk) in choosing a food. For evaluation the proposed method, we newly collected a dataset containing user surveys on Japanese food and assessments of foreign food regarding comfort and curiosity. Comparing our methods against the existing method, the Wilcoxon signed-rank test showed that when estimating comfort from taste and curiosity from ingredients, the MDS-based method outperformed the Baseline, while the KDS-based method showed no significant differences. When estimating curiosity from taste and comfort from ingredients, both methods outperformed the Baseline. The MDS-based method consistently outperformed KDS in ROC-AUC values.
☆ RAU: Towards Regularized Alignment and Uniformity for Representation Learning in Recommendation
Recommender systems (RecSys) have become essential in modern society, driving user engagement and satisfaction across diverse online platforms. Most RecSys focuses on designing a powerful encoder to embed users and items into high-dimensional vector representation space, with loss functions optimizing their representation distributions. Recent studies reveal that directly optimizing key properties of the representation distribution, such as alignment and uniformity, can outperform complex encoder designs. However, existing methods for optimizing critical attributes overlook the impact of dataset sparsity on the model: limited user-item interactions lead to sparse alignment, while excessive interactions result in uneven uniformity, both of which degrade performance. In this paper, we identify the sparse alignment and uneven uniformity issues, and further propose Regularized Alignment and Uniformity (RAU) to cope with these two issues accordingly. RAU consists of two novel regularization methods for alignment and uniformity to learn better user/item representation. 1) Center-strengthened alignment further aligns the average in-batch user/item representation to provide an enhanced alignment signal and further minimize the disparity between user and item representation. 2) Low-variance-guided uniformity minimizes the variance of pairwise distances along with uniformity, which provides extra guidance to a more stabilized uniformity increase during training. We conducted extensive experiments on three real-world datasets, and the proposed RAU resulted in significant performance improvements compared to current state-of-the-art CF methods, which confirms the advantages of the two proposed regularization methods.
♻ ☆ Large Language Models Empowered Personalized Web Agents WWW 2025
Web agents have emerged as a promising direction to automate Web task completion based on user instructions, significantly enhancing user experience. Recently, Web agents have evolved from traditional agents to Large Language Models (LLMs)-based Web agents. Despite their success, existing LLM-based Web agents overlook the importance of personalized data (e.g., user profiles and historical Web behaviors) in assisting the understanding of users' personalized instructions and executing customized actions. To overcome the limitation, we first formulate the task of LLM-empowered personalized Web agents, which integrate personalized data and user instructions to personalize instruction comprehension and action execution. To address the absence of a comprehensive evaluation benchmark, we construct a Personalized Web Agent Benchmark (PersonalWAB), featuring user instructions, personalized user data, Web functions, and two evaluation paradigms across three personalized Web tasks. Moreover, we propose a Personalized User Memory-enhanced Alignment (PUMA) framework to adapt LLMs to the personalized Web agent task. PUMA utilizes a memory bank with a task-specific retrieval strategy to filter relevant historical Web behaviors. Based on the behaviors, PUMA then aligns LLMs for personalized action execution through fine-tuning and direct preference optimization. Extensive experiments validate the superiority of PUMA over existing Web agents on PersonalWAB.
comment: Accepted to WWW 2025. The code and data are available on the project website https://hongrucai.github.io/PersonalWAB/
♻ ☆ Characterizing User Behavior: The Interplay Between Mobility Patterns and Mobile Traffic
Mobile devices have become essential for capturing human activity, and eXtended Data Records (XDRs) offer rich opportunities for detailed user behavior modeling, which is useful for designing personalized digital services. Previous studies have primarily focused on aggregated mobile traffic and mobility analyses, often neglecting individual-level insights. This paper introduces a novel approach that explores the dependency between traffic and mobility behaviors at the user level. By analyzing 13 individual features that encompass traffic patterns and various mobility aspects, we enhance the understanding of how these behaviors interact. Our advanced user modeling framework integrates traffic and mobility behaviors over time, allowing for fine-grained dependencies while maintaining population heterogeneity through user-specific signatures. Furthermore, we develop a Markov model that infers traffic behavior from mobility and vice versa, prioritizing significant dependencies while addressing privacy concerns. Using a week-long XDR dataset from 1,337,719 users across several provinces in Chile, we validate our approach, demonstrating its robustness and applicability in accurately inferring user behavior and matching mobility and traffic profiles across diverse urban contexts.
♻ ☆ Behavior Modeling Space Reconstruction for E-Commerce Search
Delivering superior search services is crucial for enhancing customer experience and driving revenue growth. Conventionally, search systems model user behaviors by combining user preference and query item relevance statically, often through a fixed logical 'and' relationship. This paper reexamines existing approaches through a unified lens using both causal graphs and Venn diagrams, uncovering two prevalent yet significant issues: entangled preference and relevance effects, and a collapsed modeling space. To surmount these challenges, our research introduces a novel framework, DRP, which enhances search accuracy through two components to reconstruct the behavior modeling space. Specifically, we implement preference editing to proactively remove the relevance effect from preference predictions, yielding untainted user preferences. Additionally, we employ adaptive fusion, which dynamically adjusts fusion criteria to align with the varying patterns of relevance and preference, facilitating more nuanced and tailored behavior predictions within the reconstructed modeling space. Empirical validation on two public datasets and a proprietary search dataset underscores the superiority of our proposed methodology, demonstrating marked improvements in performance over existing approaches.
♻ ☆ Efficient Long Sequential Low-rank Adaptive Attention for Click-through rate Prediction
In the context of burgeoning user historical behavior data, Accurate click-through rate(CTR) prediction requires effective modeling of lengthy user behavior sequences. As the volume of such data keeps swelling, the focus of research has shifted towards developing effective long-term behavior modeling methods to capture latent user interests. Nevertheless, the complexity introduced by large scale data brings about computational hurdles. There is a pressing need to strike a balance between achieving high model performance and meeting the strict response time requirements of online services. While existing retrieval-based methods (e.g., similarity filtering or attention approximation) achieve practical runtime efficiency, they inherently compromise information fidelity through aggressive sequence truncation or attention sparsification. This paper presents a novel attention mechanism. It overcomes the shortcomings of existing methods while ensuring computational efficiency. This mechanism learn compressed representation of sequence with length $L$ via low-rank projection matrices (rank $r \ll L$), reducing attention complexity from $O(L)$ to $O(r)$. It also integrates a uniquely designed loss function to preserve nonlinearity of attention. In the inference stage, the mechanism adopts matrix absorption and prestorage strategies. These strategies enable it to effectively satisfy online constraints. Comprehensive offline and online experiments demonstrate that the proposed method outperforms current state-of-the-art solutions.
Machine Learning 158
☆ Aether: Geometric-Aware Unified World Modeling
The integration of geometric reconstruction and generative modeling remains a critical challenge in developing AI systems capable of human-like spatial reasoning. This paper proposes Aether, a unified framework that enables geometry-aware reasoning in world models by jointly optimizing three core capabilities: (1) 4D dynamic reconstruction, (2) action-conditioned video prediction, and (3) goal-conditioned visual planning. Through task-interleaved feature learning, Aether achieves synergistic knowledge sharing across reconstruction, prediction, and planning objectives. Building upon video generation models, our framework demonstrates unprecedented synthetic-to-real generalization despite never observing real-world data during training. Furthermore, our approach achieves zero-shot generalization in both action following and reconstruction tasks, thanks to its intrinsic geometric modeling. Remarkably, even without real-world data, its reconstruction performance far exceeds that of domain-specific models. Additionally, Aether leverages a geometry-informed action space to seamlessly translate predictions into actions, enabling effective autonomous trajectory planning. We hope our work inspires the community to explore new frontiers in physically-reasonable world modeling and its applications.
comment: Project Page: https://aether-world.github.io/
AdaWorld: Learning Adaptable World Models with Latent Actions
World models aim to learn action-controlled prediction models and have proven essential for the development of intelligent agents. However, most existing world models rely heavily on substantial action-labeled data and costly training, making it challenging to adapt to novel environments with heterogeneous actions through limited interactions. This limitation can hinder their applicability across broader domains. To overcome this challenge, we propose AdaWorld, an innovative world model learning approach that enables efficient adaptation. The key idea is to incorporate action information during the pretraining of world models. This is achieved by extracting latent actions from videos in a self-supervised manner, capturing the most critical transitions between frames. We then develop an autoregressive world model that conditions on these latent actions. This learning paradigm enables highly adaptable world models, facilitating efficient transfer and learning of new actions even with limited interactions and finetuning. Our comprehensive experiments across multiple environments demonstrate that AdaWorld achieves superior performance in both simulation quality and visual planning.
comment: Project page: https://adaptable-world-model.github.io/
Trajectory Balance with Asynchrony: Decoupling Exploration and Learning for Fast, Scalable LLM Post-Training
Reinforcement learning (RL) is a critical component of large language model (LLM) post-training. However, existing on-policy algorithms used for post-training are inherently incompatible with the use of experience replay buffers, which can be populated scalably by distributed off-policy actors to enhance exploration as compute increases. We propose efficiently obtaining this benefit of replay buffers via Trajectory Balance with Asynchrony (TBA), a massively scalable LLM RL system. In contrast to existing approaches, TBA uses a larger fraction of compute on search, constantly generating off-policy data for a central replay buffer. A training node simultaneously samples data from this buffer based on reward or recency to update the policy using Trajectory Balance (TB), a diversity-seeking RL objective introduced for GFlowNets. TBA offers three key advantages: (1) decoupled training and search, speeding up training wall-clock time by 4x or more; (2) improved diversity through large-scale off-policy sampling; and (3) scalable search for sparse reward settings. On mathematical reasoning, preference-tuning, and automated red-teaming (diverse and representative post-training tasks), TBA produces speed and performance improvements over strong baselines.
☆ FFN Fusion: Rethinking Sequential Computation in Large Language Models
We introduce FFN Fusion, an architectural optimization technique that reduces sequential computation in large language models by identifying and exploiting natural opportunities for parallelization. Our key insight is that sequences of Feed-Forward Network (FFN) layers, particularly those remaining after the removal of specific attention layers, can often be parallelized with minimal accuracy impact. We develop a principled methodology for identifying and fusing such sequences, transforming them into parallel operations that significantly reduce inference latency while preserving model behavior. Applying these techniques to Llama-3.1-405B-Instruct, we create Llama-Nemotron-Ultra-253B-Base (Ultra-253B-Base), an efficient and soon-to-be publicly available model that achieves a 1.71X speedup in inference latency and 35X lower per-token cost while maintaining strong performance across benchmarks. Through extensive experiments on models from 49B to 253B parameters, we demonstrate that FFN Fusion becomes increasingly effective at larger scales and can complement existing optimization techniques like quantization and pruning. Most intriguingly, we find that even full transformer blocks containing both attention and FFN layers can sometimes be parallelized, suggesting new directions for neural architecture design.
☆ Building Blocks for Robust and Effective Semi-Supervised Real-World Object Detection
Semi-supervised object detection (SSOD) based on pseudo-labeling significantly reduces dependence on large labeled datasets by effectively leveraging both labeled and unlabeled data. However, real-world applications of SSOD often face critical challenges, including class imbalance, label noise, and labeling errors. We present an in-depth analysis of SSOD under real-world conditions, uncovering causes of suboptimal pseudo-labeling and key trade-offs between label quality and quantity. Based on our findings, we propose four building blocks that can be seamlessly integrated into an SSOD framework. Rare Class Collage (RCC): a data augmentation method that enhances the representation of rare classes by creating collages of rare objects. Rare Class Focus (RCF): a stratified batch sampling strategy that ensures a more balanced representation of all classes during training. Ground Truth Label Correction (GLC): a label refinement method that identifies and corrects false, missing, and noisy ground truth labels by leveraging the consistency of teacher model predictions. Pseudo-Label Selection (PLS): a selection method for removing low-quality pseudo-labeled images, guided by a novel metric estimating the missing detection rate while accounting for class rarity. We validate our methods through comprehensive experiments on autonomous driving datasets, resulting in up to 6% increase in SSOD performance. Overall, our investigation and novel, data-centric, and broadly applicable building blocks enable robust and effective SSOD in complex, real-world scenarios. Code is available at https://mos-ks.github.io/publications.
comment: Accepted to Transactions on Machine Learning Research (TMLR). OpenReview: https://openreview.net/forum?id=vRYt8QLKqK
☆ xKV: Cross-Layer SVD for KV-Cache Compression
Large Language Models (LLMs) with long context windows enable powerful applications but come at the cost of high memory consumption to store the Key and Value states (KV-Cache). Recent studies attempted to merge KV-cache from multiple layers into shared representations, yet these approaches either require expensive pretraining or rely on assumptions of high per-token cosine similarity across layers which generally does not hold in practice. We find that the dominant singular vectors are remarkably well-aligned across multiple layers of the KV-Cache. Exploiting this insight, we propose xKV, a simple post-training method that applies Singular Value Decomposition (SVD) on the KV-Cache of grouped layers. xKV consolidates the KV-Cache of multiple layers into a shared low-rank subspace, significantly reducing KV-Cache sizes. Through extensive evaluations on the RULER long-context benchmark with widely-used LLMs (e.g., Llama-3.1 and Qwen2.5), xKV achieves up to 6.8x higher compression rates than state-of-the-art inter-layer technique while improving accuracy by 2.7%. Moreover, xKV is compatible with the emerging Multi-Head Latent Attention (MLA) (e.g., DeepSeek-Coder-V2), yielding a notable 3x compression rates on coding tasks without performance degradation. These results highlight xKV's strong capability and versatility in addressing memory bottlenecks for long-context LLM inference. Our code is publicly available at: https://github.com/abdelfattah-lab/xKV.
☆ SimpleRL-Zoo: Investigating and Taming Zero Reinforcement Learning for Open Base Models in the Wild
DeepSeek-R1 has shown that long chain-of-thought (CoT) reasoning can naturally emerge through a simple reinforcement learning (RL) framework with rule-based rewards, where the training may directly start from the base models-a paradigm referred to as zero RL training. Most recent efforts to reproduce zero RL training have primarily focused on the Qwen2.5 model series, which may not be representative as we find the base models already exhibit strong instruction-following and self-reflection abilities. In this work, we investigate zero RL training across 10 diverse base models, spanning different families and sizes including LLama3-8B, Mistral-7B/24B, DeepSeek-Math-7B, Qwen2.5-math-7B, and all Qwen2.5 models from 0.5B to 32B. Leveraging several key design strategies-such as adjusting format reward and controlling query difficulty-we achieve substantial improvements in both reasoning accuracy and response length across most settings. However, by carefully monitoring the training dynamics, we observe that different base models exhibit distinct patterns during training. For instance, the increased response length does not always correlate with the emergence of certain cognitive behaviors such as verification (i.e., the "aha moment"). Notably, we observe the "aha moment" for the first time in small models not from the Qwen family. We share the key designs that enable successful zero RL training, along with our findings and practices. To facilitate further research, we open-source the code, models, and analysis tools.
☆ A semantic communication-based workload-adjustable transceiver for wireless AI-generated content (AIGC) delivery
With the significant advances in generative AI (GAI) and the proliferation of mobile devices, providing high-quality AI-generated content (AIGC) services via wireless networks is becoming the future direction. However, the primary challenges of AIGC service delivery in wireless networks lie in unstable channels, limited bandwidth resources, and unevenly distributed computational resources. In this paper, we employ semantic communication (SemCom) in diffusion-based GAI models to propose a Resource-aware wOrkload-adjUstable TransceivEr (ROUTE) for AIGC delivery in dynamic wireless networks. Specifically, to relieve the communication resource bottleneck, SemCom is utilized to prioritize semantic information of the generated content. Then, to improve computational resource utilization in both edge and local and reduce AIGC semantic distortion in transmission, modified diffusion-based models are applied to adjust the computing workload and semantic density in cooperative content generation. Simulations verify the superiority of our proposed ROUTE in terms of latency and content quality compared to conventional AIGC approaches.
☆ Bootstrapped Model Predictive Control ICLR 2025
Model Predictive Control (MPC) has been demonstrated to be effective in continuous control tasks. When a world model and a value function are available, planning a sequence of actions ahead of time leads to a better policy. Existing methods typically obtain the value function and the corresponding policy in a model-free manner. However, we find that such an approach struggles with complex tasks, resulting in poor policy learning and inaccurate value estimation. To address this problem, we leverage the strengths of MPC itself. In this work, we introduce Bootstrapped Model Predictive Control (BMPC), a novel algorithm that performs policy learning in a bootstrapped manner. BMPC learns a network policy by imitating an MPC expert, and in turn, uses this policy to guide the MPC process. Combined with model-based TD-learning, our policy learning yields better value estimation and further boosts the efficiency of MPC. We also introduce a lazy reanalyze mechanism, which enables computationally efficient imitation learning. Our method achieves superior performance over prior works on diverse continuous control tasks. In particular, on challenging high-dimensional locomotion tasks, BMPC significantly improves data efficiency while also enhancing asymptotic performance and training stability, with comparable training time and smaller network sizes. Code is available at https://github.com/wertyuilife2/bmpc.
comment: Published as a conference paper at ICLR 2025
☆ Reasoning to Learn from Latent Thoughts
Compute scaling for language model (LM) pretraining has outpaced the growth of human-written texts, leading to concerns that data will become the bottleneck to LM scaling. To continue scaling pretraining in this data-constrained regime, we propose that explicitly modeling and inferring the latent thoughts that underlie the text generation process can significantly improve pretraining data efficiency. Intuitively, our approach views web text as the compressed final outcome of a verbose human thought process and that the latent thoughts contain important contextual knowledge and reasoning steps that are critical to data-efficient learning. We empirically demonstrate the effectiveness of our approach through data-constrained continued pretraining for math. We first show that synthetic data approaches to inferring latent thoughts significantly improve data efficiency, outperforming training on the same amount of raw data (5.7\% $\rightarrow$ 25.4\% on MATH). Furthermore, we demonstrate latent thought inference without a strong teacher, where an LM bootstraps its own performance by using an EM algorithm to iteratively improve the capability of the trained LM and the quality of thought-augmented pretraining data. We show that a 1B LM can bootstrap its performance across at least three iterations and significantly outperform baselines trained on raw data, with increasing gains from additional inference compute when performing the E-step. The gains from inference scaling and EM iterations suggest new opportunities for scaling data-constrained pretraining.
☆ Exploring the Integration of Key-Value Attention Into Pure and Hybrid Transformers for Semantic Segmentation
While CNNs were long considered state of the art for image processing, the introduction of Transformer architectures has challenged this position. While achieving excellent results in image classification and segmentation, Transformers remain inherently reliant on large training datasets and remain computationally expensive. A newly introduced Transformer derivative named KV Transformer shows promising results in synthetic, NLP, and image classification tasks, while reducing complexity and memory usage. This is especially conducive to use cases where local inference is required, such as medical screening applications. We endeavoured to further evaluate the merit of KV Transformers on semantic segmentation tasks, specifically in the domain of medical imaging. By directly comparing traditional and KV variants of the same base architectures, we provide further insight into the practical tradeoffs of reduced model complexity. We observe a notable reduction in parameter count and multiply accumulate operations, while achieving similar performance from most of the KV variant models when directly compared to their QKV implementation.
comment: 6 pages, 3 figures, Preprint. Final version published in: Bildverarbeitung f\"ur die Medizin 2025, Springer. DOI: https://doi.org/10.1007/978-3-658-47422-5_71
☆ MODIS: Multi-Omics Data Integration for Small and Unpaired Datasets
A key challenge today lies in the ability to efficiently handle multi-omics data since such multimodal data may provide a more comprehensive overview of the underlying processes in a system. Yet it comes with challenges: multi-omics data are most often unpaired and only partially labeled, moreover only small amounts of data are available in some situation such as rare diseases. We propose MODIS which stands for Multi-Omics Data Integration for Small and unpaired datasets, a semi supervised approach to account for these particular settings. MODIS learns a probabilistic coupling of heterogeneous data modalities and learns a shared latent space where modalities are aligned. We rely on artificial data to build controlled experiments to explore how much supervision is needed for an accurate alignment of modalities, and how our approach enables dealing with new conditions for which few data are available. The code is available athttps://github.com/VILLOUTREIXLab/MODIS.
☆ Self-Organizing Graph Reasoning Evolves into a Critical State for Continuous Discovery Through Structural-Semantic Dynamics
We report fundamental insights into how agentic graph reasoning systems spontaneously evolve toward a critical state that sustains continuous semantic discovery. By rigorously analyzing structural (Von Neumann graph entropy) and semantic (embedding) entropy, we identify a subtle yet robust regime in which semantic entropy persistently dominates over structural entropy. This interplay is quantified by a dimensionless Critical Discovery Parameter that stabilizes at a small negative value, indicating a consistent excess of semantic entropy. Empirically, we observe a stable fraction (12%) of "surprising" edges, links between semantically distant concepts, providing evidence of long-range or cross-domain connections that drive continuous innovation. Concomitantly, the system exhibits scale-free and small-world topological features, alongside a negative cross-correlation between structural and semantic measures, reinforcing the analogy to self-organized criticality. These results establish clear parallels with critical phenomena in physical, biological, and cognitive complex systems, revealing an entropy-based principle governing adaptability and continuous innovation. Crucially, semantic richness emerges as the underlying driver of sustained exploration, despite not being explicitly used by the reasoning process. Our findings provide interdisciplinary insights and practical strategies for engineering intelligent systems with intrinsic capacities for long-term discovery and adaptation, and offer insights into how model training strategies can be developed that reinforce critical discovery.
☆ Unsupervised Detection of Fraudulent Transactions in E-commerce Using Contrastive Learning
With the rapid development of e-commerce, e-commerce platforms are facing an increasing number of fraud threats. Effectively identifying and preventing these fraudulent activities has become a critical research problem. Traditional fraud detection methods typically rely on supervised learning, which requires large amounts of labeled data. However, such data is often difficult to obtain, and the continuous evolution of fraudulent activities further reduces the adaptability and effectiveness of traditional methods. To address this issue, this study proposes an unsupervised e-commerce fraud detection algorithm based on SimCLR. The algorithm leverages the contrastive learning framework to effectively detect fraud by learning the underlying representations of transaction data in an unlabeled setting. Experimental results on the eBay platform dataset show that the proposed algorithm outperforms traditional unsupervised methods such as K-means, Isolation Forest, and Autoencoders in terms of accuracy, precision, recall, and F1 score, demonstrating strong fraud detection capabilities. The results confirm that the SimCLR-based unsupervised fraud detection method has broad application prospects in e-commerce platform security, improving both detection accuracy and robustness. In the future, with the increasing scale and diversity of datasets, the model's performance will continue to improve, and it could be integrated with real-time monitoring systems to provide more efficient security for e-commerce platforms.
☆ Interpretable and Fair Mechanisms for Abstaining Classifiers ECML
Abstaining classifiers have the option to refrain from providing a prediction for instances that are difficult to classify. The abstention mechanism is designed to trade off the classifier's performance on the accepted data while ensuring a minimum number of predictions. In this setting, often fairness concerns arise when the abstention mechanism solely reduces errors for the majority groups of the data, resulting in increased performance differences across demographic groups. While there exist a bunch of methods that aim to reduce discrimination when abstaining, there is no mechanism that can do so in an explainable way. In this paper, we fill this gap by introducing Interpretable and Fair Abstaining Classifier IFAC, an algorithm that can reject predictions both based on their uncertainty and their unfairness. By rejecting possibly unfair predictions, our method reduces error and positive decision rate differences across demographic groups of the non-rejected data. Since the unfairness-based rejections are based on an interpretable-by-design method, i.e., rule-based fairness checks and situation testing, we create a transparent process that can empower human decision-makers to review the unfair predictions and make more just decisions for them. This explainable aspect is especially important in light of recent AI regulations, mandating that any high-risk decision task should be overseen by human experts to reduce discrimination risks.
comment: 25 pages, 8 figures. In: Machine Learning and Knowledge Discovery in Databases. Research Track. ECML PKDD 2024
☆ Classical Planning with LLM-Generated Heuristics: Challenging the State of the Art with Python Code
In recent years, large language models (LLMs) have shown remarkable capabilities in various artificial intelligence problems. However, they fail to plan reliably, even when prompted with a detailed definition of the planning task. Attempts to improve their planning capabilities, such as chain-of-thought prompting, fine-tuning, and explicit "reasoning" still yield incorrect plans and usually fail to generalize to larger tasks. In this paper, we show how to use LLMs to generate correct plans, even for out-of-distribution tasks of increasing size. For a given planning domain, we ask an LLM to generate several domain-dependent heuristic functions in the form of Python code, evaluate them on a set of training tasks within a greedy best-first search, and choose the strongest one. The resulting LLM-generated heuristics solve many more unseen test tasks than state-of-the-art domain-independent heuristics for classical planning. They are even competitive with the strongest learning algorithm for domain-dependent planning. These findings are especially remarkable given that our proof-of-concept implementation is based on an unoptimized Python planner and the baselines all build upon highly optimized C++ code. In some domains, the LLM-generated heuristics expand fewer states than the baselines, revealing that they are not only efficiently computable, but sometimes even more informative than the state-of-the-art heuristics. Overall, our results show that sampling a set of planning heuristic function programs can significantly improve the planning capabilities of LLMs.
☆ Streaming Federated Learning with Markovian Data
Federated learning (FL) is now recognized as a key framework for communication-efficient collaborative learning. Most theoretical and empirical studies, however, rely on the assumption that clients have access to pre-collected data sets, with limited investigation into scenarios where clients continuously collect data. In many real-world applications, particularly when data is generated by physical or biological processes, client data streams are often modeled by non-stationary Markov processes. Unlike standard i.i.d. sampling, the performance of FL with Markovian data streams remains poorly understood due to the statistical dependencies between client samples over time. In this paper, we investigate whether FL can still support collaborative learning with Markovian data streams. Specifically, we analyze the performance of Minibatch SGD, Local SGD, and a variant of Local SGD with momentum. We answer affirmatively under standard assumptions and smooth non-convex client objectives: the sample complexity is proportional to the inverse of the number of clients with a communication complexity comparable to the i.i.d. scenario. However, the sample complexity for Markovian data streams remains higher than for i.i.d. sampling.
comment: Work under review
☆ Sample-Efficient Reinforcement Learning of Koopman eNMPC
Reinforcement learning (RL) can be used to tune data-driven (economic) nonlinear model predictive controllers ((e)NMPCs) for optimal performance in a specific control task by optimizing the dynamic model or parameters in the policy's objective function or constraints, such as state bounds. However, the sample efficiency of RL is crucial, and to improve it, we combine a model-based RL algorithm with our published method that turns Koopman (e)NMPCs into automatically differentiable policies. We apply our approach to an eNMPC case study of a continuous stirred-tank reactor (CSTR) model from the literature. The approach outperforms benchmark methods, i.e., data-driven eNMPCs using models based on system identification without further RL tuning of the resulting policy, and neural network controllers trained with model-based RL, by achieving superior control performance and higher sample efficiency. Furthermore, utilizing partial prior knowledge about the system dynamics via physics-informed learning further increases sample efficiency.
comment: 24 pages, 9 figures, 2 tables
☆ The case for delegated AI autonomy for Human AI teaming in healthcare
In this paper we propose an advanced approach to integrating artificial intelligence (AI) into healthcare: autonomous decision support. This approach allows the AI algorithm to act autonomously for a subset of patient cases whilst serving a supportive role in other subsets of patient cases based on defined delegation criteria. By leveraging the complementary strengths of both humans and AI, it aims to deliver greater overall performance than existing human-AI teaming models. It ensures safe handling of patient cases and potentially reduces clinician review time, whilst being mindful of AI tool limitations. After setting the approach within the context of current human-AI teaming models, we outline the delegation criteria and apply them to a specific AI-based tool used in histopathology. The potential impact of the approach and the regulatory requirements for its successful implementation are then discussed.
☆ Mechanistic Interpretability of Fine-Tuned Vision Transformers on Distorted Images: Decoding Attention Head Behavior for Transparent and Trustworthy AI
Mechanistic interpretability improves the safety, reliability, and robustness of large AI models. This study examined individual attention heads in vision transformers (ViTs) fine tuned on distorted 2D spectrogram images containing non relevant content (axis labels, titles, color bars). By introducing extraneous features, the study analyzed how transformer components processed unrelated information, using mechanistic interpretability to debug issues and reveal insights into transformer architectures. Attention maps assessed head contributions across layers. Heads in early layers (1 to 3) showed minimal task impact with ablation increased MSE loss slightly ({\mu}=0.11%, {\sigma}=0.09%), indicating focus on less critical low level features. In contrast, deeper heads (e.g., layer 6) caused a threefold higher loss increase ({\mu}=0.34%, {\sigma}=0.02%), demonstrating greater task importance. Intermediate layers (6 to 11) exhibited monosemantic behavior, attending exclusively to chirp regions. Some early heads (1 to 4) were monosemantic but non task relevant (e.g. text detectors, edge or corner detectors). Attention maps distinguished monosemantic heads (precise chirp localization) from polysemantic heads (multiple irrelevant regions). These findings revealed functional specialization in ViTs, showing how heads processed relevant vs. extraneous information. By decomposing transformers into interpretable components, this work enhanced model understanding, identified vulnerabilities, and advanced safer, more transparent AI.
comment: 15 pages, 8 figures
☆ On the Optimality of Single-label and Multi-label Neural Network Decoders
We investigate the design of two neural network (NN) architectures recently proposed as decoders for forward error correction: the so-called single-label NN (SLNN) and multi-label NN (MLNN) decoders. These decoders have been reported to achieve near-optimal codeword- and bit-wise performance, respectively. Results in the literature show near-optimality for a variety of short codes. In this paper, we analytically prove that certain SLNN and MLNN architectures can, in fact, always realize optimal decoding, regardless of the code. These optimal architectures and their binary weights are shown to be defined by the codebook, i.e., no training or network optimization is required. Our proposed architectures are in fact not NNs, but a different way of implementing the maximum likelihood decoding rule. Optimal performance is numerically demonstrated for Hamming $(7,4)$, Polar $(16,8)$, and BCH $(31,21)$ codes. The results show that our optimal architectures are less complex than the SLNN and MLNN architectures proposed in the literature, which in fact only achieve near-optimal performance. Extension to longer codes is still hindered by the curse of dimensionality. Therefore, even though SLNN and MLNN can perform maximum likelihood decoding, such architectures cannot be used for medium and long codes.
comment: 8 pages, 8 figures
☆ EgoSurgery-HTS: A Dataset for Egocentric Hand-Tool Segmentation in Open Surgery Videos
Egocentric open-surgery videos capture rich, fine-grained details essential for accurately modeling surgical procedures and human behavior in the operating room. A detailed, pixel-level understanding of hands and surgical tools is crucial for interpreting a surgeon's actions and intentions. We introduce EgoSurgery-HTS, a new dataset with pixel-wise annotations and a benchmark suite for segmenting surgical tools, hands, and interacting tools in egocentric open-surgery videos. Specifically, we provide a labeled dataset for (1) tool instance segmentation of 14 distinct surgical tools, (2) hand instance segmentation, and (3) hand-tool segmentation to label hands and the tools they manipulate. Using EgoSurgery-HTS, we conduct extensive evaluations of state-of-the-art segmentation methods and demonstrate significant improvements in the accuracy of hand and hand-tool segmentation in egocentric open-surgery videos compared to existing datasets. The dataset will be released at https://github.com/Fujiry0/EgoSurgery.
☆ Simulation-Driven Balancing of Competitive Game Levels with Reinforcement Learning
The balancing process for game levels in competitive two-player contexts involves a lot of manual work and testing, particularly for non-symmetrical game levels. In this work, we frame game balancing as a procedural content generation task and propose an architecture for automatically balancing of tile-based levels within the PCGRL framework (procedural content generation via reinforcement learning). Our architecture is divided into three parts: (1) a level generator, (2) a balancing agent, and (3) a reward modeling simulation. Through repeated simulations, the balancing agent receives rewards for adjusting the level towards a given balancing objective, such as equal win rates for all players. To this end, we propose new swap-based representations to improve the robustness of playability, thereby enabling agents to balance game levels more effectively and quickly compared to traditional PCGRL. By analyzing the agent's swapping behavior, we can infer which tile types have the most impact on the balance. We validate our approach in the Neural MMO (NMMO) environment in a competitive two-player scenario. In this extended conference paper, we present improved results, explore the applicability of the method to various forms of balancing beyond equal balancing, compare the performance to another search-based approach, and discuss the application of existing fairness metrics to game balancing.
comment: Preprint of the journal (IEEE Transactions on Games) paper of the same name
☆ Thermalizer: Stable autoregressive neural emulation of spatiotemporal chaos
Autoregressive surrogate models (or \textit{emulators}) of spatiotemporal systems provide an avenue for fast, approximate predictions, with broad applications across science and engineering. At inference time, however, these models are generally unable to provide predictions over long time rollouts due to accumulation of errors leading to diverging trajectories. In essence, emulators operate out of distribution, and controlling the online distribution quickly becomes intractable in large-scale settings. To address this fundamental issue, and focusing on time-stationary systems admitting an invariant measure, we leverage diffusion models to obtain an implicit estimator of the score of this invariant measure. We show that this model of the score function can be used to stabilize autoregressive emulator rollouts by applying on-the-fly denoising during inference, a process we call \textit{thermalization}. Thermalizing an emulator rollout is shown to extend the time horizon of stable predictions by an order of magnitude in complex systems exhibiting turbulent and chaotic behavior, opening up a novel application of diffusion models in the context of neural emulation.
☆ Differentially Private Joint Independence Test
Identification of joint dependence among more than two random vectors plays an important role in many statistical applications, where the data may contain sensitive or confidential information. In this paper, we consider the the d-variable Hilbert-Schmidt independence criterion (dHSIC) in the context of differential privacy. Given the limiting distribution of the empirical estimate of dHSIC is complicated Gaussian chaos, constructing tests in the non-privacy regime is typically based on permutation and bootstrap. To detect joint dependence in privacy, we propose a dHSIC-based testing procedure by employing a differentially private permutation methodology. Our method enjoys privacy guarantee, valid level and pointwise consistency, while the bootstrap counterpart suffers inconsistent power. We further investigate the uniform power of the proposed test in dHSIC metric and $L_2$ metric, indicating that the proposed test attains the minimax optimal power across different privacy regimes. As a byproduct, our results also contain the pointwise and uniform power of the non-private permutation dHSIC, addressing an unsolved question remained in Pfister et al. (2018).
comment: 51pages
☆ Accenture-NVS1: A Novel View Synthesis Dataset
This paper introduces ACC-NVS1, a specialized dataset designed for research on Novel View Synthesis specifically for airborne and ground imagery. Data for ACC-NVS1 was collected in Austin, TX and Pittsburgh, PA in 2023 and 2024. The collection encompasses six diverse real-world scenes captured from both airborne and ground cameras, resulting in a total of 148,000 images. ACC-NVS1 addresses challenges such as varying altitudes and transient objects. This dataset is intended to supplement existing datasets, providing additional resources for comprehensive research, rather than serving as a benchmark.
comment: 6 pages, 7 figures
☆ Energy-Efficient Dynamic Training and Inference for GNN-Based Network Modeling
Efficient network modeling is essential for resource optimization and network planning in next-generation large-scale complex networks. Traditional approaches, such as queuing theory-based modeling and packet-based simulators, can be inefficient due to the assumption made and the computational expense, respectively. To address these challenges, we propose an innovative energy-efficient dynamic orchestration of Graph Neural Networks (GNN) based model training and inference framework for context-aware network modeling and predictions. We have developed a low-complexity solution framework, QAG, that is a Quantum approximation optimization (QAO) algorithm for Adaptive orchestration of GNN-based network modeling. We leverage the tripartite graph model to represent a multi-application system with many compute nodes. Thereafter, we apply the constrained graph-cutting using QAO to find the feasible energy-efficient configurations of the GNN-based model and deploying them on the available compute nodes to meet the network modeling application requirements. The proposed QAG scheme closely matches the optimum and offers atleast a 50% energy saving while meeting the application requirements with 60% lower churn-rate.
comment: Accepted in IEEE WCNC 2025
☆ OCRT: Boosting Foundation Models in the Open World with Object-Concept-Relation Triad CVPR 2025
Although foundation models (FMs) claim to be powerful, their generalization ability significantly decreases when faced with distribution shifts, weak supervision, or malicious attacks in the open world. On the other hand, most domain generalization or adversarial fine-tuning methods are task-related or model-specific, ignoring the universality in practical applications and the transferability between FMs. This paper delves into the problem of generalizing FMs to the out-of-domain data. We propose a novel framework, the Object-Concept-Relation Triad (OCRT), that enables FMs to extract sparse, high-level concepts and intricate relational structures from raw visual inputs. The key idea is to bind objects in visual scenes and a set of object-centric representations through unsupervised decoupling and iterative refinement. To be specific, we project the object-centric representations onto a semantic concept space that the model can readily interpret and estimate their importance to filter out irrelevant elements. Then, a concept-based graph, which has a flexible degree, is constructed to incorporate the set of concepts and their corresponding importance, enabling the extraction of high-order factors from informative concepts and facilitating relational reasoning among these concepts. Extensive experiments demonstrate that OCRT can substantially boost the generalizability and robustness of SAM and CLIP across multiple downstream tasks.
comment: Accepted by CVPR 2025
☆ TARDIS: Mitigate Temporal Misalignment via Representation Steering
Language models often struggle with temporal misalignment, performance degradation caused by shifts in the temporal distribution of data. Continuously updating models to avoid degradation is expensive. Can models be adapted without updating model weights? We present TARDIS, an unsupervised representation editing method that addresses this challenge. TARDIS extracts steering vectors from unlabeled data and adjusts the model's representations to better align with the target time period's distribution. Our experiments reveal that TARDIS enhances downstream task performance without the need for fine-tuning, can mitigate temporal misalignment even when exact target time period data is unavailable, and remains efficient even when the temporal information of the target data points is unknown at inference time.
☆ Feature Qualification by Deep Nets: A Constructive Approach
The great success of deep learning has stimulated avid research activities in verifying the power of depth in theory, a common consensus of which is that deep net are versatile in approximating and learning numerous functions. Such a versatility certainly enhances the understanding of the power of depth, but makes it difficult to judge which data features are crucial in a specific learning task. This paper proposes a constructive approach to equip deep nets for the feature qualification purpose. Using the product-gate nature and localized approximation property of deep nets with sigmoid activation (deep sigmoid nets), we succeed in constructing a linear deep net operator that possesses optimal approximation performance in approximating smooth and radial functions. Furthermore, we provide theoretical evidences that the constructed deep net operator is capable of qualifying multiple features such as the smoothness and radialness of the target functions.
☆ Geometric Preference Elicitation for Minimax Regret Optimization in Uncertainty Matroids
This paper presents an efficient preference elicitation framework for uncertain matroid optimization, where precise weight information is unavailable, but insights into possible weight values are accessible. The core innovation of our approach lies in its ability to systematically elicit user preferences, aligning the optimization process more closely with decision-makers' objectives. By incrementally querying preferences between pairs of elements, we iteratively refine the parametric uncertainty regions, leveraging the structural properties of matroids. Our method aims to achieve the exact optimum by reducing regret with a few elicitation rounds. Additionally, our approach avoids the computation of Minimax Regret and the use of Linear programming solvers at every iteration, unlike previous methods. Experimental results on four standard matroids demonstrate that our method reaches optimality more quickly and with fewer preference queries than existing techniques.
☆ Adaptive Machine Learning for Resource-Constrained Environments KDD 2024
The Internet of Things is an example domain where data is perpetually generated in ever-increasing quantities, reflecting the proliferation of connected devices and the formation of continuous data streams over time. Consequently, the demand for ad-hoc, cost-effective machine learning solutions must adapt to this evolving data influx. This study tackles the task of offloading in small gateways, exacerbated by their dynamic availability over time. An approach leveraging CPU utilization metrics using online and continual machine learning techniques is proposed to predict gateway availability. These methods are compared to popular machine learning algorithms and a recent time-series foundation model, Lag-Llama, for fine-tuned and zero-shot setups. Their performance is benchmarked on a dataset of CPU utilization measurements over time from an IoT gateway and focuses on model metrics such as prediction errors, training and inference times, and memory consumption. Our primary objective is to study new efficient ways to predict CPU performance in IoT environments. Across various scenarios, our findings highlight that ensemble and online methods offer promising results for this task in terms of accuracy while maintaining a low resource footprint.
comment: 17 pages, 11 figures, accepted at DELTA 2024 (Workshop on Discovering Drift Phenomena in Evolving Landscapes), co-located with ACM SIGKDD 2024. This preprint has not undergone peer review. The Version of Record is available at https://doi.org/10.1007/978-3-031-82346-6_1
☆ Scaling Laws for Emulation of Stellar Spectra
Neural network-based emulators for the inference of stellar parameters and elemental abundances represent an increasingly popular methodology in modern spectroscopic surveys. However, these approaches are often constrained by their emulation precision and domain transfer capabilities. Greater generalizability has previously been achieved only with significantly larger model architectures, as demonstrated by Transformer-based models in natural language processing. This observation aligns with neural scaling laws, where model performance predictably improves with increased model size, computational resources allocated to model training, and training data volume. In this study, we demonstrate that these scaling laws also apply to Transformer-based spectral emulators in astronomy. Building upon our previous work with TransformerPayne and incorporating Maximum Update Parametrization techniques from natural language models, we provide training guidelines for scaling models to achieve optimal performance. Our results show that within the explored parameter space, clear scaling relationships emerge. These findings suggest that optimal computational resource allocation requires balanced scaling. Specifically, given a tenfold increase in training compute, achieving an optimal seven-fold reduction in mean squared error necessitates an approximately 2.5-fold increase in dataset size and a 3.8-fold increase in model size. This study establishes a foundation for developing spectral foundational models with enhanced domain transfer capabilities.
comment: 25 pages, 11 figures, submitted to OJA
☆ Adventurer: Exploration with BiGAN for Deep Reinforcement Learning
Recent developments in deep reinforcement learning have been very successful in learning complex, previously intractable problems. Sample efficiency and local optimality, however, remain significant challenges. To address these challenges, novelty-driven exploration strategies have emerged and shown promising potential. Unfortunately, no single algorithm outperforms all others in all tasks and most of them struggle with tasks with high-dimensional and complex observations. In this work, we propose Adventurer, a novelty-driven exploration algorithm that is based on Bidirectional Generative Adversarial Networks (BiGAN), where BiGAN is trained to estimate state novelty. Intuitively, a generator that has been trained on the distribution of visited states should only be able to generate a state coming from the distribution of visited states. As a result, novel states using the generator to reconstruct input states from certain latent representations would lead to larger reconstruction errors. We show that BiGAN performs well in estimating state novelty for complex observations. This novelty estimation method can be combined with intrinsic-reward-based exploration. Our empirical results show that Adventurer produces competitive results on a range of popular benchmark tasks, including continuous robotic manipulation tasks (e.g. Mujoco robotics) and high-dimensional image-based tasks (e.g. Atari games).
comment: Accepted at Applied Intelligence
☆ AutoBayes: A Compositional Framework for Generalized Variational Inference
We introduce a new compositional framework for generalized variational inference, clarifying the different parts of a model, how they interact, and how they compose. We explain that both exact Bayesian inference and the loss functions typical of variational inference (such as variational free energy and its generalizations) satisfy chain rules akin to that of reverse-mode automatic differentiation, and we advocate for exploiting this to build and optimize models accordingly. To this end, we construct a series of compositional tools: for building models; for constructing their inversions; for attaching local loss functions; and for exposing parameters. Finally, we explain how the resulting parameterized statistical games may be optimized locally, too. We illustrate our framework with a number of classic examples, pointing to new areas of extensibility that are revealed.
comment: 15 pages
☆ Reinforcement Learning in Switching Non-Stationary Markov Decision Processes: Algorithms and Convergence Analysis
Reinforcement learning in non-stationary environments is challenging due to abrupt and unpredictable changes in dynamics, often causing traditional algorithms to fail to converge. However, in many real-world cases, non-stationarity has some structure that can be exploited to develop algorithms and facilitate theoretical analysis. We introduce one such structure, Switching Non-Stationary Markov Decision Processes (SNS-MDP), where environments switch over time based on an underlying Markov chain. Under a fixed policy, the value function of an SNS-MDP admits a closed-form solution determined by the Markov chain's statistical properties, and despite the inherent non-stationarity, Temporal Difference (TD) learning methods still converge to the correct value function. Furthermore, policy improvement can be performed, and it is shown that policy iteration converges to the optimal policy. Moreover, since Q-learning converges to the optimal Q-function, it likewise yields the corresponding optimal policy. To illustrate the practical advantages of SNS-MDPs, we present an example in communication networks where channel noise follows a Markovian pattern, demonstrating how this framework can effectively guide decision-making in complex, time-varying contexts.
☆ Oaken: Fast and Efficient LLM Serving with Online-Offline Hybrid KV Cache Quantization
Modern Large Language Model serving system batches multiple requests to achieve high throughput, while batching attention operations is challenging, rendering memory bandwidth a critical bottleneck. The community relies on high-end GPUs with multiple high-bandwidth memory channels. Unfortunately, HBM's high bandwidth often comes at the expense of limited memory capacity, which reduces core utilization and increases costs. Recent advancements enabling longer contexts for LLMs have substantially increased the key-value cache size, further intensifying the pressures on memory capacity. The literature has explored KV cache quantization techniques, which commonly use low bitwidth for most values, selectively using higher bitwidth for outlier values. While this approach helps achieve high accuracy and low bitwidth simultaneously, it comes with the limitation that cost for online outlier detection is excessively high, negating the advantages. We propose Oaken, an acceleration solution that achieves high accuracy and high performance simultaneously through co-designing algorithm and hardware. To effectively find a sweet spot in the accuracy-performance trade-off space of KV cache quantization, Oaken employs an online-offline hybrid approach, setting outlier thresholds offline, which are then used to determine the quantization scale online. To translate the proposed algorithmic technique into tangible performance gains, Oaken also comes with custom quantization engines and memory management units that can be integrated with any LLM accelerators. We built an Oaken accelerator on top of an LLM accelerator, LPU, and conducted a comprehensive evaluation. Our experiments show that for a batch size of 256, Oaken achieves up to 1.58x throughput improvement over NVIDIA A100 GPU, incurring a minimal accuracy loss of only 0.54\% on average, compared to state-of-the-art KV cache quantization techniques.
comment: 15 pages, 14 figures, and 4 tables
☆ Adaptive Unimodal Regulation for Balanced Multimodal Information Acquisition CVPR2025
Sensory training during the early ages is vital for human development. Inspired by this cognitive phenomenon, we observe that the early training stage is also important for the multimodal learning process, where dataset information is rapidly acquired. We refer to this stage as the prime learning window. However, based on our observation, this prime learning window in multimodal learning is often dominated by information-sufficient modalities, which in turn suppresses the information acquisition of information-insufficient modalities. To address this issue, we propose Information Acquisition Regulation (InfoReg), a method designed to balance information acquisition among modalities. Specifically, InfoReg slows down the information acquisition process of information-sufficient modalities during the prime learning window, which could promote information acquisition of information-insufficient modalities. This regulation enables a more balanced learning process and improves the overall performance of the multimodal network. Experiments show that InfoReg outperforms related multimodal imbalanced methods across various datasets, achieving superior model performance. The code is available at https://github.com/GeWu-Lab/InfoReg_CVPR2025.
comment: 10pages, 16 figures, CVPR2025
☆ The Role of Artificial Intelligence in Enhancing Insulin Recommendations and Therapy Outcomes
The growing worldwide incidence of diabetes requires more effective approaches for managing blood glucose levels. Insulin delivery systems have advanced significantly, with artificial intelligence (AI) playing a key role in improving their precision and adaptability. AI algorithms, particularly those based on reinforcement learning, allow for personalised insulin dosing by continuously adapting to an individual's responses. Despite these advancements, challenges such as data privacy, algorithm transparency, and accessibility still need to be addressed. Continued progress and validation in AI-driven insulin delivery systems promise to improve therapy outcomes further, offering people more effective and individualised management of their diabetes. This paper presents an overview of current strategies, key challenges, and future directions.
☆ A Universal Model Combining Differential Equations and Neural Networks for Ball Trajectory Prediction
This paper presents a data driven universal ball trajectory prediction method integrated with physics equations. Existing methods are designed for specific ball types and struggle to generalize. This challenge arises from three key factors. First, learning-based models require large datasets but suffer from accuracy drops in unseen scenarios. Second, physics-based models rely on complex formulas and detailed inputs, yet accurately obtaining ball states, such as spin, is often impractical. Third, integrating physical principles with neural networks to achieve high accuracy, fast inference, and strong generalization remains difficult. To address these issues, we propose an innovative approach that incorporates physics-based equations and neural networks. We first derive three generalized physical formulas. Then, using a neural network and observed trajectory points, we infer certain parameters while fitting the remaining ones. These formulas enable precise trajectory prediction with minimal training data: only a few dozen samples. Extensive experiments demonstrate our method superiority in generalization, real-time performance, and accuracy.
☆ Galaxy Walker: Geometry-aware VLMs For Galaxy-scale Understanding
Modern vision-language models (VLMs) develop patch embedding and convolution backbone within vector space, especially Euclidean ones, at the very founding. When expanding VLMs to a galaxy scale for understanding astronomical phenomena, the integration of spherical space for planetary orbits and hyperbolic spaces for black holes raises two formidable challenges. a) The current pre-training model is confined to Euclidean space rather than a comprehensive geometric embedding. b) The predominant architecture lacks suitable backbones for anisotropic physical geometries. In this paper, we introduced Galaxy-Walker, a geometry-aware VLM, for the universe-level vision understanding tasks. We proposed the geometry prompt that generates geometry tokens by random walks across diverse spaces on a multi-scale physical graph, along with a geometry adapter that compresses and reshapes the space anisotropy in a mixture-of-experts manner. Extensive experiments demonstrate the effectiveness of our approach, with Galaxy-Walker achieving state-of-the-art performance in both galaxy property estimation ($R^2$ scores up to $0.91$) and morphology classification tasks (up to $+0.17$ F1 improvement in challenging features), significantly outperforming both domain-specific models and general-purpose VLMs.
☆ Parametric Dynamic Mode Decomposition with multi-linear interpolation for prediction of thermal fields of Al2O3-water nanofluid flows at unseen parameters
The study proposes a data-driven model which combines the Dynamic Mode Decomposition with multi-linear interpolation to predict the thermal fields of nanofluid flows at unseen Reynolds numbers (Re) and particle volume concentrations ($\epsilon$). The flow, considered for the study, is laminar and incompressible. The study employs an in-house Fortran-based solver to predict the thermal fields of Al$_2$O$_3$-water nanofluid flow through a two-dimensional rectangular channel, with the bottom wall subjected to a uniform heat flux. The performance of two models operating in one- and two-dimensional parametric spaces are investigated. Initially, a DMD with linear interpolation (DMD-LI) based solver is used for prediction of temperature of the nanofluid at any Re $>$ 100. The DMD-LI based model, predicts temperature fields with a maximum percentage difference of just 0.0273\%, in comparison with the CFD-based solver at Re =960, and $\epsilon$ = 1.0\%. The corresponding difference in the average Nusselt numbers is only 0.39\%. Following that a DMD with bi-linear interpolation (DMD-BLI) based solver is used for prediction of temperature of the nanofluid at any Re $>$ 100 and $\epsilon$ $>$ 0.5\%. The performance of two different ways of stacking the data are also examined. When compared to the CFD-based model, the DMD-BLI-based model predicts the temperature fields with a maximum percentage difference of 0.21 \%, at Re = 800 and $\epsilon$ = 1.35\%. And the corresponding percentage difference in the average Nusselt number prediction is only 6.08\%. All the results are reported in detail. Along side the important conclusions, the future scope of the study is also listed.
comment: 19 pages, 19 figures
☆ Anchor-based oversampling for imbalanced tabular data via contrastive and adversarial learning
Imbalanced data represent a distribution with more frequencies of one class (majority) than the other (minority). This phenomenon occurs across various domains, such as security, medical care and human activity. In imbalanced learning, classification algorithms are typically inclined to classify the majority class accurately, resulting in artificially high accuracy rates. As a result, many minority samples are mistakenly labelled as majority-class instances, resulting in a bias that benefits the majority class. This study presents a framework based on boundary anchor samples to tackle the imbalance learning challenge. First, we select and use anchor samples to train a multilayer perceptron (MLP) classifier, which acts as a prior knowledge model and aids the adversarial and contrastive learning procedures. Then, we designed a novel deep generative model called Anchor Stabilized Conditional Generative Adversarial Network or Anch-SCGAN in short. Anch-SCGAN is supported with two generators for the minority and majority classes and a discriminator incorporating additional class-specific information from the pre-trained feature extractor MLP. In addition, we facilitate the generator's training procedure in two ways. First, we define a new generator loss function based on reprocessed anchor samples and contrastive learning. Second, we apply a scoring strategy to stabilize the adversarial training part in generators. We train Anch-SCGAN and further finetune it with anchor samples to improve the precision of the generated samples. Our experiments on 16 real-world imbalanced datasets illustrate that Anch-SCGAN outperforms the renowned methods in imbalanced learning.
☆ Distil-xLSTM: Learning Attention Mechanisms through Recurrent Structures
The current era of Natural Language Processing (NLP) is dominated by Transformer models. However, novel architectures relying on recurrent mechanisms, such as xLSTM and Mamba, have been proposed as alternatives to attention-based models. Although computation is done differently than with the attention mechanism mechanism, these recurrent models yield good results and sometimes even outperform state-of-the-art attention-based models. In this work, we propose Distil-xLSTM, an xLSTM-based Small Language Model (SLM) trained by distilling knowledge from a Large Language Model (LLM) that shows promising results while being compute and scale efficient. Our Distil-xLSTM focuses on approximating a transformer-based model attention parametrization using its recurrent sequence mixing components and shows good results with minimal training.
☆ Self-Reported Confidence of Large Language Models in Gastroenterology: Analysis of Commercial, Open-Source, and Quantized Models
This study evaluated self-reported response certainty across several large language models (GPT, Claude, Llama, Phi, Mistral, Gemini, Gemma, and Qwen) using 300 gastroenterology board-style questions. The highest-performing models (GPT-o1 preview, GPT-4o, and Claude-3.5-Sonnet) achieved Brier scores of 0.15-0.2 and AUROC of 0.6. Although newer models demonstrated improved performance, all exhibited a consistent tendency towards overconfidence. Uncertainty estimation presents a significant challenge to the safe use of LLMs in healthcare. Keywords: Large Language Models; Confidence Elicitation; Artificial Intelligence; Gastroenterology; Uncertainty Quantification
comment: 35 pages, 5 figures, 1 table, 7 supplementary figures
☆ Discriminative protein sequence modelling with Latent Space Diffusion
We explore a framework for protein sequence representation learning that decomposes the task between manifold learning and distributional modelling. Specifically we present a Latent Space Diffusion architecture which combines a protein sequence autoencoder with a denoising diffusion model operating on its latent space. We obtain a one-parameter family of learned representations from the diffusion model, along with the autoencoder's latent representation. We propose and evaluate two autoencoder architectures: a homogeneous model forcing amino acids of the same type to be identically distributed in the latent space, and an inhomogeneous model employing a noise-based variant of masking. As a baseline we take a latent space learned by masked language modelling, and evaluate discriminative capability on a range of protein property prediction tasks. Our finding is twofold: the diffusion models trained on both our proposed variants display higher discriminative power than the one trained on the masked language model baseline, none of the diffusion representations achieve the performance of the masked language model embeddings themselves.
☆ RLCAD: Reinforcement Learning Training Gym for Revolution Involved CAD Command Sequence Generation
A CAD command sequence is a typical parametric design paradigm in 3D CAD systems where a model is constructed by overlaying 2D sketches with operations such as extrusion, revolution, and Boolean operations. Although there is growing academic interest in the automatic generation of command sequences, existing methods and datasets only support operations such as 2D sketching, extrusion,and Boolean operations. This limitation makes it challenging to represent more complex geometries. In this paper, we present a reinforcement learning (RL) training environment (gym) built on a CAD geometric engine. Given an input boundary representation (B-Rep) geometry, the policy network in the RL algorithm generates an action. This action, along with previously generated actions, is processed within the gym to produce the corresponding CAD geometry, which is then fed back into the policy network. The rewards, determined by the difference between the generated and target geometries within the gym, are used to update the RL network. Our method supports operations beyond sketches, Boolean, and extrusion, including revolution operations. With this training gym, we achieve state-of-the-art (SOTA) quality in generating command sequences from B-Rep geometries. In addition, our method can significantly improve the efficiency of command sequence generation by a factor of 39X compared with the previous training gym.
☆ Natural Language Processing for Electronic Health Records in Scandinavian Languages: Norwegian, Swedish, and Danish
Background: Clinical natural language processing (NLP) refers to the use of computational methods for extracting, processing, and analyzing unstructured clinical text data, and holds a huge potential to transform healthcare in various clinical tasks. Objective: The study aims to perform a systematic review to comprehensively assess and analyze the state-of-the-art NLP methods for the mainland Scandinavian clinical text. Method: A literature search was conducted in various online databases including PubMed, ScienceDirect, Google Scholar, ACM digital library, and IEEE Xplore between December 2022 and February 2024. Further, relevant references to the included articles were also used to solidify our search. The final pool includes articles that conducted clinical NLP in the mainland Scandinavian languages and were published in English between 2010 and 2024. Results: Out of the 113 articles, 18% (n=21) focus on Norwegian clinical text, 64% (n=72) on Swedish, 10% (n=11) on Danish, and 8% (n=9) focus on more than one language. Generally, the review identified positive developments across the region despite some observable gaps and disparities between the languages. There are substantial disparities in the level of adoption of transformer-based models. In essential tasks such as de-identification, there is significantly less research activity focusing on Norwegian and Danish compared to Swedish text. Further, the review identified a low level of sharing resources such as data, experimentation code, pre-trained models, and rate of adaptation and transfer learning in the region. Conclusion: The review presented a comprehensive assessment of the state-of-the-art Clinical NLP for electronic health records (EHR) text in mainland Scandinavian languages and, highlighted the potential barriers and challenges that hinder the rapid advancement of the field in the region.
comment: 45 pages including the appendix, 9 figures in the main manuscript and 11 figures in the Appendix
☆ Parental Guidance: Efficient Lifelong Learning through Evolutionary Distillation
Developing robotic agents that can perform well in diverse environments while showing a variety of behaviors is a key challenge in AI and robotics. Traditional reinforcement learning (RL) methods often create agents that specialize in narrow tasks, limiting their adaptability and diversity. To overcome this, we propose a preliminary, evolution-inspired framework that includes a reproduction module, similar to natural species reproduction, balancing diversity and specialization. By integrating RL, imitation learning (IL), and a coevolutionary agent-terrain curriculum, our system evolves agents continuously through complex tasks. This approach promotes adaptability, inheritance of useful traits, and continual learning. Agents not only refine inherited skills but also surpass their predecessors. Our initial experiments show that this method improves exploration efficiency and supports open-ended learning, offering a scalable solution where sparse reward coupled with diverse terrain environments induces a multi-task setting.
comment: 4 pages, 3 figures, CoRL 2024 Workshop MAPoDeL
☆ k-NN as a Simple and Effective Estimator of Transferability
How well can one expect transfer learning to work in a new setting where the domain is shifted, the task is different, and the architecture changes? Many transfer learning metrics have been proposed to answer this question. But how accurate are their predictions in a realistic new setting? We conducted an extensive evaluation involving over 42,000 experiments comparing 23 transferability metrics across 16 different datasets to assess their ability to predict transfer performance. Our findings reveal that none of the existing metrics perform well across the board. However, we find that a simple k-nearest neighbor evaluation -- as is commonly used to evaluate feature quality for self-supervision -- not only surpasses existing metrics, but also offers better computational efficiency and ease of implementation.
☆ Global Convergence of Continual Learning on Non-IID Data
Continual learning, which aims to learn multiple tasks sequentially, has gained extensive attention. However, most existing work focuses on empirical studies, and the theoretical aspect remains under-explored. Recently, a few investigations have considered the theory of continual learning only for linear regressions, establishes the results based on the strict independent and identically distributed (i.i.d.) assumption and the persistent excitation on the feature data that may be difficult to verify or guarantee in practice. To overcome this fundamental limitation, in this paper, we provide a general and comprehensive theoretical analysis for continual learning of regression models. By utilizing the stochastic Lyapunov function and martingale estimation techniques, we establish the almost sure convergence results of continual learning under a general data condition for the first time. Additionally, without any excitation condition imposed on the data, the convergence rates for the forgetting and regret metrics are provided.
comment: We establish the almost sure convergence results of continual learning under a general data condition
☆ Neuro-symbolic Weak Supervision: Theory and Semantics
Weak supervision allows machine learning models to learn from limited or noisy labels, but it introduces challenges in interpretability and reliability - particularly in multi-instance partial label learning (MI-PLL), where models must resolve both ambiguous labels and uncertain instance-label mappings. We propose a semantics for neuro-symbolic framework that integrates Inductive Logic Programming (ILP) to improve MI-PLL by providing structured relational constraints that guide learning. Within our semantic characterization, ILP defines a logical hypothesis space for label transitions, clarifies classifier semantics, and establishes interpretable performance standards. This hybrid approach improves robustness, transparency, and accountability in weakly supervised settings, ensuring neural predictions align with domain knowledge. By embedding weak supervision into a logical framework, we enhance both interpretability and learning, making weak supervision more suitable for real-world, high-stakes applications.
☆ Deterministic Certification of Graph Neural Networks against Graph Poisoning Attacks with Arbitrary Perturbations CVPR 2025
Graph neural networks (GNNs) are becoming the de facto method to learn on the graph data and have achieved the state-of-the-art on node and graph classification tasks. However, recent works show GNNs are vulnerable to training-time poisoning attacks -- marginally perturbing edges, nodes, or/and node features of training graph(s) can largely degrade GNNs' testing performance. Most previous defenses against graph poisoning attacks are empirical and are soon broken by adaptive / stronger ones. A few provable defenses provide robustness guarantees, but have large gaps when applied in practice: 1) restrict the attacker on only one type of perturbation; 2) design for a particular GNN architecture or task; and 3) robustness guarantees are not 100\% accurate. In this work, we bridge all these gaps by developing PGNNCert, the first certified defense of GNNs against poisoning attacks under arbitrary (edge, node, and node feature) perturbations with deterministic robustness guarantees. Extensive evaluations on multiple node and graph classification datasets and GNNs demonstrate the effectiveness of PGNNCert to provably defend against arbitrary poisoning perturbations. PGNNCert is also shown to significantly outperform the state-of-the-art certified defenses against edge perturbation or node perturbation during GNN training.
comment: Accepted at CVPR 2025
☆ Learning a Class of Mixed Linear Regressions: Global Convergence under General Data Conditions
Mixed linear regression (MLR) has attracted increasing attention because of its great theoretical and practical importance in capturing nonlinear relationships by utilizing a mixture of linear regression sub-models. Although considerable efforts have been devoted to the learning problem of such systems, i.e., estimating data labels and identifying model parameters, most existing investigations employ the offline algorithm, impose the strict independent and identically distributed (i.i.d.) or persistent excitation (PE) conditions on the regressor data, and provide local convergence results only. In this paper, we investigate the recursive estimation and data clustering problems for a class of stochastic MLRs with two components. To address this inherently nonconvex optimization problem, we propose a novel two-step recursive identification algorithm to estimate the true parameters, where the direction vector and the scaling coefficient of the unknown parameters are estimated by the least squares and the expectation-maximization (EM) principles, respectively. Under a general data condition, which is much weaker than the traditional i.i.d. and PE conditions, we establish the global convergence and the convergence rate of the proposed identification algorithm for the first time. Furthermore, we prove that, without any excitation condition on the regressor data, the data clustering performance including the cumulative mis-classification error and the within-cluster error can be optimal asymptotically. Finally, we provide a numerical example to illustrate the performance of the proposed learning algorithm.
☆ Statistically Testing Training Data for Unwanted Error Patterns using Rule-Oriented Regression
Artificial intelligence models trained from data can only be as good as the underlying data is. Biases in training data propagating through to the output of a machine learning model are a well-documented and well-understood phenomenon, but the machinery to prevent these undesired effects is much less developed. Efforts to ensure data is clean during collection, such as using bias-aware sampling, are most effective when the entity controlling data collection also trains the AI. In cases where the data is already available, how do we find out if the data was already manipulated, i.e., ``poisoned'', so that an undesired behavior would be trained into a machine learning model? This is a challenge fundamentally different to (just) improving approximation accuracy or efficiency, and we provide a method to test training data for flaws, to establish a trustworthy ground-truth for a subsequent training of machine learning models (of any kind). Unlike the well-studied problem of approximating data using fuzzy rules that are generated from the data, our method hinges on a prior definition of rules to happen before seeing the data to be tested. Therefore, the proposed method can also discover hidden error patterns, which may also have substantial influence. Our approach extends the abilities of conventional statistical testing by letting the ``test-condition'' be any Boolean condition to describe a pattern in the data, whose presence we wish to determine. The method puts fuzzy inference into a regression model, to get the best of the two: explainability from fuzzy logic with statistical properties and diagnostics from the regression, and finally also being applicable to ``small data'', hence not requiring large datasets as deep learning methods do. We provide an open source implementation for demonstration and experiments.
☆ Verbal Process Supervision Elicits Better Coding Agents
The emergence of large language models and their applications as AI agents have significantly advanced state-of-the-art code generation benchmarks, transforming modern software engineering tasks. However, even with test-time computed reasoning models, these systems still struggle with complex software engineering challenges. This work introduces CURA, a code understanding and reasoning agent system enhanced with verbal process supervision (VPS), achieving a 3.65\% improvement over baseline models on challenging benchmarks like BigCodeBench. Furthermore, CURA, when paired with the o3-mini model and VPS techniques, attains state-of-the-art performance. This work represents a step forward in integrating reasoning-driven architectures with LLM-based code generation, enabling agentic reasoning for language models to solve complex software engineering tasks.
☆ Whispering in Amharic: Fine-tuning Whisper for Low-resource Language
This work explores fine-tuning OpenAI's Whisper automatic speech recognition (ASR) model for Amharic, a low-resource language, to improve transcription accuracy. While the foundational Whisper model struggles with Amharic due to limited representation in its training data, we fine-tune it using datasets like Mozilla Common Voice, FLEURS, and the BDU-speech dataset. The best-performing model, Whispersmall-am, significantly improves when finetuned on a mix of existing FLEURS data and new, unseen Amharic datasets. Training solely on new data leads to poor performance, but combining it with FLEURS data reinforces the model, enabling better specialization in Amharic. We also demonstrate that normalizing Amharic homophones significantly enhances Word Error Rate (WER) and Bilingual Evaluation Understudy (BLEU) scores. This study underscores the importance of fine-tuning strategies and dataset composition for improving ASR in low-resource languages, providing insights for future Amharic speech recognition research.
☆ PALATE: Peculiar Application of the Law of Total Expectation to Enhance the Evaluation of Deep Generative Models
Deep generative models (DGMs) have caused a paradigm shift in the field of machine learning, yielding noteworthy advancements in domains such as image synthesis, natural language processing, and other related areas. However, a comprehensive evaluation of these models that accounts for the trichotomy between fidelity, diversity, and novelty in generated samples remains a formidable challenge. A recently introduced solution that has emerged as a promising approach in this regard is the Feature Likelihood Divergence (FLD), a method that offers a theoretically motivated practical tool, yet also exhibits some computational challenges. In this paper, we propose PALATE, a novel enhancement to the evaluation of DGMs that addresses limitations of existing metrics. Our approach is based on a peculiar application of the law of total expectation to random variables representing accessible real data. When combined with the MMD baseline metric and DINOv2 feature extractor, PALATE offers a holistic evaluation framework that matches or surpasses state-of-the-art solutions while providing superior computational efficiency and scalability to large-scale datasets. Through a series of experiments, we demonstrate the effectiveness of the PALATE enhancement, contributing a computationally efficient, holistic evaluation approach that advances the field of DGMs assessment, especially in detecting sample memorization and evaluating generalization capabilities.
☆ InPO: Inversion Preference Optimization with Reparametrized DDIM for Efficient Diffusion Model Alignment CVPR2025
Without using explicit reward, direct preference optimization (DPO) employs paired human preference data to fine-tune generative models, a method that has garnered considerable attention in large language models (LLMs). However, exploration of aligning text-to-image (T2I) diffusion models with human preferences remains limited. In comparison to supervised fine-tuning, existing methods that align diffusion model suffer from low training efficiency and subpar generation quality due to the long Markov chain process and the intractability of the reverse process. To address these limitations, we introduce DDIM-InPO, an efficient method for direct preference alignment of diffusion models. Our approach conceptualizes diffusion model as a single-step generative model, allowing us to fine-tune the outputs of specific latent variables selectively. In order to accomplish this objective, we first assign implicit rewards to any latent variable directly via a reparameterization technique. Then we construct an Inversion technique to estimate appropriate latent variables for preference optimization. This modification process enables the diffusion model to only fine-tune the outputs of latent variables that have a strong correlation with the preference dataset. Experimental results indicate that our DDIM-InPO achieves state-of-the-art performance with just 400 steps of fine-tuning, surpassing all preference aligning baselines for T2I diffusion models in human preference evaluation tasks.
comment: Accepted by CVPR2025
☆ Distributionally Robust Federated Learning: An ADMM Algorithm
Federated learning (FL) aims to train machine learning (ML) models collaboratively using decentralized data, bypassing the need for centralized data aggregation. Standard FL models often assume that all data come from the same unknown distribution. However, in practical situations, decentralized data frequently exhibit heterogeneity. We propose a novel FL model, Distributionally Robust Federated Learning (DRFL), that applies distributionally robust optimization to overcome the challenges posed by data heterogeneity and distributional ambiguity. We derive a tractable reformulation for DRFL and develop a novel solution method based on the alternating direction method of multipliers (ADMM) algorithm to solve this problem. Our experimental results demonstrate that DRFL outperforms standard FL models under data heterogeneity and ambiguity.
☆ Teaching LLMs for Step-Level Automatic Math Correction via Reinforcement Learning
Automatic math correction aims to check students' solutions to mathematical problems via artificial intelligence technologies. Most existing studies focus on judging the final answer at the problem level, while they ignore detailed feedback on each step in a math problem-solving process, which requires abilities of semantic understanding and reasoning. In this paper, we propose a reinforcement learning (RL)-based method to boost large language model (LLM) for step-level automatic math correction, named StepAMC. Particularly, we convert the step-level automatic math correction within the text classification task into an RL problem to enhance the reasoning capabilities of LLMs. Then, we design a space-constrained policy network to improve the stability of RL. Then, we introduce a fine-grained reward network to convert the binary human feedback into a continuous value. We conduct extensive experiments over two benchmark datasets and the results show that our model outperforms the eleven strong baselines.
☆ Offline Meteorology-Pollution Coupling Global Air Pollution Forecasting Model with Bilinear Pooling
Air pollution has become a major threat to human health, making accurate forecasting crucial for pollution control. Traditional physics-based models forecast global air pollution by coupling meteorology and pollution processes, using either online or offline methods depending on whether fully integrated with meteorological models and run simultaneously. However, the high computational demands of both methods severely limit real-time prediction efficiency. Existing deep learning (DL) solutions employ online coupling strategies for global air pollution forecasting, which finetune pollution forecasting based on pretrained atmospheric models, requiring substantial training resources. This study pioneers a DL-based offline coupling framework that utilizes bilinear pooling to achieve offline coupling between meteorological fields and pollutants. The proposed model requires only 13% of the parameters of DL-based online coupling models while achieving competitive performance. Compared with the state-of-the-art global air pollution forecasting model CAMS, our approach demonstrates superiority in 63% variables across all forecast time steps and 85% variables in predictions exceeding 48 hours. This work pioneers experimental validation of the effectiveness of meteorological fields in DL-based global air pollution forecasting, demonstrating that offline coupling meteorological fields with pollutants can achieve a 15% relative reduction in RMSE across all pollution variables. The research establishes a new paradigm for real-time global air pollution warning systems and delivers critical technical support for developing more efficient and comprehensive AI-powered global atmospheric forecasting frameworks.
☆ Knowledge Graph Enhanced Generative Multi-modal Models for Class-Incremental Learning
Continual learning in computer vision faces the critical challenge of catastrophic forgetting, where models struggle to retain prior knowledge while adapting to new tasks. Although recent studies have attempted to leverage the generalization capabilities of pre-trained models to mitigate overfitting on current tasks, models still tend to forget details of previously learned categories as tasks progress, leading to misclassification. To address these limitations, we introduce a novel Knowledge Graph Enhanced Generative Multi-modal model (KG-GMM) that builds an evolving knowledge graph throughout the learning process. Our approach utilizes relationships within the knowledge graph to augment the class labels and assigns different relations to similar categories to enhance model differentiation. During testing, we propose a Knowledge Graph Augmented Inference method that locates specific categories by analyzing relationships within the generated text, thereby reducing the loss of detailed information about old classes when learning new knowledge and alleviating forgetting. Experiments demonstrate that our method effectively leverages relational information to help the model correct mispredictions, achieving state-of-the-art results in both conventional CIL and few-shot CIL settings, confirming the efficacy of knowledge graphs at preserving knowledge in the continual learning scenarios.
☆ Solving Situation Puzzles with Large Language Model and External Reformulation
In recent years, large language models (LLMs) have shown an impressive ability to perform arithmetic and symbolic reasoning tasks. However, we found that LLMs (e.g., ChatGPT) cannot perform well on reasoning that requires multiple rounds of dialogue, especially when solving situation puzzles. Specifically, LLMs intend to ask very detailed questions focusing on a specific aspect or same/similar questions after several rounds of Q&As. To help LLMs get out of the above dilemma, we propose a novel external reformulation methodology, where the situation puzzle will be reformulated after several rounds of Q&A or when the LLMs raise an incorrect guess. Experiments show superior performance (e.g., win rate, number of question/guess attempts) of our method than directly using LLMs for solving situation puzzles, highlighting the potential of strategic problem reformulation to enhance the reasoning capabilities of LLMs in complex interactive scenarios.
☆ Finite-Time Bounds for Two-Time-Scale Stochastic Approximation with Arbitrary Norm Contractions and Markovian Noise
Two-time-scale Stochastic Approximation (SA) is an iterative algorithm with applications in reinforcement learning and optimization. Prior finite time analysis of such algorithms has focused on fixed point iterations with mappings contractive under Euclidean norm. Motivated by applications in reinforcement learning, we give the first mean square bound on non linear two-time-scale SA where the iterations have arbitrary norm contractive mappings and Markovian noise. We show that the mean square error decays at a rate of $O(1/n^{2/3})$ in the general case, and at a rate of $O(1/n)$ in a special case where the slower timescale is noiseless. Our analysis uses the generalized Moreau envelope to handle the arbitrary norm contractions and solutions of Poisson equation to deal with the Markovian noise. By analyzing the SSP Q-Learning algorithm, we give the first $O(1/n)$ bound for an algorithm for asynchronous control of MDPs under the average reward criterion. We also obtain a rate of $O(1/n)$ for Q-Learning with Polyak-averaging and provide an algorithm for learning Generalized Nash Equilibrium (GNE) for strongly monotone games which converges at a rate of $O(1/n^{2/3})$.
comment: Submitted to IEEE Conference on Decision and Control (CDC) 2025
☆ RoCA: Robust Contrastive One-class Time Series Anomaly Detection with Contaminated Data
The accumulation of time-series signals and the absence of labels make time-series Anomaly Detection (AD) a self-supervised task of deep learning. Methods based on normality assumptions face the following three limitations: (1) A single assumption could hardly characterize the whole normality or lead to some deviation. (2) Some assumptions may go against the principle of AD. (3) Their basic assumption is that the training data is uncontaminated (free of anomalies), which is unrealistic in practice, leading to a decline in robustness. This paper proposes a novel robust approach, RoCA, which is the first to address all of the above three challenges, as far as we are aware. It fuses the separated assumptions of one-class classification and contrastive learning in a single training process to characterize a more complete so-called normality. Additionally, it monitors the training data and computes a carefully designed anomaly score throughout the training process. This score helps identify latent anomalies, which are then used to define the classification boundary, inspired by the concept of outlier exposure. The performance on AIOps datasets improved by 6% compared to when contamination was not considered (COCA). On two large and high-dimensional multivariate datasets, the performance increased by 5% to 10%. RoCA achieves the highest average performance on both univariate and multivariate datasets. The source code is available at https://github.com/ruiking04/RoCA.
☆ Maximum Redundancy Pruning: A Principle-Driven Layerwise Sparsity Allocation for LLMs
Large language models (LLMs) have demonstrated impressive capabilities, but their enormous size poses significant challenges for deployment in real-world applications. To address this issue, researchers have sought to apply network pruning techniques to LLMs. A critical challenge in pruning is allocation the sparsity for each layer. Recent sparsity allocation methods is often based on heuristics or search that can easily lead to suboptimal performance. In this paper, we conducted an extensive investigation into various LLMs and revealed three significant discoveries: (1) the layerwise pruning sensitivity (LPS) of LLMs is highly non-uniform, (2) the choice of pruning metric affects LPS, and (3) the performance of a sparse model is related to the uniformity of its layerwise redundancy level. Based on these observations, we propose that the layerwise sparsity of LLMs should adhere to three principles: \emph{non-uniformity}, \emph{pruning metric dependency}, and \emph{uniform layerwise redundancy level} in the pruned model. To this end, we proposed Maximum Redundancy Pruning (MRP), an iterative pruning algorithm that prunes in the most redundant layers (\emph{i.e.}, those with the highest non-outlier ratio) at each iteration. The achieved layerwise sparsity aligns with the outlined principles. We conducted extensive experiments on publicly available LLMs, including the LLaMA2 and OPT, across various benchmarks. Experimental results validate the effectiveness of MRP, demonstrating its superiority over previous methods.
☆ ALWNN Empowered Automatic Modulation Classification: Conquering Complexity and Scarce Sample Conditions
In Automatic Modulation Classification (AMC), deep learning methods have shown remarkable performance, offering significant advantages over traditional approaches and demonstrating their vast potential. Nevertheless, notable drawbacks, particularly in their high demands for storage, computational resources, and large-scale labeled data, which limit their practical application in real-world scenarios. To tackle this issue, this paper innovatively proposes an automatic modulation classification model based on the Adaptive Lightweight Wavelet Neural Network (ALWNN) and the few-shot framework (MALWNN). The ALWNN model, by integrating the adaptive wavelet neural network and depth separable convolution, reduces the number of model parameters and computational complexity. The MALWNN framework, using ALWNN as an encoder and incorporating prototype network technology, decreases the model's dependence on the quantity of samples. Simulation results indicate that this model performs remarkably well on mainstream datasets. Moreover, in terms of Floating Point Operations Per Second (FLOPS) and Normalized Multiply - Accumulate Complexity (NMACC), ALWNN significantly reduces computational complexity compared to existing methods. This is further validated by real-world system tests on USRP and Raspberry Pi platforms. Experiments with MALWNN show its superior performance in few-shot learning scenarios compared to other algorithms.
☆ Do Your Best and Get Enough Rest for Continual Learning CVPR
According to the forgetting curve theory, we can enhance memory retention by learning extensive data and taking adequate rest. This means that in order to effectively retain new knowledge, it is essential to learn it thoroughly and ensure sufficient rest so that our brain can memorize without forgetting. The main takeaway from this theory is that learning extensive data at once necessitates sufficient rest before learning the same data again. This aspect of human long-term memory retention can be effectively utilized to address the continual learning of neural networks. Retaining new knowledge for a long period of time without catastrophic forgetting is the critical problem of continual learning. Therefore, based on Ebbinghaus' theory, we introduce the view-batch model that adjusts the learning schedules to optimize the recall interval between retraining the same samples. The proposed view-batch model allows the network to get enough rest to learn extensive knowledge from the same samples with a recall interval of sufficient length. To this end, we specifically present two approaches: 1) a replay method that guarantees the optimal recall interval, and 2) a self-supervised learning that acquires extensive knowledge from a single training sample at a time. We empirically show that these approaches of our method are aligned with the forgetting curve theory, which can enhance long-term memory. In our experiments, we also demonstrate that our method significantly improves many state-of-the-art continual learning methods in various protocols and scenarios. We open-source this project at https://github.com/hankyul2/ViewBatchModel.
comment: Accepted by IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) 2025
☆ MoST: Efficient Monarch Sparse Tuning for 3D Representation Learning
We introduce Monarch Sparse Tuning (MoST), the first reparameterization-based parameter-efficient fine-tuning (PEFT) method tailored for 3D representation learning. Unlike existing adapter-based and prompt-tuning 3D PEFT methods, MoST introduces no additional inference overhead and is compatible with many 3D representation learning backbones. At its core, we present a new family of structured matrices for 3D point clouds, Point Monarch, which can capture local geometric features of irregular points while offering high expressiveness. MoST reparameterizes the dense update weight matrices as our sparse Point Monarch matrices, significantly reducing parameters while retaining strong performance. Experiments on various backbones show that MoST is simple, effective, and highly generalizable. It captures local features in point clouds, achieving state-of-the-art results on multiple benchmarks, e.g., 97.5% acc. on ScanObjectNN (PB_50_RS) and 96.2% on ModelNet40 classification, while it can also combine with other matrix decompositions (e.g., Low-rank, Kronecker) to further reduce parameters.
comment: 11 pages, 6 figures, 6 tables. Code and weights are available at https://github.com/xhanxu/MoST
☆ Latent Embedding Adaptation for Human Preference Alignment in Diffusion Planners
This work addresses the challenge of personalizing trajectories generated in automated decision-making systems by introducing a resource-efficient approach that enables rapid adaptation to individual users' preferences. Our method leverages a pretrained conditional diffusion model with Preference Latent Embeddings (PLE), trained on a large, reward-free offline dataset. The PLE serves as a compact representation for capturing specific user preferences. By adapting the pretrained model using our proposed preference inversion method, which directly optimizes the learnable PLE, we achieve superior alignment with human preferences compared to existing solutions like Reinforcement Learning from Human Feedback (RLHF) and Low-Rank Adaptation (LoRA). To better reflect practical applications, we create a benchmark experiment using real human preferences on diverse, high-reward trajectories.
comment: 8 pages
☆ A New Stochastic Approximation Method for Gradient-based Simulated Parameter Estimation
This paper tackles the challenge of parameter calibration in stochastic models, particularly in scenarios where the likelihood function is unavailable in an analytical form. We introduce a gradient-based simulated parameter estimation framework, which employs a multi-time scale stochastic approximation algorithm. This approach effectively addresses the ratio bias that arises in both maximum likelihood estimation and posterior density estimation problems. The proposed algorithm enhances estimation accuracy and significantly reduces computational costs, as demonstrated through extensive numerical experiments. Our work extends the GSPE framework to handle complex models such as hidden Markov models and variational inference-based problems, offering a robust solution for parameter estimation in challenging stochastic environments.
☆ Improved Rates of Differentially Private Nonconvex-Strongly-Concave Minimax Optimization AAAI 2025
In this paper, we study the problem of (finite sum) minimax optimization in the Differential Privacy (DP) model. Unlike most of the previous studies on the (strongly) convex-concave settings or loss functions satisfying the Polyak-Lojasiewicz condition, here we mainly focus on the nonconvex-strongly-concave one, which encapsulates many models in deep learning such as deep AUC maximization. Specifically, we first analyze a DP version of Stochastic Gradient Descent Ascent (SGDA) and show that it is possible to get a DP estimator whose $l_2$-norm of the gradient for the empirical risk function is upper bounded by $\tilde{O}(\frac{d^{1/4}}{({n\epsilon})^{1/2}})$, where $d$ is the model dimension and $n$ is the sample size. We then propose a new method with less gradient noise variance and improve the upper bound to $\tilde{O}(\frac{d^{1/3}}{(n\epsilon)^{2/3}})$, which matches the best-known result for DP Empirical Risk Minimization with non-convex loss. We also discussed several lower bounds of private minimax optimization. Finally, experiments on AUC maximization, generative adversarial networks, and temporal difference learning with real-world data support our theoretical analysis.
comment: Published in AAAI 2025
☆ LoTUS: Large-Scale Machine Unlearning with a Taste of Uncertainty CVPR 2025
We present LoTUS, a novel Machine Unlearning (MU) method that eliminates the influence of training samples from pre-trained models, avoiding retraining from scratch. LoTUS smooths the prediction probabilities of the model -- up to an information theoretic bound -- mitigating its over-confidence that stems from data memorization. We evaluate LoTUS on the Transformer and ResNet18 models, against eight baseline methods, on five public datasets. Beyond established MU benchmarks, we evaluate unlearning on a large-scale dataset (ImageNet1k) which deters retraining, simulating real-world conditions. Moreover, we introduce the novel Retrain-Free Jensen-Shannon Divergence (RF-JSD) metric to enable evaluation under real-world conditions. Experimental results show that LoTUS outperforms state-of-the-art methods in terms of both efficiency and effectiveness. Code: https://github.com/cspartalis/LoTUS.
comment: Accepted as a main conference paper at CVPR 2025
☆ Efficient Transformed Gaussian Process State-Space Models for Non-Stationary High-Dimensional Dynamical Systems
Gaussian process state-space models (GPSSMs) have emerged as a powerful framework for modeling dynamical systems, offering interpretable uncertainty quantification and inherent regularization. However, existing GPSSMs face significant challenges in handling high-dimensional, non-stationary systems due to computational inefficiencies, limited scalability, and restrictive stationarity assumptions. In this paper, we propose an efficient transformed Gaussian process state-space model (ETGPSSM) to address these limitations. Our approach leverages a single shared Gaussian process (GP) combined with normalizing flows and Bayesian neural networks, enabling efficient modeling of complex, high-dimensional state transitions while preserving scalability. To address the lack of closed-form expressions for the implicit process in the transformed GP, we follow its generative process and introduce an efficient variational inference algorithm, aided by the ensemble Kalman filter (EnKF), to enable computationally tractable learning and inference. Extensive empirical evaluations on synthetic and real-world datasets demonstrate the superior performance of our ETGPSSM in system dynamics learning, high-dimensional state estimation, and time-series forecasting, outperforming existing GPSSMs and neural network-based methods in both accuracy and computational efficiency.
comment: 13 pages, 6 figures
☆ DiffMove: Group Mobility Tendency Enhanced Trajectory Recovery via Diffusion Model
In the real world, trajectory data is often sparse and incomplete due to low collection frequencies or limited device coverage. Trajectory recovery aims to recover these missing trajectory points, making the trajectories denser and more complete. However, this task faces two key challenges: 1) The excessive sparsity of individual trajectories makes it difficult to effectively leverage historical information for recovery; 2) Sparse trajectories make it harder to capture complex individual mobility preferences. To address these challenges, we propose a novel method called DiffMove. Firstly, we harness crowd wisdom for trajectory recovery. Specifically, we construct a group tendency graph using the collective trajectories of all users and then integrate the group mobility trends into the location representations via graph embedding. This solves the challenge of sparse trajectories being unable to rely on individual historical trajectories for recovery. Secondly, we capture individual mobility preferences from both historical and current perspectives. Finally, we integrate group mobility tendencies and individual preferences into the spatiotemporal distribution of the trajectory to recover high-quality trajectories. Extensive experiments on two real-world datasets demonstrate that DiffMove outperforms existing state-of-the-art methods. Further analysis validates the robustness of our method.
☆ Byzantine-Resilient Over-the-Air Federated Learning under Zero-Trust Architecture
Over-the-air computation (AirComp) has emerged as an essential approach for enabling communication-efficient federated learning (FL) over wireless networks. Nonetheless, the inherent analog transmission mechanism in AirComp-based FL (AirFL) intensifies challenges posed by potential Byzantine attacks. In this paper, we propose a novel Byzantine-robust FL paradigm for over-the-air transmissions, referred to as federated learning with secure adaptive clustering (FedSAC). FedSAC aims to protect a portion of the devices from attacks through zero trust architecture (ZTA) based Byzantine identification and adaptive device clustering. By conducting a one-step convergence analysis, we theoretically characterize the convergence behavior with different device clustering mechanisms and uneven aggregation weighting factors for each device. Building upon our analytical results, we formulate a joint optimization problem for the clustering and weighting factors in each communication round. To facilitate the targeted optimization, we propose a dynamic Byzantine identification method using historical reputation based on ZTA. Furthermore, we introduce a sequential clustering method, transforming the joint optimization into a weighting optimization problem without sacrificing the optimality. To optimize the weighting, we capitalize on the penalty convex-concave procedure (P-CCP) to obtain a stationary solution. Numerical results substantiate the superiority of the proposed FedSAC over existing methods in terms of both test accuracy and convergence rate.
comment: Accepted by IEEE JSAC
☆ Analyzing Islamophobic Discourse Using Semi-Coded Terms and LLMs
Islamophobia started evolving into a global phenomenon by attracting followers across the globe, particularly in Western societies. Thus, understanding Islamophobia's global spread and online dissemination is crucial. This paper performs a large-scale analysis of specialized, semi-coded Islamophobic terms such as (muzrat, pislam, mudslime, mohammedan, muzzies) floated on extremist social platforms, i.e., 4Chan, Gab, Telegram, etc. First, we use large language models (LLMs) to show their ability to understand these terms. Second, using Google Perspective API, we also find that Islamophobic text is more toxic compared to other kinds of hate speech. Finally, we use BERT topic modeling approach to extract different topics and Islamophobic discourse on these social platforms. Our findings indicate that LLMs understand these Out-Of-Vocabulary (OOV) slurs; however, measures are still required to control such discourse. Our topic modeling also indicates that Islamophobic text is found across various political, conspiratorial, and far-right movements and is particularly directed against Muslim immigrants. Taken altogether, we performed the first study on Islamophobic semi-coded terms and shed a global light on Islamophobia.
☆ Risk Management for Distributed Arbitrage Systems: Integrating Artificial Intelligence
Effective risk management solutions become absolutely crucial when financial markets embrace distributed technology and decentralized financing (DeFi). This study offers a thorough survey and comparative analysis of the integration of artificial intelligence (AI) in risk management for distributed arbitrage systems. We examine several modern caching techniques namely in memory caching, distributed caching, and proxy caching and their functions in enhancing performance in decentralized settings. Through literature review we examine the utilization of AI techniques for alleviating risks related to market volatility, liquidity challenges, operational failures, regulatory compliance, and security threats. This comparison research evaluates various case studies from prominent DeFi technologies, emphasizing critical performance metrics like latency reduction, load balancing, and system resilience. Additionally, we examine the problems and trade offs associated with these technologies, emphasizing their effects on consistency, scalability, and fault tolerance. By meticulously analyzing real world applications, specifically centering on the Aave platform as our principal case study, we illustrate how the purposeful amalgamation of AI with contemporary caching methodologies has revolutionized risk management in distributed arbitrage systems.
comment: International Conference on AI and Financial Innovation AIFI-2025
☆ PNN: A Novel Progressive Neural Network for Fault Classification in Rotating Machinery under Small Dataset Constraint
Fault detection in rotating machinery is a complex task, particularly in small and heterogeneous dataset scenarios. Variability in sensor placement, machinery configurations, and structural differences further increase the complexity of the problem. Conventional deep learning approaches often demand large, homogeneous datasets, limiting their applicability in data-scarce industrial environments. While transfer learning and few-shot learning have shown potential, however, they are often constrained by the need for extensive fault datasets. This research introduces a unified framework leveraging a novel progressive neural network (PNN) architecture designed to address these challenges. The PNN sequentially estimates the fixed-size refined features of the higher order with the help of all previously estimated features and appends them to the feature set. This fixed-size feature output at each layer controls the complexity of the PNN and makes it suitable for effective learning from small datasets. The framework's effectiveness is validated on eight datasets, including six open-source datasets, one in-house fault simulator, and one real-world industrial dataset. The PNN achieves state-of-the-art performance in fault detection across varying dataset sizes and machinery types, highlighting superior generalization and classification capabilities.
☆ Bridging Emotions and Architecture: Sentiment Analysis in Modern Distributed Systems
Sentiment analysis is a field within NLP that has gained importance because it is applied in various areas such as; social media surveillance, customer feedback evaluation and market research. At the same time, distributed systems allow for effective processing of large amounts of data. Therefore, this paper examines how sentiment analysis converges with distributed systems by concentrating on different approaches, challenges and future investigations. Furthermore, we do an extensive experiment where we train sentiment analysis models using both single node configuration and distributed architecture to bring out the benefits and shortcomings of each method in terms of performance and accuracy.
comment: IEEE 3rd International Conference on Advancements in Smart, Secure and Intelligent Computing (ASSIC)
☆ Severing Spurious Correlations with Data Pruning ICLR 2025
Deep neural networks have been shown to learn and rely on spurious correlations present in the data that they are trained on. Reliance on such correlations can cause these networks to malfunction when deployed in the real world, where these correlations may no longer hold. To overcome the learning of and reliance on such correlations, recent studies propose approaches that yield promising results. These works, however, study settings where the strength of the spurious signal is significantly greater than that of the core, invariant signal, making it easier to detect the presence of spurious features in individual training samples and allow for further processing. In this paper, we identify new settings where the strength of the spurious signal is relatively weaker, making it difficult to detect any spurious information while continuing to have catastrophic consequences. We also discover that spurious correlations are learned primarily due to only a handful of all the samples containing the spurious feature and develop a novel data pruning technique that identifies and prunes small subsets of the training data that contain these samples. Our proposed technique does not require inferred domain knowledge, information regarding the sample-wise presence or nature of spurious information, or human intervention. Finally, we show that such data pruning attains state-of-the-art performance on previously studied settings where spurious information is identifiable.
comment: ICLR 2025, Spotlight
☆ DiffGED: Computing Graph Edit Distance via Diffusion-based Graph Matching
The Graph Edit Distance (GED) problem, which aims to compute the minimum number of edit operations required to transform one graph into another, is a fundamental challenge in graph analysis with wide-ranging applications. However, due to its NP-hard nature, traditional A* approaches often suffer from scalability issue, making them computationally intractable for large graphs. Many recent deep learning frameworks address GED by formulating it as a regression task, which, while efficient, fails to recover the edit path -- a central interest in GED. Furthermore, recent hybrid approaches that combine deep learning with traditional methods to recover the edit path often yield poor solution quality. These methods also struggle to generate candidate solutions in parallel, resulting in increased running times.In this paper, we present a novel approach, DiffGED, that leverages generative diffusion model to solve GED and recover the corresponding edit path. Specifically, we first generate multiple diverse node matching matrices in parallel through a diffusion-based graph matching model. Next, node mappings are extracted from each generated matching matrices in parallel, and each extracted node mapping can be simply transformed into an edit path. Benefiting from the generative diversity provided by the diffusion model, DiffGED is less likely to fall into local sub-optimal solutions, thereby achieving superior overall solution quality close to the exact solution. Experimental results on real-world datasets demonstrate that DiffGED can generate multiple diverse edit paths with exceptionally high accuracy comparable to exact solutions while maintaining a running time shorter than most of hybrid approaches.
♻ ☆ Modern Hopfield Networks with Continuous-Time Memories
Recent research has established a connection between modern Hopfield networks (HNs) and transformer attention heads, with guarantees of exponential storage capacity. However, these models still face challenges scaling storage efficiently. Inspired by psychological theories of continuous neural resource allocation in working memory, we propose an approach that compresses large discrete Hopfield memories into smaller, continuous-time memories. Leveraging continuous attention, our new energy function modifies the update rule of HNs, replacing the traditional softmax-based probability mass function with a probability density, over the continuous memory. This formulation aligns with modern perspectives on human executive function, offering a principled link between attractor dynamics in working memory and resource-efficient memory allocation. Our framework maintains competitive performance with HNs while leveraging a compressed memory, reducing computational costs across synthetic and video datasets.
♻ ☆ Gradient-Guided Annealing for Domain Generalization CVPR2025
Domain Generalization (DG) research has gained considerable traction as of late, since the ability to generalize to unseen data distributions is a requirement that eludes even state-of-the-art training algorithms. In this paper we observe that the initial iterations of model training play a key role in domain generalization effectiveness, since the loss landscape may be significantly different across the training and test distributions, contrary to the case of i.i.d. data. Conflicts between gradients of the loss components of each domain lead the optimization procedure to undesirable local minima that do not capture the domain-invariant features of the target classes. We propose alleviating domain conflicts in model optimization, by iteratively annealing the parameters of a model in the early stages of training and searching for points where gradients align between domains. By discovering a set of parameter values where gradients are updated towards the same direction for each data distribution present in the training set, the proposed Gradient-Guided Annealing (GGA) algorithm encourages models to seek out minima that exhibit improved robustness against domain shifts. The efficacy of GGA is evaluated on five widely accepted and challenging image classification domain generalization benchmarks, where its use alone is able to establish highly competitive or even state-of-the-art performance. Moreover, when combined with previously proposed domain-generalization algorithms it is able to consistently improve their effectiveness by significant margins.
comment: Paper accepted in CVPR2025
♻ ☆ Noisy Low Rank Column-wise Sensing
This letter studies the AltGDmin algorithm for solving the noisy low rank column-wise sensing (LRCS) problem. Our sample complexity guarantee improves upon the best existing one by a factor $\max(r, \log(1/\epsilon))/r$ where $r$ is the rank of the unknown matrix and $\epsilon$ is the final desired accuracy. A second contribution of this work is a detailed comparison of guarantees from all work that studies the exact same mathematical problem as LRCS, but refers to it by different names.
comment: 9 pages
♻ ☆ Fine-tuning machine-learned particle-flow reconstruction for new detector geometries in future colliders
We demonstrate transfer learning capabilities in a machine-learned algorithm trained for particle-flow reconstruction in high energy particle colliders. This paper presents a cross-detector fine-tuning study, where we initially pre-train the model on a large full simulation dataset from one detector design, and subsequently fine-tune the model on a sample with a different collider and detector design. Specifically, we use the Compact Linear Collider detector (CLICdet) model for the initial training set, and demonstrate successful knowledge transfer to the CLIC-like detector (CLD) proposed for the Future Circular Collider in electron-positron mode (FCC-ee). We show that with an order of magnitude less samples from the second dataset, we can achieve the same performance as a costly training from scratch, across particle-level and event-level performance metrics, including jet and missing transverse momentum resolution. Furthermore, we find that the fine-tuned model achieves comparable performance to the traditional rule-based particle-flow approach on event-level metrics after training on 100,000 CLD events, whereas a model trained from scratch requires at least 1 million CLD events to achieve similar reconstruction performance. To our knowledge, this represents the first full-simulation cross-detector transfer learning study for particle-flow reconstruction. These findings offer valuable insights towards building large foundation models that can be fine-tuned across different detector designs and geometries, helping to accelerate the development cycle for new detectors and opening the door to rapid detector design and optimization using machine learning.
comment: 20 pages, 13 figures
♻ ☆ Subversion Strategy Eval: Can language models statelessly strategize to subvert control protocols?
An AI control protocol is a plan for usefully deploying AI systems that aims to prevent an AI from intentionally causing some unacceptable outcome. This paper investigates how well AI systems can generate and act on their own strategies for subverting control protocols whilst operating statelessly (without shared memory between contexts). To do this, an AI system may need to reliably generate optimal plans in each context, take actions with well-calibrated probabilities, and coordinate plans with other instances of itself without communicating. We develop Subversion Strategy Eval, a suite of eight environments, covering a range of protocols and strategic capabilities, and six sets of affordances that help isolate individual capabilities. We implement the evaluation in Inspect-AI and release it open-source. We evaluate Claude 3.5 models, including helpful-only versions, as well as OpenAI reasoning models. None of the models demonstrate substantial capability in strategizing to subvert control protocols statelessly. However, providing models with additional affordances, such as the ability to share a plan between contexts, can substantially improve performance. We hope our evaluations can act as a leading indicator for when models are capable of subverting control protocols and also relax the worst-case assumption of perfect strategic ability in AI control evaluations.
♻ ☆ FlowDock: Geometric Flow Matching for Generative Protein-Ligand Docking and Affinity Prediction
Powerful generative AI models of protein-ligand structure have recently been proposed, but few of these methods support both flexible protein-ligand docking and affinity estimation. Of those that do, none can directly model multiple binding ligands concurrently or have been rigorously benchmarked on pharmacologically relevant drug targets, hindering their widespread adoption in drug discovery efforts. In this work, we propose FlowDock, the first deep geometric generative model based on conditional flow matching that learns to directly map unbound (apo) structures to their bound (holo) counterparts for an arbitrary number of binding ligands. Furthermore, FlowDock provides predicted structural confidence scores and binding affinity values with each of its generated protein-ligand complex structures, enabling fast virtual screening of new (multi-ligand) drug targets. For the well-known PoseBusters Benchmark dataset, FlowDock outperforms single-sequence AlphaFold 3 with a 51% blind docking success rate using unbound (apo) protein input structures and without any information derived from multiple sequence alignments, and for the challenging new DockGen-E dataset, FlowDock outperforms single-sequence AlphaFold 3 and matches single-sequence Chai-1 for binding pocket generalization. Additionally, in the ligand category of the 16th community-wide Critical Assessment of Techniques for Structure Prediction (CASP16), FlowDock ranked among the top-5 methods for pharmacological binding affinity estimation across 140 protein-ligand complexes, demonstrating the efficacy of its learned representations in virtual screening. Source code, data, and pre-trained models are available at https://github.com/BioinfoMachineLearning/FlowDock.
comment: 15 pages, 2 tables, 2 algorithms, 11 figures. Code, data, pre-trained models, and baseline method predictions are available at https://github.com/BioinfoMachineLearning/FlowDock
♻ ☆ Optimistic Q-learning for average reward and episodic reinforcement learning
We present an optimistic Q-learning algorithm for regret minimization in average reward reinforcement learning under an additional assumption on the underlying MDP that for all policies, the time to visit some frequent state $s_0$ is finite and upper bounded by $H$, either in expectation or with constant probability. Our setting strictly generalizes the episodic setting and is significantly less restrictive than the assumption of bounded hitting time \textit{for all states} made by most previous literature on model-free algorithms in average reward settings. We demonstrate a regret bound of $\tilde{O}(H^5 S\sqrt{AT})$, where $S$ and $A$ are the numbers of states and actions, and $T$ is the horizon. A key technical novelty of our work is the introduction of an $\overline{L}$ operator defined as $\overline{L} v = \frac{1}{H} \sum_{h=1}^H L^h v$ where $L$ denotes the Bellman operator. Under the given assumption, we show that the $\overline{L}$ operator has a strict contraction (in span) even in the average-reward setting where the discount factor is $1$. Our algorithm design uses ideas from episodic Q-learning to estimate and apply this operator iteratively. Thus, we provide a unified view of regret minimization in episodic and non-episodic settings, which may be of independent interest.
comment: 37 pages, simplified proofs
♻ ☆ PRISM: Privacy-Preserving Improved Stochastic Masking for Federated Generative Models
Despite recent advancements in federated learning (FL), the integration of generative models into FL has been limited due to challenges such as high communication costs and unstable training in heterogeneous data environments. To address these issues, we propose PRISM, a FL framework tailored for generative models that ensures (i) stable performance in heterogeneous data distributions and (ii) resource efficiency in terms of communication cost and final model size. The key of our method is to search for an optimal stochastic binary mask for a random network rather than updating the model weights, identifying a sparse subnetwork with high generative performance; i.e., a ``strong lottery ticket''. By communicating binary masks in a stochastic manner, PRISM minimizes communication overhead. This approach, combined with the utilization of maximum mean discrepancy (MMD) loss and a mask-aware dynamic moving average aggregation method (MADA) on the server side, facilitates stable and strong generative capabilities by mitigating local divergence in FL scenarios. Moreover, thanks to its sparsifying characteristic, PRISM yields a lightweight model without extra pruning or quantization, making it ideal for environments such as edge devices. Experiments on MNIST, FMNIST, CelebA, and CIFAR10 demonstrate that PRISM outperforms existing methods, while maintaining privacy with minimal communication costs. PRISM is the first to successfully generate images under challenging non-IID and privacy-preserving FL environments on complex datasets, where previous methods have struggled.
♻ ☆ What Constitutes a Less Discriminatory Algorithm? NeurIPS
Disparate impact doctrine offers an important legal apparatus for targeting discriminatory data-driven algorithmic decisions. A recent body of work has focused on conceptualizing one particular construct from this doctrine: the less discriminatory alternative, an alternative policy that reduces disparities while meeting the same business needs of a status quo or baseline policy. However, attempts to operationalize this construct in the algorithmic setting must grapple with some thorny challenges and ambiguities. In this paper, we attempt to raise and resolve important questions about less discriminatory algorithms (LDAs). How should we formally define LDAs, and how does this interact with different societal goals they might serve? And how feasible is it for firms or plaintiffs to computationally search for candidate LDAs? We find that formal LDA definitions face fundamental challenges when they attempt to evaluate and compare predictive models in the absence of held-out data. As a result, we argue that LDA definitions cannot be purely quantitative, and must rely on standards of "reasonableness." We then identify both mathematical and computational constraints on firms' ability to efficiently conduct a proactive search for LDAs, but we provide evidence that these limits are "weak" in a formal sense. By defining LDAs formally, we put forward a framework in which both firms and plaintiffs can search for alternative models that comport with societal goals.
comment: 25 pages, 4 figures, 3 tables. Published at ACM CS&Law '25. Prior versions appeared at NeurIPS Algorithmic Fairness Through the Lens of Metrics and Evaluation Workshop (AFME '24) and Regulatable ML Workshop (RegML '24)
♻ ☆ Adaptive Collaborative Correlation Learning-based Semi-Supervised Multi-Label Feature Selection
Semi-supervised multi-label feature selection has recently been developed to solve the curse of dimensionality problem in high-dimensional multi-label data with certain samples missing labels. Although many efforts have been made, most existing methods use a predefined graph approach to capture the sample similarity or the label correlation. In this manner, the presence of noise and outliers within the original feature space can undermine the reliability of the resulting sample similarity graph. It also fails to precisely depict the label correlation due to the existence of unknown labels. Besides, these methods only consider the discriminative power of selected features, while neglecting their redundancy. In this paper, we propose an Adaptive Collaborative Correlation lEarning-based Semi-Supervised Multi-label Feature Selection (Access-MFS) method to address these issues. Specifically, a generalized regression model equipped with an extended uncorrelated constraint is introduced to select discriminative yet irrelevant features and maintain consistency between predicted and ground-truth labels in labeled data, simultaneously. Then, the instance correlation and label correlation are integrated into the proposed regression model to adaptively learn both the sample similarity graph and the label similarity graph, which mutually enhance feature selection performance. Extensive experimental results demonstrate the superiority of the proposed Access-MFS over other state-of-the-art methods.
♻ ☆ CarbonSense: A Multimodal Dataset and Baseline for Carbon Flux Modelling
Terrestrial carbon fluxes provide vital information about our biosphere's health and its capacity to absorb anthropogenic CO$_2$ emissions. The importance of predicting carbon fluxes has led to the emerging field of data-driven carbon flux modelling (DDCFM), which uses statistical techniques to predict carbon fluxes from biophysical data. However, the field lacks a standardized dataset to promote comparisons between models. To address this gap, we present CarbonSense, the first machine learning-ready dataset for DDCFM. CarbonSense integrates measured carbon fluxes, meteorological predictors, and satellite imagery from 385 locations across the globe, offering comprehensive coverage and facilitating robust model training. Additionally, we provide a baseline model using a current state-of-the-art DDCFM approach and a novel transformer based model. Our experiments illustrate the potential gains that multimodal deep learning techniques can bring to this domain. By providing these resources, we aim to lower the barrier to entry for other deep learning researchers to develop new models and drive new advances in carbon flux modelling.
comment: 9 content pages, 11 reference pages, 9 appendix pages
♻ ☆ HybridNorm: Towards Stable and Efficient Transformer Training via Hybrid Normalization
Transformers have become the de facto architecture for a wide range of machine learning tasks, particularly in large language models (LLMs). Despite their remarkable performance, challenges remain in training deep transformer networks, especially regarding the location of layer normalization. While Pre-Norm structures facilitate easier training due to their more prominent identity path, they often yield suboptimal performance compared to Post-Norm. In this paper, we propose $\textbf{HybridNorm}$, a straightforward yet effective hybrid normalization strategy that integrates the advantages of both Pre-Norm and Post-Norm approaches. Specifically, HybridNorm employs QKV normalization within the attention mechanism and Post-Norm in the feed-forward network (FFN) of each transformer block. This design not only stabilizes training but also enhances performance, particularly in the context of LLMs. Comprehensive experiments in both dense and sparse architectures show that HybridNorm consistently outperforms both Pre-Norm and Post-Norm approaches, achieving state-of-the-art results across various benchmarks. These findings highlight the potential of HybridNorm as a more stable and effective technique for improving the training and performance of deep transformer models. Code is available at https://github.com/BryceZhuo/HybridNorm.
♻ ☆ MF-CLIP: Leveraging CLIP as Surrogate Models for No-box Adversarial Attacks
The vulnerability of Deep Neural Networks (DNNs) to adversarial attacks poses a significant challenge to their deployment in safety-critical applications. While extensive research has addressed various attack scenarios, the no-box attack setting where adversaries have no prior knowledge, including access to training data of the target model, remains relatively underexplored despite its practical relevance. This work presents a systematic investigation into leveraging large-scale Vision-Language Models (VLMs), particularly CLIP, as surrogate models for executing no-box attacks. Our theoretical and empirical analyses reveal a key limitation in the execution of no-box attacks stemming from insufficient discriminative capabilities for direct application of vanilla CLIP as a surrogate model. To address this limitation, we propose MF-CLIP: a novel framework that enhances CLIP's effectiveness as a surrogate model through margin-aware feature space optimization. Comprehensive evaluations across diverse architectures and datasets demonstrate that MF-CLIP substantially advances the state-of-the-art in no-box attacks, surpassing existing baselines by 15.23% on standard models and achieving a 9.52% improvement on adversarially trained models. Our code will be made publicly available to facilitate reproducibility and future research in this direction.
♻ ☆ Quantifying Climate Change Impacts on Renewable Energy Generation: A Super-Resolution Recurrent Diffusion Model
Driven by global climate change and the ongoing energy transition, the coupling between power supply capabilities and meteorological factors has become increasingly significant. Over the long term, accurately quantifying the power generation of renewable energy under the influence of climate change is essential for the development of sustainable power systems. However, due to interdisciplinary differences in data requirements, climate data often lacks the necessary hourly resolution to capture the short-term variability and uncertainties of renewable energy resources. To address this limitation, a super-resolution recurrent diffusion model (SRDM) has been developed to enhance the temporal resolution of climate data and model the short-term uncertainty. The SRDM incorporates a pre-trained decoder and a denoising network, that generates long-term, high-resolution climate data through a recurrent coupling mechanism. The high-resolution climate data is then converted into power value using the mechanism model, enabling the simulation of wind and photovoltaic (PV) power generation on future long-term scales. Case studies were conducted in the Ejina region of Inner Mongolia, China, using fifth-generation reanalysis (ERA5) and coupled model intercomparison project (CMIP6) data under two climate pathways: SSP126 and SSP585. The results demonstrate that the SRDM outperforms existing generative models in generating super-resolution climate data. Furthermore, the research highlights the estimation biases introduced when low-resolution climate data is used for power conversion.
♻ ☆ MaRS: A Fast Sampler for Mean Reverting Diffusion based on ODE and SDE Solvers ICLR 2025
In applications of diffusion models, controllable generation is of practical significance, but is also challenging. Current methods for controllable generation primarily focus on modifying the score function of diffusion models, while Mean Reverting (MR) Diffusion directly modifies the structure of the stochastic differential equation (SDE), making the incorporation of image conditions simpler and more natural. However, current training-free fast samplers are not directly applicable to MR Diffusion. And thus MR Diffusion requires hundreds of NFEs (number of function evaluations) to obtain high-quality samples. In this paper, we propose a new algorithm named MaRS (MR Sampler) to reduce the sampling NFEs of MR Diffusion. We solve the reverse-time SDE and the probability flow ordinary differential equation (PF-ODE) associated with MR Diffusion, and derive semi-analytical solutions. The solutions consist of an analytical function and an integral parameterized by a neural network. Based on this solution, we can generate high-quality samples in fewer steps. Our approach does not require training and supports all mainstream parameterizations, including noise prediction, data prediction and velocity prediction. Extensive experiments demonstrate that MR Sampler maintains high sampling quality with a speedup of 10 to 20 times across ten different image restoration tasks. Our algorithm accelerates the sampling procedure of MR Diffusion, making it more practical in controllable generation.
comment: Accepted by ICLR 2025
♻ ☆ VeriMind: Agentic LLM for Automated Verilog Generation with a Novel Evaluation Metric
Designing Verilog modules requires meticulous attention to correctness, efficiency, and adherence to design specifications. However, manually writing Verilog code remains a complex and time-consuming task that demands both expert knowledge and iterative refinement. Leveraging recent advancements in large language models (LLMs) and their structured text generation capabilities, we propose VeriMind, an agentic LLM framework for Verilog code generation that significantly automates and optimizes the synthesis process. Unlike traditional LLM-based code generators, VeriMind employs a structured reasoning approach: given a user-provided prompt describing design requirements, the system first formulates a detailed train of thought before the final Verilog code is generated. This multi-step methodology enhances interpretability, accuracy, and adaptability in hardware design. In addition, we introduce a novel evaluation metric-pass@ARC-which combines the conventional pass@k measure with Average Refinement Cycles (ARC) to capture both success rate and the efficiency of iterative refinement. Experimental results on diverse hardware design tasks demonstrated that our approach achieved up to $8.3\%$ improvement on pass@k metric and $8.1\%$ on pass@ARC metric. These findings underscore the transformative potential of agentic LLMs in automated hardware design, RTL development, and digital system synthesis.
♻ ☆ On Using Certified Training towards Empirical Robustness
Adversarial training is arguably the most popular way to provide empirical robustness against specific adversarial examples. While variants based on multi-step attacks incur significant computational overhead, single-step variants are vulnerable to a failure mode known as catastrophic overfitting, which hinders their practical utility for large perturbations. A parallel line of work, certified training, has focused on producing networks amenable to formal guarantees of robustness against any possible attack. However, the wide gap between the best-performing empirical and certified defenses has severely limited the applicability of the latter. Inspired by recent developments in certified training, which rely on a combination of adversarial attacks with network over-approximations, and by the connections between local linearity and catastrophic overfitting, we present experimental evidence on the practical utility and limitations of using certified training towards empirical robustness. We show that, when tuned for the purpose, a recent certified training algorithm can prevent catastrophic overfitting on single-step attacks, and that it can bridge the gap to multi-step baselines under appropriate experimental settings. Finally, we present a conceptually simple regularizer for network over-approximations that can achieve similar effects while markedly reducing runtime.
comment: Transactions on Machine Learning Research, 2025
♻ ☆ Optimal Approximation of Zonoids and Uniform Approximation by Shallow Neural Networks
We study the following two related problems. The first is to determine to what error an arbitrary zonoid in $\mathbb{R}^{d+1}$ can be approximated in the Hausdorff distance by a sum of $n$ line segments. The second is to determine optimal approximation rates in the uniform norm for shallow ReLU$^k$ neural networks on their variation spaces. The first of these problems has been solved for $d\neq 2,3$, but when $d=2,3$ a logarithmic gap between the best upper and lower bounds remains. We close this gap, which completes the solution in all dimensions. For the second problem, our techniques significantly improve upon existing approximation rates when $k\geq 1$, and enable uniform approximation of both the target function and its derivatives.
♻ ☆ Beyond Training: Dynamic Token Merging for Zero-Shot Video Understanding
Recent advancements in multimodal large language models (MLLMs) have opened new avenues for video understanding. However, achieving high fidelity in zero-shot video tasks remains challenging. Traditional video processing methods rely heavily on fine-tuning to capture nuanced spatial-temporal details, which incurs significant data and computation costs. In contrast, training-free approaches, though efficient, often lack robustness in preserving context-rich features across complex video content. To this end, we propose DYTO, a novel dynamic token merging framework for zero-shot video understanding that adaptively optimizes token efficiency while preserving crucial scene details. DYTO integrates a hierarchical frame selection and a bipartite token merging strategy to dynamically cluster key frames and selectively compress token sequences, striking a balance between computational efficiency with semantic richness. Extensive experiments across multiple benchmarks demonstrate the effectiveness of DYTO, achieving superior performance compared to both fine-tuned and training-free methods and setting a new state-of-the-art for zero-shot video understanding.
comment: Code is available at https://github.com/Jam1ezhang/DYTO
♻ ☆ RankCLIP: Ranking-Consistent Language-Image Pretraining
Self-supervised contrastive learning models, such as CLIP, have set new benchmarks for vision-language models in many downstream tasks. However, their dependency on rigid one-to-one mappings overlooks the complex and often multifaceted relationships between and within texts and images. To this end, we introduce RankCLIP, a novel pre-training method that extends beyond the rigid one-to-one matching framework of CLIP and its variants. By extending the traditional pair-wise loss to list-wise, and leveraging both in-modal and cross-modal ranking consistency, RankCLIP improves the alignment process, enabling it to capture the nuanced many-to-many relationships between and within each modality. Through comprehensive experiments, we demonstrate the effectiveness of RankCLIP in various downstream tasks, notably achieving significant gains in zero-shot classifications over state-of-the-art methods, underscoring the importance of this enhanced learning process.
comment: Code and model checkpoints are available at https://github.com/Jam1ezhang/RankCLIP
♻ ☆ Benchmarking Post-Training Quantization in LLMs: Comprehensive Taxonomy, Unified Evaluation, and Comparative Analysis
Post-training Quantization (PTQ) technique has been extensively adopted for large language models (LLMs) compression owing to its efficiency and low resource requirement. However, current research lacks a in-depth analysis of the superior and applicable scenarios of each PTQ strategy. In addition, existing algorithms focus primarily on performance, overlooking the trade-off among model size, performance, and quantization bitwidth. To mitigate these confusions, we provide a novel benchmark for LLMs PTQ in this paper. Firstly, in order to support our benchmark, we propose a comprehensive taxonomy for existing mainstream methods by scrutinizing their computational strategies (e.g., optimization-based, compensation-based, etc.). Then, we conduct extensive experiments with the baseline within each class, covering models with various sizes (7B-70B), bitwidths, training levels (LLaMA1/2/3/3.1), architectures (Mixtral, DeepSeekMoE and Mamba) and modality (LLaVA1.5 and VILA1.5) on a wide range of evaluation metrics.Through comparative analysis on the results, we summarize the superior of each PTQ strategy and modelsize-bitwidth trade-off considering the performance. For example, our benchmark reveals that compensation-based technique demonstrates outstanding cross-architecture robustness and extremely low-bit PTQ for ultra large models should be reexamined. Finally, we further accordingly claim that a practical combination of compensation and other PTQ strategy can achieve SOTA various robustness. We believe that our benchmark will provide valuable recommendations for the deployment of LLMs and future research on PTQ approaches.We conduct an repository for our benchmark at https://github.com/zjq0455/PTQ_Benchmark.
comment: 17 pages, 3 fugures
♻ ☆ Learnability in Online Kernel Selection with Memory Constraint via Data-dependent Regret Analysis
Online kernel selection is a fundamental problem of online kernel methods.In this paper,we study online kernel selection with memory constraint in which the memory of kernel selection and online prediction procedures is limited to a fixed budget. An essential question is what is the intrinsic relationship among online learnability, memory constraint, and data complexity? To answer the question,it is necessary to show the trade-offs between regret and memory constraint.Previous work gives a worst-case lower bound depending on the data size,and shows learning is impossible within a small memory constraint.In contrast, we present distinct results by offering data-dependent upper bounds that rely on two data complexities:kernel alignment and the cumulative losses of competitive hypothesis.We propose an algorithmic framework giving data-dependent upper bounds for two types of loss functions.For the hinge loss function,our algorithm achieves an expected upper bound depending on kernel alignment.For smooth loss functions,our algorithm achieves a high-probability upper bound depending on the cumulative losses of competitive hypothesis.We also prove a matching lower bound for smooth loss functions.Our results show that if the two data complexities are sub-linear,then learning is possible within a small memory constraint.Our algorithmic framework depends on a new buffer maintaining framework and a reduction from online kernel selection to prediction with expert advice. Finally,we empirically verify the prediction performance of our algorithms on benchmark datasets.
♻ ☆ DUNE: Distilling a Universal Encoder from Heterogeneous 2D and 3D Teachers CVPR-2025
Recent multi-teacher distillation methods have unified the encoders of multiple foundation models into a single encoder, achieving competitive performance on core vision tasks like classification, segmentation, and depth estimation. This led us to ask: Could similar success be achieved when the pool of teachers also includes vision models specialized in diverse tasks across both 2D and 3D perception? In this paper, we define and investigate the problem of heterogeneous teacher distillation, or co-distillation, a challenging multi-teacher distillation scenario where teacher models vary significantly in both (a) their design objectives and (b) the data they were trained on. We explore data-sharing strategies and teacher-specific encoding, and introduce DUNE, a single encoder excelling in 2D vision, 3D understanding, and 3D human perception. Our model achieves performance comparable to that of its larger teachers, sometimes even outperforming them, on their respective tasks. Notably, DUNE surpasses MASt3R in Map-free Visual Relocalization with a much smaller encoder.
comment: Accepted to CVPR-2025. Project page: https://europe.naverlabs.com/dune
♻ ☆ DiTEC-WDN: A Large-Scale Dataset of Hydraulic Scenarios across Multiple Water Distribution Networks
Privacy restrictions hinder the sharing of real-world Water Distribution Network (WDN) models, limiting the application of emerging data-driven machine learning, which typically requires extensive observations. To address this challenge, we propose the dataset DiTEC-WDN that comprises 36,000 unique scenarios simulated over either short-term (24 hours) or long-term (1 year) periods. We constructed this dataset using an automated pipeline that optimizes crucial parameters (e.g., pressure, flow rate, and demand patterns), facilitates large-scale simulations, and records discrete, synthetic but hydraulically realistic states under standard conditions via rule validation and post-hoc analysis. With a total of 228 million generated graph-based states, DiTEC-WDN can support a variety of machine-learning tasks, including graph-level, node-level, and link-level regression, as well as time-series forecasting. This contribution, released under a public license, encourages open scientific research in the critical water sector, eliminates the risk of exposing sensitive data, and fulfills the need for a large-scale water distribution network benchmark for study comparisons and scenario analysis.
comment: Submitted to Nature Scientific Data. Huy Truong and Andr\'es Tello contributed equally to this work. For the dataset, see https://huggingface.co/datasets/rugds/ditec-wdn
♻ ☆ Understanding Model Calibration -- A gentle introduction and visual exploration of calibration and the expected calibration error (ECE)
To be considered reliable, a model must be calibrated so that its confidence in each decision closely reflects its true outcome. In this blogpost we'll take a look at the most commonly used definition for calibration and then dive into a frequently used evaluation measure for model calibration. We'll then cover some of the drawbacks of this measure and how these surfaced the need for additional notions of calibration, which require their own new evaluation measures. This post is not intended to be an in-depth dissection of all works on calibration, nor does it focus on how to calibrate models. Instead, it is meant to provide a gentle introduction to the different notions and their evaluation measures as well as to re-highlight some issues with a measure that is still widely used to evaluate calibration.
♻ ☆ GAEA: A Geolocation Aware Conversational Model
Image geolocalization, in which, traditionally, an AI model predicts the precise GPS coordinates of an image is a challenging task with many downstream applications. However, the user cannot utilize the model to further their knowledge other than the GPS coordinate; the model lacks an understanding of the location and the conversational ability to communicate with the user. In recent days, with tremendous progress of large multimodal models (LMMs) -- proprietary and open-source -- researchers have attempted to geolocalize images via LMMs. However, the issues remain unaddressed; beyond general tasks, for more specialized downstream tasks, one of which is geolocalization, LMMs struggle. In this work, we propose to solve this problem by introducing a conversational model GAEA that can provide information regarding the location of an image, as required by a user. No large-scale dataset enabling the training of such a model exists. Thus we propose GAEA-1.6M, a comprehensive dataset with 800K images and around 1.6M question-answer pairs constructed by leveraging OpenStreetMap (OSM) attributes and geographical context clues. For quantitative evaluation, we propose a diverse benchmark, GAEA-Bench, comprising 4K image-text pairs to evaluate conversational capabilities equipped with diverse question types. We consider 11 state-of-the-art open-source and proprietary LMMs and demonstrate that GAEA significantly outperforms the best open-source model, LLaVA-OneVision by 25.69% and the best proprietary model, GPT-4o by 8.28%. Our dataset, model and codes are available.
comment: The dataset and code used in this submission is available at: https://ucf-crcv.github.io/GAEA/
♻ ☆ Free Hunch: Denoiser Covariance Estimation for Diffusion Models Without Extra Costs
The covariance for clean data given a noisy observation is an important quantity in many training-free guided generation methods for diffusion models. Current methods require heavy test-time computation, altering the standard diffusion training process or denoiser architecture, or making heavy approximations. We propose a new framework that sidesteps these issues by using covariance information that is available for free from training data and the curvature of the generative trajectory, which is linked to the covariance through the second-order Tweedie's formula. We integrate these sources of information using (i) a novel method to transfer covariance estimates across noise levels and (ii) low-rank updates in a given noise level. We validate the method on linear inverse problems, where it outperforms recent baselines, especially with fewer diffusion steps.
comment: 24 pages, 11 figures
♻ ☆ ETAP: Event-based Tracking of Any Point
Tracking any point (TAP) recently shifted the motion estimation paradigm from focusing on individual salient points with local templates to tracking arbitrary points with global image contexts. However, while research has mostly focused on driving the accuracy of models in nominal settings, addressing scenarios with difficult lighting conditions and high-speed motions remains out of reach due to the limitations of the sensor. This work addresses this challenge with the first event camera-based TAP method. It leverages the high temporal resolution and high dynamic range of event cameras for robust high-speed tracking, and the global contexts in TAP methods to handle asynchronous and sparse event measurements. We further extend the TAP framework to handle event feature variations induced by motion -- thereby addressing an open challenge in purely event-based tracking -- with a novel feature-alignment loss which ensures the learning of motion-robust features. Our method is trained with data from a new data generation pipeline and systematically ablated across all design decisions. Our method shows strong cross-dataset generalization and performs 136% better on the average Jaccard metric than the baselines. Moreover, on an established feature tracking benchmark, it achieves a 20% improvement over the previous best event-only method and even surpasses the previous best events-and-frames method by 4.1%. Our code is available at https://github.com/tub-rip/ETAP
comment: 17 pages, 15 figures, 8 tables. Project page: https://github.com/tub-rip/ETAP
♻ ☆ CarPlanner: Consistent Auto-regressive Trajectory Planning for Large-scale Reinforcement Learning in Autonomous Driving CVPR 2025
Trajectory planning is vital for autonomous driving, ensuring safe and efficient navigation in complex environments. While recent learning-based methods, particularly reinforcement learning (RL), have shown promise in specific scenarios, RL planners struggle with training inefficiencies and managing large-scale, real-world driving scenarios. In this paper, we introduce \textbf{CarPlanner}, a \textbf{C}onsistent \textbf{a}uto-\textbf{r}egressive \textbf{Planner} that uses RL to generate multi-modal trajectories. The auto-regressive structure enables efficient large-scale RL training, while the incorporation of consistency ensures stable policy learning by maintaining coherent temporal consistency across time steps. Moreover, CarPlanner employs a generation-selection framework with an expert-guided reward function and an invariant-view module, simplifying RL training and enhancing policy performance. Extensive analysis demonstrates that our proposed RL framework effectively addresses the challenges of training efficiency and performance enhancement, positioning CarPlanner as a promising solution for trajectory planning in autonomous driving. To the best of our knowledge, we are the first to demonstrate that the RL-based planner can surpass both IL- and rule-based state-of-the-arts (SOTAs) on the challenging large-scale real-world dataset nuPlan. Our proposed CarPlanner surpasses RL-, IL-, and rule-based SOTA approaches within this demanding dataset.
comment: CVPR 2025
♻ ☆ Graph Structure Refinement with Energy-based Contrastive Learning AAAI 2025
Graph Neural Networks (GNNs) have recently gained widespread attention as a successful tool for analyzing graph-structured data. However, imperfect graph structure with noisy links lacks enough robustness and may damage graph representations, therefore limiting the GNNs' performance in practical tasks. Moreover, existing generative architectures fail to fit discriminative graph-related tasks. To tackle these issues, we introduce an unsupervised method based on a joint of generative training and discriminative training to learn graph structure and representation, aiming to improve the discriminative performance of generative models. We propose an Energy-based Contrastive Learning (ECL) guided Graph Structure Refinement (GSR) framework, denoted as ECL-GSR. To our knowledge, this is the first work to combine energy-based models with contrastive learning for GSR. Specifically, we leverage ECL to approximate the joint distribution of sample pairs, which increases the similarity between representations of positive pairs while reducing the similarity between negative ones. Refined structure is produced by augmenting and removing edges according to the similarity metrics among node representations. Extensive experiments demonstrate that ECL-GSR outperforms the state-of-the-art on eight benchmark datasets in node classification. ECL-GSR achieves faster training with fewer samples and memories against the leading baseline, highlighting its simplicity and efficiency in downstream tasks.
comment: Accepted to AAAI 2025
♻ ☆ On-device Federated Learning in Smartphones for Detecting Depression from Reddit Posts
Depression detection using deep learning models has been widely explored in previous studies, especially due to the large amounts of data available from social media posts. These posts provide valuable information about individuals' mental health conditions and can be leveraged to train models and identify patterns in the data. However, distributed learning approaches have not been extensively explored in this domain. In this study, we adopt Federated Learning (FL) to facilitate decentralized training on smartphones while protecting user data privacy. We train three neural network architectures--GRU, RNN, and LSTM on Reddit posts to detect signs of depression and evaluate their performance under heterogeneous FL settings. To optimize the training process, we leverage a common tokenizer across all client devices, which reduces the computational load. Additionally, we analyze resource consumption and communication costs on smartphones to assess their impact in a real-world FL environment. Our experimental results demonstrate that the federated models achieve comparable performance to the centralized models. This study highlights the potential of FL for decentralized mental health prediction by providing a secure and efficient model training process on edge devices.
comment: 11 pages, 7 figures
♻ ☆ Entropy annealing for policy mirror descent in continuous time and space
Entropy regularization has been widely used in policy optimization algorithms to enhance exploration and the robustness of the optimal control; however it also introduces an additional regularization bias. This work quantifies the impact of entropy regularization on the convergence of policy gradient methods for stochastic exit time control problems. We analyze a continuous-time policy mirror descent dynamics, which updates the policy based on the gradient of an entropy-regularized value function and adjusts the strength of entropy regularization as the algorithm progresses. We prove that with a fixed entropy level, the mirror descent dynamics converges exponentially to the optimal solution of the regularized problem. We further show that when the entropy level decays at suitable polynomial rates, the annealed flow converges to the solution of the unregularized problem at a rate of $\mathcal O(1/S)$ for discrete action spaces and, under suitable conditions, at a rate of $\mathcal O(1/\sqrt{S})$ for general action spaces, with $S$ being the gradient flow running time. The technical challenge lies in analyzing the gradient flow in the infinite-dimensional space of Markov kernels for nonconvex objectives. This paper explains how entropy regularization improves policy optimization, even with the true gradient, from the perspective of convergence rate.
♻ ☆ Truncated Kernel Stochastic Gradient Descent on Spheres
Inspired by the structure of spherical harmonics, we propose the truncated kernel stochastic gradient descent (T-kernel SGD) algorithm with a least-square loss function for spherical data fitting. T-kernel SGD introduces a novel regularization strategy by implementing stochastic gradient descent through a closed-form solution of the projection of the stochastic gradient in a low-dimensional subspace. In contrast to traditional kernel SGD, the regularization strategy implemented by T-kernel SGD is more effective in balancing bias and variance by dynamically adjusting the hypothesis space during iterations. The most significant advantage of the proposed algorithm is that it can achieve theoretically optimal convergence rates using a constant step size (independent of the sample size) while overcoming the inherent saturation problem of kernel SGD. Additionally, we leverage the structure of spherical polynomials to derive an equivalent T-kernel SGD, significantly reducing storage and computational costs compared to kernel SGD. Typically, T-kernel SGD requires only $\mathcal{O}(n^{1+\frac{d}{d-1}\epsilon})$ computational complexity and $\mathcal{O}(n^{\frac{d}{d-1}\epsilon})$ storage to achieve optimal rates for the d-dimensional sphere, where $0<\epsilon<\frac{1}{2}$ can be arbitrarily small if the optimal fitting or the underlying space possesses sufficient regularity. This regularity is determined by the smoothness parameter of the objective function and the decaying rate of the eigenvalues of the integral operator associated with the kernel function, both of which reflect the difficulty of the estimation problem. Our main results quantitatively characterize how this prior information influences the convergence of T-kernel SGD. The numerical experiments further validate the theoretical findings presented in this paper.
comment: 74 pages, 22 figures
♻ ☆ Glivenko-Cantelli for $f$-divergence
We extend the celebrated Glivenko-Cantelli theorem, sometimes called the fundamental theorem of statistics, from its standard setting of total variation distance to all $f$-divergences. A key obstacle in this endeavor is to define $f$-divergence on a subcollection of a $\sigma$-algebra that forms a $\pi$-system but not a $\sigma$-subalgebra. This is a side contribution of our work. We will show that this notion of $f$-divergence on the $\pi$-system of rays preserves nearly all known properties of standard $f$-divergence, yields a novel integral representation of the Kolmogorov-Smirnov distance, and has a Glivenko-Cantelli theorem. We will also discuss the prospects of a Vapnik-Chervonenkis theory for $f$-divergence.
comment: 26 pages, 1 figure
♻ ☆ Adaptive UAV-Assisted Hierarchical Federated Learning: Optimizing Energy, Latency, and Resilience for Dynamic Smart IoT
A key application of HFL lies in smart Internet of Things (IoT) systems, including remote monitoring and battlefield operations, where cellular connectivity is often unavailable. In such scenarios, UAVs can act as mobile aggregators, dynamically providing connectivity to terrestrial IoT devices. Subsequently, this paper investigates an HFL architecture enabled by energy-constrained, dynamically deployed UAVs that are susceptible to communication disruptions. We propose a novel approach to minimize global training costs in such environments by formulating a joint optimization problem that integrates learning configuration, bandwidth allocation, and IoT device-to-UAV association, ensuring timely global aggregation before UAV disconnections and redeployments. The problem explicitly captures the dynamic nature of IoT devices and their intermittent connectivity to UAVs and is shown to be NP-hard. To address its complexity, we decompose the problem into three interrelated subproblems. First, we optimize learning configuration and bandwidth allocation using an augmented Lagrangian function to reduce training costs. Second, we introduce a device fitness score that accounts for data heterogeneity (via Kullback-Leibler divergence), device-to-UAV proximity, and computational resources, leveraging a Twin Delayed Deep Deterministic Policy Gradient (TD3)-based algorithm for adaptive device-to-UAV assignment. Third, we develop a low-complexity two-stage greedy strategy for UAV redeployment and global aggregator selection, ensuring efficient model aggregation despite UAV disconnections.
comment: 18 pages, 43 figures
♻ ☆ Model Predictive Task Sampling for Efficient and Robust Adaptation
Foundation models have revolutionized general-purpose problem-solving, offering rapid task adaptation through pretraining, meta-training, and finetuning. Recent crucial advances in these paradigms reveal the importance of challenging task prioritized sampling to enhance adaptation robustness under distribution shifts. However, ranking task difficulties over iteration as a preliminary step typically requires exhaustive task evaluation, which is practically unaffordable in computation and data-annotation. This study provides a novel perspective to illuminate the possibility of leveraging the dual importance of adaptation robustness and learning efficiency, particularly in scenarios where task evaluation is risky or costly, such as iterative agent-environment interactions for robotic policy evaluation or computationally intensive inference steps for finetuning foundation models. Firstly, we introduce Model Predictive Task Sampling (MPTS), a framework that bridges the task space and adaptation risk landscape, providing a theoretical foundation for robust active task sampling. MPTS employs a generative model to characterize the episodic optimization process and predicts task-specific adaptation risk via posterior inference. The resulting risk learner amortizes the costly evaluation of task adaptation performance and provably approximates task difficulty rankings. MPTS seamlessly integrates into zero-shot, few-shot, and supervised finetuning settings. Empirically, we conduct extensive experiments in pattern recognition using foundation models and sequential decision-making. Our results demonstrate that MPTS significantly enhances adaptation robustness for tail or out-of-distribution (OOD) tasks and improves learning efficiency compared to state-of-the-art (SOTA) methods. The code is available at the project site https://github.com/thu-rllab/MPTS.
♻ ☆ ConFIG: Towards Conflict-free Training of Physics Informed Neural Networks ICLR2025
The loss functions of many learning problems contain multiple additive terms that can disagree and yield conflicting update directions. For Physics-Informed Neural Networks (PINNs), loss terms on initial/boundary conditions and physics equations are particularly interesting as they are well-established as highly difficult tasks. To improve learning the challenging multi-objective task posed by PINNs, we propose the ConFIG method, which provides conflict-free updates by ensuring a positive dot product between the final update and each loss-specific gradient. It also maintains consistent optimization rates for all loss terms and dynamically adjusts gradient magnitudes based on conflict levels. We additionally leverage momentum to accelerate optimizations by alternating the back-propagation of different loss terms. We provide a mathematical proof showing the convergence of the ConFIG method, and it is evaluated across a range of challenging PINN scenarios. ConFIG consistently shows superior performance and runtime compared to baseline methods. We also test the proposed method in a classic multi-task benchmark, where the ConFIG method likewise exhibits a highly promising performance. Source code is available at https://tum-pbs.github.io/ConFIG
comment: Accepted by ICLR2025 as spotlight paper; Project homepage: https://tum-pbs.github.io/ConFIG/
♻ ☆ MERGE$^3$: Efficient Evolutionary Merging on Consumer-grade GPUs
Evolutionary model merging enables the creation of high-performing multi-task models but remains computationally prohibitive for consumer hardware. We introduce MERGE$^3$, an efficient framework that makes evolutionary merging feasible on a single GPU by reducing fitness computation costs 50$\times$ while preserving performance. MERGE$^3$ achieves this by Extracting a reduced dataset for evaluation, Estimating model abilities using Item Response Theory (IRT), and Evolving optimal merges via IRT-based performance estimators. Our method enables state-of-the-art multilingual and cross-lingual merging, transferring knowledge across languages with significantly lower computational overhead. We provide theoretical guarantees and an open-source library, democratizing high-quality model merging.
comment: 19 pages, 13 figures
♻ ☆ Understanding Oversmoothing in Diffusion-Based GNNs From the Perspective of Operator Semigroup Theory KDD 2025
This paper presents an analytical study of the oversmoothing issue in diffusion-based Graph Neural Networks (GNNs). Generalizing beyond extant approaches grounded in random walk analysis or particle systems, we approach this problem through operator semigroup theory. This theoretical framework allows us to rigorously prove that oversmoothing is intrinsically linked to the ergodicity of the diffusion operator. Relying on semigroup method, we can quantitatively analyze the dynamic of graph diffusion and give a specific mathematical form of the smoothing feature by ergodicity and invariant measure of operator, which improves previous works only show existence of oversmoothing. This finding further poses a general and mild ergodicity-breaking condition, encompassing the various specific solutions previously offered, thereby presenting a more universal and theoretically grounded approach to relieve oversmoothing in diffusion-based GNNs. Additionally, we offer a probabilistic interpretation of our theory, forging a link with prior works and broadening the theoretical horizon. Our experimental results reveal that this ergodicity-breaking term effectively mitigates oversmoothing measured by Dirichlet energy, and simultaneously enhances performance in node classification tasks.
comment: Accepted by KDD 2025
♻ ☆ Toward a method for LLM-enabled Indoor Navigation
Indoor navigation presents unique challenges due to complex layouts, lack of GPS signals, and accessibility concerns. Existing solutions often struggle with real-time adaptability and user-specific needs. In this work, we explore the potential of a Large Language Model (LLM), i.e., ChatGPT, to generate natural, context-aware navigation instructions from indoor map images. We design and evaluate test cases across different real-world environments, analyzing the effectiveness of LLMs in interpreting spatial layouts, handling user constraints, and planning efficient routes. Our findings demonstrate the potential of LLMs for supporting personalized indoor navigation, with an average of 50.54% correct indications and a maximum of 77.78%. The results do not appear to depend on the complexity of the layout or the complexity of the expected path, but rather on the number of points of interest and the abundance of visual information, which negatively affect the performance.
comment: 7 pages, 3 figures, 5 tables
♻ ☆ Generalization of the Gibbs algorithm with high probability at low temperatures
The paper gives a bound on the generalization error of the Gibbs algorithm, which recovers known data-independent bounds for the high temperature range and extends to the low-temperature range, where generalization depends critically on the data-dependent loss-landscape. It is shown, that with high probability the generalization error of a single hypothesis drawn from the Gibbs posterior decreases with the total prior volume of all hypotheses with similar or smaller empirical error. This gives theoretical support to the belief in the benefit of flat minima. The zero temperature limit is discussed and the bound is extended to a class of similar stochastic algorithms.
♻ ☆ Efficient Training of Multi-task Neural Solver for Combinatorial Optimization
Efficiently training a multi-task neural solver for various combinatorial optimization problems (COPs) has been less studied so far. Naive application of conventional multi-task learning approaches often falls short in delivering a high-quality, unified neural solver. This deficiency primarily stems from the significant computational demands and a lack of adequate consideration for the complexities inherent in COPs. In this paper, we propose a general and efficient training paradigm to deliver a unified combinatorial multi-task neural solver. To this end, we resort to the theoretical loss decomposition for multiple tasks under an encoder-decoder framework, which enables more efficient training via proper bandit task-sampling algorithms through an intra-task influence matrix. By employing theoretically grounded approximations, our method significantly enhances overall performance, regardless of whether it is within constrained training budgets, across equivalent training epochs, or in terms of generalization capabilities, when compared to conventional training schedules. On the real-world datasets of TSPLib and CVRPLib, our method also achieved the best results compared to single task learning and multi-task learning approaches. Additionally, the influence matrix provides empirical evidence supporting common practices in the field of learning to optimize, further substantiating the effectiveness of our approach. Our code is open-sourced and available at https://github.com/LOGO-CUHKSZ/MTL-COP.
♻ ☆ Survival Analysis with Machine Learning for Predicting Li-ion Battery Remaining Useful Life
The accurate prediction of RUL for lithium-ion batteries is crucial for enhancing the reliability and longevity of energy storage systems. Traditional methods for RUL prediction often struggle with issues such as data sparsity, varying battery chemistries, and the inability to capture complex degradation patterns over time. In this study, we propose a survival analysis-based framework combined with deep learning models to predict the RUL of lithium-ion batteries. Specifically, we utilize five advanced models: the Cox-type models (Cox, CoxPH, and CoxTime) and two machine-learning-based models (DeepHit and MTLR). These models address the challenges of accurate RUL estimation by transforming raw time-series battery data into survival data, including key degradation indicators such as voltage, current, and internal resistance. Advanced feature extraction techniques enhance the model's robustness in diverse real-world scenarios, including varying charging conditions and battery chemistries. Our models are tested using 10-fold cross-validation, ensuring generalizability and minimizing overfitting. Experimental results show that our survival-based framework significantly improves RUL prediction accuracy compared to traditional methods, providing a reliable tool for battery management and maintenance optimization. This study contributes to the advancement of predictive maintenance in battery technology, offering valuable insights for both researchers and industry practitioners aiming to enhance the operational lifespan of lithium-ion batteries.
♻ ☆ Active Learning for Neural PDE Solvers
Solving partial differential equations (PDEs) is a fundamental problem in science and engineering. While neural PDE solvers can be more efficient than established numerical solvers, they often require large amounts of training data that is costly to obtain. Active learning (AL) could help surrogate models reach the same accuracy with smaller training sets by querying classical solvers with more informative initial conditions and PDE parameters. While AL is more common in other domains, it has yet to be studied extensively for neural PDE solvers. To bridge this gap, we introduce AL4PDE, a modular and extensible active learning benchmark. It provides multiple parametric PDEs and state-of-the-art surrogate models for the solver-in-the-loop setting, enabling the evaluation of existing and the development of new AL methods for neural PDE solving. We use the benchmark to evaluate batch active learning algorithms such as uncertainty- and feature-based methods. We show that AL reduces the average error by up to 71% compared to random sampling and significantly reduces worst-case errors. Moreover, AL generates similar datasets across repeated runs, with consistent distributions over the PDE parameters and initial conditions. The acquired datasets are reusable, providing benefits for surrogate models not involved in the data generation.
♻ ☆ Continual Low-Rank Scaled Dot-product Attention
Transformers are widely used for their ability to capture data relations in sequence processing, with great success for a wide range of static tasks. However, the computational and memory footprint of their main component, i.e., the Scaled Dot-product Attention, is commonly overlooked. This makes their adoption in applications involving stream data processing with constraints in response latency, computational and memory resources infeasible. Some works have proposed methods to lower the computational cost of Transformers, i.e. low-rank approximations, sparsity in attention, and efficient formulations for Continual Inference. In this paper, we introduce a new formulation of the Scaled Dot-product Attention based on the Nystr\"om approximation that is suitable for Continual Inference. In experiments on Online Audio Classification and Online Action Detection tasks, the proposed Continual Scaled Dot-product Attention can lower the number of operations by up to three orders of magnitude compared to the original Transformers while retaining the predictive performance of competing models.
comment: 14 pages, 7 figures
♻ ☆ Sparse Tensor PCA via Tensor Decomposition for Unsupervised Feature Selection
Recently, introducing Tensor Decomposition (TD) techniques into unsupervised feature selection (UFS) has been an emerging research topic. A tensor structure is beneficial for mining the relations between different modes and helps relieve the computation burden. However, while existing methods exploit TD to preserve the data tensor structure, they do not consider the influence of data orientation and thus have difficulty in handling orientation-specific data such as time series. To solve the above problem, we utilize the orientation-dependent tensor-tensor product from Tensor Singular Value Decomposition based on *M-product (T-SVDM) and extend the one-dimensional Sparse Principal Component Analysis (SPCA) to a tensor form. The proposed sparse tensor PCA model can constrain sparsity at the specified mode and yield sparse tensor principal components, enhancing flexibility and accuracy in learning feature relations. To ensure fast convergence and a flexible description of feature correlation, we develop a convex version specially designed for general UFS tasks and propose an efficient slice-by-slice algorithm that performs dual optimization in the transform domain. Experimental results on real-world datasets demonstrate the effectiveness and remarkable computational efficiency of the proposed method for tensor data of diverse structures over the state-of-the-arts. With a proper combination of data orientation and transform domain, our method is promising for various applications. The codes related to our proposed methods and the experiments are available at https://github.com/zjj20212035/STPCA.git.
♻ ☆ Forensics Adapter: Adapting CLIP for Generalizable Face Forgery Detection CVPR 2025
We describe the Forensics Adapter, an adapter network designed to transform CLIP into an effective and generalizable face forgery detector. Although CLIP is highly versatile, adapting it for face forgery detection is non-trivial as forgery-related knowledge is entangled with a wide range of unrelated knowledge. Existing methods treat CLIP merely as a feature extractor, lacking task-specific adaptation, which limits their effectiveness. To address this, we introduce an adapter to learn face forgery traces -- the blending boundaries unique to forged faces, guided by task-specific objectives. Then we enhance the CLIP visual tokens with a dedicated interaction strategy that communicates knowledge across CLIP and the adapter. Since the adapter is alongside CLIP, its versatility is highly retained, naturally ensuring strong generalizability in face forgery detection. With only 5.7M trainable parameters, our method achieves a significant performance boost, improving by approximately 7% on average across five standard datasets. We believe the proposed method can serve as a baseline for future CLIP-based face forgery detection methods. The code is available at https://github.com/OUC-VAS/ForensicsAdapter.
comment: CVPR 2025
♻ ☆ Hard constraint learning approaches with trainable influence functions for evolutionary equations
This paper develops a novel deep learning approach for solving evolutionary equations, which integrates sequential learning strategies with an enhanced hard constraint strategy featuring trainable parameters, addressing the low computational accuracy of standard Physics-Informed Neural Networks (PINNs) in large temporal domains.Sequential learning strategies divide a large temporal domain into multiple subintervals and solve them one by one in a chronological order, which naturally respects the principle of causality and improves the stability of the PINN solution. The improved hard constraint strategy strictly ensures the continuity and smoothness of the PINN solution at time interval nodes, and at the same time passes the information from the previous interval to the next interval, which avoids the incorrect/trivial solution at the position far from the initial time. Furthermore, by investigating the requirements of different types of equations on hard constraints, we design a novel influence function with trainable parameters for hard constraints, which provides theoretical and technical support for the effective implementations of hard constraint strategies, and significantly improves the universality and computational accuracy of our method. In addition, an adaptive time-domain partitioning algorithm is proposed, which plays an important role in the application of the proposed method as well as in the improvement of computational efficiency and accuracy. Numerical experiments verify the performance of the method. The data and code accompanying this paper are available at https://github.com/zhizhi4452/HCS.
♻ ☆ Distilled Prompt Learning for Incomplete Multimodal Survival Prediction CVPR2025
The integration of multimodal data including pathology images and gene profiles is widely applied in precise survival prediction. Despite recent advances in multimodal survival models, collecting complete modalities for multimodal fusion still poses a significant challenge, hindering their application in clinical settings. Current approaches tackling incomplete modalities often fall short, as they typically compensate for only a limited part of the knowledge of missing modalities. To address this issue, we propose a Distilled Prompt Learning framework (DisPro) to utilize the strong robustness of Large Language Models (LLMs) to missing modalities, which employs two-stage prompting for compensation of comprehensive information for missing modalities. In the first stage, Unimodal Prompting (UniPro) distills the knowledge distribution of each modality, preparing for supplementing modality-specific knowledge of the missing modality in the subsequent stage. In the second stage, Multimodal Prompting (MultiPro) leverages available modalities as prompts for LLMs to infer the missing modality, which provides modality-common information. Simultaneously, the unimodal knowledge acquired in the first stage is injected into multimodal inference to compensate for the modality-specific knowledge of the missing modality. Extensive experiments covering various missing scenarios demonstrated the superiority of the proposed method. The code is available at https://github.com/Innse/DisPro.
comment: Accepted by CVPR2025
♻ ☆ Adapt-$\infty$: Scalable Continual Multimodal Instruction Tuning via Dynamic Data Selection
Visual instruction datasets from various distributors are released at different times and often contain a significant number of semantically redundant text-image pairs, depending on their task compositions (i.e., skills) or reference sources. This redundancy greatly limits the efficient deployment of continually adaptable multimodal large language models, hindering their ability to refine existing skills and acquire new competencies over time. We reframe the problem of lifelong Instruction Tuning (LiIT) via data selection, where the model automatically selects beneficial samples to learn from earlier and new datasets based on the current state of acquired knowledge in the model. We propose Adapt-$\infty$, a new multi-way and adaptive data selection approach that dynamically balances sample efficiency and effectiveness during LiIT. We first construct pseudo-skill clusters by grouping gradient-based sample vectors. Next, we select the best-performing data selector for each skill cluster from a pool of selector experts, including our newly proposed scoring function, Image Grounding score. This data selector samples a subset of the most important samples from each skill cluster for training. To prevent the continuous increase in the size of the dataset pool during LiIT, we introduce a cluster-wise permanent data pruning strategy to remove the most semantically redundant samples from each cluster, keeping computational requirements manageable. We validate the effectiveness and efficiency of Adapt-$\infty$ over a sequence of multimodal instruction tuning datasets with various tasks, including (Knowledge) VQA, multilingual, grounding, reasoning, language-only, and multi-image comprehension. Training with samples selected by Adapt-$\infty$ alleviates catastrophic forgetting, especially for rare tasks, and promotes forward transfer across the continuum using only a fraction of the original data.
comment: First two authors contributed equally. Code: https://github.com/adymaharana/adapt-inf
♻ ☆ ThreatCrawl: A BERT-based Focused Crawler for the Cybersecurity Domain
Publicly available information contains valuable information for Cyber Threat Intelligence (CTI). This can be used to prevent attacks that have already taken place on other systems. Ideally, only the initial attack succeeds and all subsequent ones are detected and stopped. But while there are different standards to exchange this information, a lot of it is shared in articles or blog posts in non-standardized ways. Manually scanning through multiple online portals and news pages to discover new threats and extracting them is a time-consuming task. To automize parts of this scanning process, multiple papers propose extractors that use Natural Language Processing (NLP) to extract Indicators of Compromise (IOCs) from documents. However, while this already solves the problem of extracting the information out of documents, the search for these documents is rarely considered. In this paper, a new focused crawler is proposed called ThreatCrawl, which uses Bidirectional Encoder Representations from Transformers (BERT)-based models to classify documents and adapt its crawling path dynamically. While ThreatCrawl has difficulties to classify the specific type of Open Source Intelligence (OSINT) named in texts, e.g., IOC content, it can successfully find relevant documents and modify its path accord ingly. It yields harvest rates of up to 52%, which are, to the best of our knowledge, better than the current state of the art. The results and source code will be made publicly available upon acceptance.
comment: 11 pages, 9 figures, 5 tables
♻ ☆ Sample-Efficient "Clustering and Conquer" Procedures for Parallel Large-Scale Ranking and Selection
This work seeks to break the sample efficiency bottleneck in parallel large-scale ranking and selection (R&S) problems by leveraging correlation information. We modify the commonly used "divide and conquer" framework in parallel computing by adding a correlation-based clustering step, transforming it into "clustering and conquer". This seemingly simple modification achieves the optimal sample complexity reduction for a widely used class of efficient large-scale R&S procedures. Our approach enjoys two key advantages: 1) it does not require highly accurate correlation estimation or precise clustering, and 2) it allows for seamless integration with various existing R&S procedures, while achieving optimal sample complexity. Theoretically, we develop a novel gradient analysis framework to analyze sample efficiency and guide the design of large-scale R&S procedures. We also introduce a new parallel clustering algorithm tailored for large-scale scenarios. Finally, in large-scale AI applications such as neural architecture search, our methods demonstrate superior performance.
♻ ☆ Learning Causal Transition Matrix for Instance-dependent Label Noise
Noisy labels are both inevitable and problematic in machine learning methods, as they negatively impact models' generalization ability by causing overfitting. In the context of learning with noise, the transition matrix plays a crucial role in the design of statistically consistent algorithms. However, the transition matrix is often considered unidentifiable. One strand of methods typically addresses this problem by assuming that the transition matrix is instance-independent; that is, the probability of mislabeling a particular instance is not influenced by its characteristics or attributes. This assumption is clearly invalid in complex real-world scenarios. To better understand the transition relationship and relax this assumption, we propose to study the data generation process of noisy labels from a causal perspective. We discover that an unobservable latent variable can affect either the instance itself, the label annotation procedure, or both, which complicates the identification of the transition matrix. To address various scenarios, we have unified these observations within a new causal graph. In this graph, the input instance is divided into a noise-resistant component and a noise-sensitive component based on whether they are affected by the latent variable. These two components contribute to identifying the ``causal transition matrix'', which approximates the true transition matrix with theoretical guarantee. In line with this, we have designed a novel training framework that explicitly models this causal relationship and, as a result, achieves a more accurate model for inferring the clean label.
♻ ☆ Cross-domain Random Pre-training with Prototypes for Reinforcement Learning
This work has been submitted to the IEEE for possible publication. Copyright may be transferred without notice, after which this version may no longer be accessible. Unsupervised cross-domain Reinforcement Learning (RL) pre-training shows great potential for challenging continuous visual control but poses a big challenge. In this paper, we propose \textbf{C}ross-domain \textbf{R}andom \textbf{P}re-\textbf{T}raining with \textbf{pro}totypes (CRPTpro), a novel, efficient, and effective self-supervised cross-domain RL pre-training framework. CRPTpro decouples data sampling from encoder pre-training, proposing decoupled random collection to easily and quickly generate a qualified cross-domain pre-training dataset. Moreover, a novel prototypical self-supervised algorithm is proposed to pre-train an effective visual encoder that is generic across different domains. Without finetuning, the cross-domain encoder can be implemented for challenging downstream tasks defined in different domains, either seen or unseen. Compared with recent advanced methods, CRPTpro achieves better performance on downstream policy learning without extra training on exploration agents for data collection, greatly reducing the burden of pre-training. We conduct extensive experiments across eight challenging continuous visual-control domains, including balance control, robot locomotion, and manipulation. CRPTpro significantly outperforms the next best Proto-RL(C) on 11/12 cross-domain downstream tasks with only 54.5\% wall-clock pre-training time, exhibiting state-of-the-art pre-training performance with greatly improved pre-training efficiency.
comment: This work has been submitted to the IEEE for possible publication
♻ ☆ Using Anomaly Detection to Detect Poisoning Attacks in Federated Learning Applications
Adversarial attacks such as poisoning attacks have attracted the attention of many machine learning researchers. Traditionally, poisoning attacks attempt to inject adversarial training data in order to manipulate the trained model. In federated learning (FL), data poisoning attacks can be generalized to model poisoning attacks, which cannot be detected by simpler methods due to the lack of access to local training data by the detector. State-of-the-art poisoning attack detection methods for FL have various weaknesses, e.g., the number of attackers has to be known or not high enough, working with i.i.d. data only, and high computational complexity. To overcome above weaknesses, we propose a novel framework for detecting poisoning attacks in FL, which employs a reference model based on a public dataset and an auditor model to detect malicious updates. We implemented a detector based on the proposed framework and using a one-class support vector machine (OC-SVM), which reaches the lowest possible computational complexity O(K) where K is the number of clients. We evaluated our detector's performance against state-of-the-art (SOTA) poisoning attacks for two typical applications of FL: electrocardiograph (ECG) classification and human activity recognition (HAR). Our experimental results validated the performance of our detector over other SOTA detection methods.
comment: We will updated this article soon
♻ ☆ A Deep Learning Model for Coronary Artery Segmentation and Quantitative Stenosis Detection in Angiographic Images
Coronary artery disease (CAD) is a leading cause of cardiovascular-related mortality, and accurate stenosis detection is crucial for effective clinical decision-making. Coronary angiography remains the gold standard for diagnosing CAD, but manual analysis of angiograms is prone to errors and subjectivity. This study aims to develop a deep learning-based approach for the automatic segmentation of coronary arteries from angiographic images and the quantitative detection of stenosis, thereby improving the accuracy and efficiency of CAD diagnosis. We propose a novel deep learning-based method for the automatic segmentation of coronary arteries in angiographic images, coupled with a dynamic cohort method for stenosis detection. The segmentation model combines the MedSAM and VM-UNet architectures to achieve high-performance results. After segmentation, the vascular centerline is extracted, vessel diameter is computed, and the degree of stenosis is measured with high precision, enabling accurate identification of arterial stenosis. On the mixed dataset (including the ARCADE, DCA1, and GH datasets), the model achieved an average IoU of 0.6308, with sensitivity and specificity of 0.9772 and 0.9903, respectively. On the ARCADE dataset, the average IoU was 0.6303, with sensitivity of 0.9832 and specificity of 0.9933. Additionally, the stenosis detection algorithm achieved a true positive rate (TPR) of 0.5867 and a positive predictive value (PPV) of 0.5911, demonstrating the effectiveness of our model in analyzing coronary angiography images. SAM-VMNet offers a promising tool for the automated segmentation and detection of coronary artery stenosis. The model's high accuracy and robustness provide significant clinical value for the early diagnosis and treatment planning of CAD. The code and examples are available at https://github.com/qimingfan10/SAM-VMNet.
♻ ☆ Lessons and Insights from a Unifying Study of Parameter-Efficient Fine-Tuning (PEFT) in Visual Recognition CVPR 2025
Parameter-efficient fine-tuning (PEFT) has attracted significant attention due to the growth of pre-trained model sizes and the need to fine-tune (FT) them for superior downstream performance. Despite a surge in new PEFT methods, a systematic study to understand their performance and suitable application scenarios is lacking, leaving questions like "when to apply PEFT" and "which method to use" largely unanswered, especially in visual recognition. In this paper, we conduct a unifying empirical study of representative PEFT methods with Vision Transformers. We systematically tune their hyper-parameters to fairly compare their accuracy on downstream tasks. Our study offers a practical user guide and unveils several new insights. First, if tuned carefully, different PEFT methods achieve similar accuracy in the low-shot benchmark VTAB-1K. This includes simple approaches like FT the bias terms that were reported inferior. Second, despite similar accuracy, we find that PEFT methods make different mistakes and high-confidence predictions, likely due to their different inductive biases. Such an inconsistency (or complementarity) opens up the opportunity for ensemble methods, and we make preliminary attempts at this. Third, going beyond the commonly used low-shot tasks, we find that PEFT is also useful in many-shot regimes, achieving comparable or better accuracy than full FT while using significantly fewer parameters. Lastly, we investigate PEFT's ability to preserve a pre-trained model's robustness to distribution shifts (e.g., CLIP). Perhaps not surprisingly, PEFT approaches outperform full FT alone. However, with weight-space ensembles, full FT can better balance target distribution and distribution shift performance, suggesting a future research direction for robust PEFT.
comment: CVPR 2025. The code is available at https://github.com/OSU-MLB/ViT_PEFT_Vision
♻ ☆ On the Performance Analysis of Momentum Method: A Frequency Domain Perspective ICLR 2025
Momentum-based optimizers are widely adopted for training neural networks. However, the optimal selection of momentum coefficients remains elusive. This uncertainty impedes a clear understanding of the role of momentum in stochastic gradient methods. In this paper, we present a frequency domain analysis framework that interprets the momentum method as a time-variant filter for gradients, where adjustments to momentum coefficients modify the filter characteristics. Our experiments support this perspective and provide a deeper understanding of the mechanism involved. Moreover, our analysis reveals the following significant findings: high-frequency gradient components are undesired in the late stages of training; preserving the original gradient in the early stages, and gradually amplifying low-frequency gradient components during training both enhance performance. Based on these insights, we propose Frequency Stochastic Gradient Descent with Momentum (FSGDM), a heuristic optimizer that dynamically adjusts the momentum filtering characteristic with an empirically effective dynamic magnitude response. Experimental results demonstrate the superiority of FSGDM over conventional momentum optimizers.
comment: ICLR 2025. 22 pages, 14 figures. Keywords: Momentum Method, Stochastic Gradient Descent, Z-Transform, Frequency Domain Analysis, Deep Learning
♻ ☆ Reinforcement Learning: A Comprehensive Overview
This manuscript gives a big-picture, up-to-date overview of the field of (deep) reinforcement learning and sequential decision making, covering value-based method, policy-gradient methods, model-based methods, and various other topics (e.g., multi-agent RL, RL+LLMs, and RL+inference).
♻ ☆ Implementing Fairness in AI Classification: The Role of Explainability
In this paper, we propose a philosophical and experimental investigation of the problem of AI fairness in classification. We argue that implementing fairness in AI classification involves more work than just operationalizing a fairness metric. It requires establishing the explainability of the classification model chosen and of the principles behind it. Specifically, it involves making the training processes transparent, determining what outcomes the fairness criteria actually produce, and assessing their trade-offs by comparison with closely related models that would lead to a different outcome. To exemplify this methodology, we trained a model and developed a tool for disparity detection and fairness interventions, the package FairDream. While FairDream is set to enforce Demographic Parity, experiments reveal that it fulfills the constraint of Equalized Odds. The algorithm is thus more conservative than the user might expect. To justify this outcome, we first clarify the relation between Demographic Parity and Equalized Odds as fairness criteria. We then explain FairDream's reweighting method and justify the trade-offs reached by FairDream by a benchmark comparison with closely related GridSearch models. We draw conclusions regarding the way in which these explanatory steps can make an AI model trustworthy.
♻ ☆ Parameter Efficient Mamba Tuning via Projector-targeted Diagonal-centric Linear Transformation CVPR 2025
Despite the growing interest in Mamba architecture as a potential replacement for Transformer architecture, parameter-efficient fine-tuning (PEFT) approaches for Mamba remain largely unexplored. In our study, we introduce two key insights-driven strategies for PEFT in Mamba architecture: (1) While state-space models (SSMs) have been regarded as the cornerstone of Mamba architecture, then expected to play a primary role in transfer learning, our findings reveal that Projectors -- not SSMs -- are the predominant contributors to transfer learning. (2) Based on our observation, we propose a novel PEFT method specialized to Mamba architecture: Projector-targeted Diagonal-centric Linear Transformation (ProDiaL). ProDiaL focuses on optimizing only the pretrained Projectors for new tasks through diagonal-centric linear transformation matrices, without directly fine-tuning the Projector weights. This targeted approach allows efficient task adaptation, utilizing less than 1% of the total parameters, and exhibits strong performance across both vision and language Mamba models, highlighting its versatility and effectiveness.
comment: accepted in CVPR 2025
♻ ☆ AnDB: Breaking Boundaries with an AI-Native Database for Universal Semantic Analysis
In this demonstration, we present AnDB, an AI-native database that supports traditional OLTP workloads and innovative AI-driven tasks, enabling unified semantic analysis across structured and unstructured data. While structured data analytics is mature, challenges remain in bridging the semantic gap between user queries and unstructured data. AnDB addresses these issues by leveraging cutting-edge AI-native technologies, allowing users to perform semantic queries using intuitive SQL-like statements without requiring AI expertise. This approach eliminates the ambiguity of traditional text-to-SQL systems and provides a seamless end-to-end optimization for analyzing all data types. AnDB automates query processing by generating multiple execution plans and selecting the optimal one through its optimizer, which balances accuracy, execution time, and financial cost based on user policies and internal optimizing mechanisms. AnDB future-proofs data management infrastructure, empowering users to effectively and efficiently harness the full potential of all kinds of data without starting from scratch.
comment: 4 pages, 5 figures
♻ ☆ A Mechanistic Explanatory Strategy for XAI
Despite significant advancements in XAI, scholars note a persistent lack of solid conceptual foundations and integration with broader discourse on scientific explanation. In response, emerging XAI research draws on explanatory strategies from various sciences and philosophy of science literature to fill these gaps. This paper outlines a mechanistic strategy for explaining the functional organization of deep learning systems, situating recent advancements in AI explainability within a broader philosophical context. According to the mechanistic approach, the explanation of opaque AI systems involves identifying mechanisms that drive decision-making. For deep neural networks, this means discerning functionally relevant components - such as neurons, layers, circuits, or activation patterns - and understanding their roles through decomposition, localization, and recomposition. Proof-of-principle case studies from image recognition and language modeling align these theoretical approaches with the latest research from AI labs like OpenAI and Anthropic. The paper suggests that such a systematic approach to studying model organization can reveal elements that individual explainability techniques might miss, fostering more thoroughly explainable AI.
comment: Forthcoming in M\"uller, V. C., Dung, L., L\"ohr, G., & Rumana, A. (Eds.). Philosophy of Artificial Intelligence: The State of the Art, Synthese Library, Springer Nature. Please cite the published version
♻ ☆ TokenFormer: Rethinking Transformer Scaling with Tokenized Model Parameters ICLR
Transformers have become the predominant architecture in foundation models due to their excellent performance across various domains. However, the substantial cost of scaling these models remains a significant concern. This problem arises primarily from their dependence on a fixed number of parameters within linear projections. When architectural modifications (e.g., channel dimensions) are introduced, the entire model typically requires retraining from scratch. As model sizes continue growing, this strategy results in increasingly high computational costs and becomes unsustainable. To overcome this problem, we introduce TokenFormer, a natively scalable architecture that leverages the attention mechanism not only for computations among input tokens but also for interactions between tokens and model parameters, thereby enhancing architectural flexibility. By treating model parameters as tokens, we replace all the linear projections in Transformers with our token-parameter attention layer, where input tokens act as queries and model parameters as keys and values. This reformulation allows for progressive and efficient scaling without necessitating retraining from scratch. Our model scales from 124M to 1.4B parameters by incrementally adding new key-value parameter pairs, achieving performance comparable to Transformers trained from scratch while greatly reducing training costs. Code and models are available at https://github.com/Haiyang-W/TokenFormer.
comment: Accepted by ICLR for a spotlight presentation
♻ ☆ Leveraging Allophony in Self-Supervised Speech Models for Atypical Pronunciation Assessment NAACL 2025
Allophony refers to the variation in the phonetic realization of a phoneme based on its phonetic environment. Modeling allophones is crucial for atypical pronunciation assessment, which involves distinguishing atypical from typical pronunciations. However, recent phoneme classifier-based approaches often simplify this by treating various realizations as a single phoneme, bypassing the complexity of modeling allophonic variation. Motivated by the acoustic modeling capabilities of frozen self-supervised speech model (S3M) features, we propose MixGoP, a novel approach that leverages Gaussian mixture models to model phoneme distributions with multiple subclusters. Our experiments show that MixGoP achieves state-of-the-art performance across four out of five datasets, including dysarthric and non-native speech. Our analysis further suggests that S3M features capture allophonic variation more effectively than MFCCs and Mel spectrograms, highlighting the benefits of integrating MixGoP with S3M features.
comment: Accepted to NAACL 2025. Codebase available at https://github.com/juice500ml/acoustic-units-for-ood
♻ ☆ Distributional bias compromises leave-one-out cross-validation
Cross-validation is a common method for estimating the predictive performance of machine learning models. In a data-scarce regime, where one typically wishes to maximize the number of instances used for training the model, an approach called "leave-one-out cross-validation" is often used. In this design, a separate model is built for predicting each data instance after training on all other instances. Since this results in a single test instance available per model trained, predictions are aggregated across the entire dataset to calculate common performance metrics such as the area under the receiver operating characteristic or R2 scores. In this work, we demonstrate that this approach creates a negative correlation between the average label of each training fold and the label of its corresponding test instance, a phenomenon that we term distributional bias. As machine learning models tend to regress to the mean of their training data, this distributional bias tends to negatively impact performance evaluation and hyperparameter optimization. We show that this effect generalizes to leave-P-out cross-validation and persists across a wide range of modeling and evaluation approaches, and that it can lead to a bias against stronger regularization. To address this, we propose a generalizable rebalanced cross-validation approach that corrects for distributional bias for both classification and regression. We demonstrate that our approach improves cross-validation performance evaluation in synthetic simulations, across machine learning benchmarks, and in several published leave-one-out analyses.
comment: 29 pages, 6 figures, supplementary information
♻ ☆ Distributionally Robust Learning for Multi-source Unsupervised Domain Adaptation
Empirical risk minimization often performs poorly when the distribution of the target domain differs from those of source domains. To address such potential distribution shifts, we develop an unsupervised domain adaptation approach that leverages labeled data from multiple source domains and unlabeled data from the target domain. We introduce a distributionally robust model that optimizes an adversarial reward based on the explained variance across a class of target distributions, ensuring generalization to the target domain. We show that the proposed robust model is a weighted average of conditional outcome models from source domains. This formulation allows us to compute the robust model through the aggregation of source models, which can be estimated using various machine learning algorithms of the users' choice, such as random forests, boosting, and neural networks. Additionally, we introduce a bias-correction step to obtain a more accurate aggregation weight, which is effective for various machine learning algorithms. Our framework can be interpreted as a distributionally robust federated learning approach that satisfies privacy constraints while providing insights into the importance of each source for prediction on the target domain. The performance of our method is evaluated on both simulated and real data.
♻ ☆ Neural Representation for Wireless Radiation Field Reconstruction: A 3D Gaussian Splatting Approach
Wireless channel modeling plays a pivotal role in designing, analyzing, and optimizing wireless communication systems. Nevertheless, developing an effective channel modeling approach has been a long-standing challenge. This issue has been escalated due to denser network deployment, larger antenna arrays, and broader bandwidth in next-generation networks. To address this challenge, we put forth WRF-GS, a novel framework for channel modeling based on wireless radiation field (WRF) reconstruction using 3D Gaussian splatting (3D-GS). WRF-GS employs 3D Gaussian primitives and neural networks to capture the interactions between the environment and radio signals, enabling efficient WRF reconstruction and visualization of the propagation characteristics. The reconstructed WRF can then be used to synthesize the spatial spectrum for comprehensive wireless channel characterization. While WRF-GS demonstrates remarkable effectiveness, it faces limitations in capturing high-frequency signal variations caused by complex multipath effects. To overcome these limitations, we propose WRF-GS+, an enhanced framework that integrates electromagnetic wave physics into the neural network design. WRF-GS+ leverages deformable 3D Gaussians to model both static and dynamic components of the WRF, significantly improving its ability to characterize signal variations. In addition, WRF-GS+ enhances the splatting process by simplifying the 3D-GS modeling process and improving computational efficiency. Experimental results demonstrate that both WRF-GS and WRF-GS+ outperform baselines for spatial spectrum synthesis, including ray tracing and other deep-learning approaches. Notably, WRF-GS+ achieves state-of-the-art performance in the received signal strength indication (RSSI) and channel state information (CSI) prediction tasks, surpassing existing methods by more than 0.7 dB and 3.36 dB, respectively.
comment: This is an extended journal version of our previous conference paper that was accepted to the IEEE INFOCOM 2025 at arXiv:2412.04832v2. The code for this version is available at https://github.com/wenchaozheng/WRF-GSplus
♻ ☆ Similarity-Dissimilarity Loss for Multi-label Supervised Contrastive Learning
Supervised contrastive learning has been explored in making use of label information for multi-label classification, but determining positive samples in multi-label scenario remains challenging. Previous studies have examined strategies for identifying positive samples, considering label overlap proportion between anchors and samples. However, they ignore various relations between given anchors and samples, as well as how to dynamically adjust the weights in contrastive loss functions based on different relations, leading to great ambiguity. In this paper, we introduce five distinct relations between multi-label samples and propose a Similarity-Dissimilarity Loss with contrastive learning for multi-label classification. Our loss function re-weights the loss by computing the similarity and dissimilarity between positive samples and a given anchor based on the introduced relations. We mainly conduct experiments for multi-label text classification on MIMIC datasets, then further extend the evaluation on MS-COCO. The Experimental results show that our proposed loss effectively improves the performance on all encoders under supervised contrastive learning paradigm, demonstrating its effectiveness and robustness.
♻ ☆ Design and Implementation of an FPGA-Based Tiled Matrix Multiplication Accelerator for Transformer Self-Attention on the Xilinx KV260 SoM
Transformer-based large language models (LLMs) rely heavily on intensive matrix multiplications for attention and feed-forward layers, with the Q, K, and V linear projections in the Multi-Head Self-Attention (MHA) module constituting a decisive performance bottleneck. In this work, we introduce a highly optimized tiled matrix multiplication accelerator on a resource-constrained Xilinx KV260 FPGA that not only addresses this challenge but sets a new standard for efficiency and performance. Our design exploits persistent on-chip storage, a robust two-level tiling strategy for maximal data reuse, and a systolic-like unrolled compute engine that together deliver unparalleled speed and energy efficiency. Integrated with DistilBERT for Q, K, and V projections, our accelerator achieves an unequivocal 7x speedup over ARM CPU implementations (PyTorch) and an extraordinary 200x improvement over naive NumPy, reaching a throughput of up to 3.1~GFLOPs for matrix multiplications on (64,768) x (768,3072) matrices while operating at a conservative 100 MHz. These results decisively demonstrate the transformative potential of FPGA-based acceleration for critical Transformer operations, paving the way for scalable and energy-efficient deep learning inference on edge devices.
comment: 7 pages, 4 figures, 2 tables. Prepared in ACM conference style. Preprint under review
♻ ☆ Emulating Full Participation: An Effective and Fair Client Selection Strategy for Federated Learning
In federated learning, client selection is a critical problem that significantly impacts both model performance and fairness. Prior studies typically treat these two objectives separately, or balance them using simple weighting schemes. However, we observe that commonly used metrics for model performance and fairness often conflict with each other, and a straightforward weighted combination is insufficient to capture their complex interactions. To address this, we first propose two guiding principles that directly tackle the inherent conflict between the two metrics while reinforcing each other. Based on these principles, we formulate the client selection problem as a long-term optimization task, leveraging the Lyapunov function and the submodular nature of the problem to solve it effectively. Experiments show that the proposed method improves both model performance and fairness, guiding the system to converge comparably to full client participation. This improvement can be attributed to the fact that both model performance and fairness benefit from the diversity of the selected clients' data distributions. Our approach adaptively enhances this diversity by selecting clients based on their data distributions, thereby improving both model performance and fairness.
♻ ☆ Near-Optimal Sample Complexity for Iterated CVaR Reinforcement Learning with a Generative Model AISTATS 2025
In this work, we study the sample complexity problem of risk-sensitive Reinforcement Learning (RL) with a generative model, where we aim to maximize the Conditional Value at Risk (CVaR) with risk tolerance level $\tau$ at each step, a criterion we refer to as Iterated CVaR. We first build a connection between Iterated CVaR RL and $(s, a)$-rectangular distributional robust RL with a specific uncertainty set for CVaR. We establish nearly matching upper and lower bounds on the sample complexity of this problem. Specifically, we first prove that a value iteration-based algorithm, ICVaR-VI, achieves an $\epsilon$-optimal policy with at most $\tilde{O} \left(\frac{SA}{(1-\gamma)^4\tau^2\epsilon^2} \right)$ samples, where $\gamma$ is the discount factor, and $S, A$ are the sizes of the state and action spaces. Furthermore, when $\tau \geq \gamma$, the sample complexity improves to $\tilde{O} \left( \frac{SA}{(1-\gamma)^3\epsilon^2} \right)$. We further show a minimax lower bound of $\tilde{O} \left(\frac{(1-\gamma \tau)SA}{(1-\gamma)^4\tau\epsilon^2} \right)$. For a fixed risk level $\tau \in (0,1]$, our upper and lower bounds match, demonstrating the tightness and optimality of our analysis. We also investigate a limiting case with a small risk level $\tau$, called Worst-Path RL, where the objective is to maximize the minimum possible cumulative reward. We develop matching upper and lower bounds of $\tilde{O} \left(\frac{SA}{p_{\min}} \right)$, where $p_{\min}$ denotes the minimum non-zero reaching probability of the transition kernel.
comment: Accepted as a conference paper at AISTATS 2025
♻ ☆ Limits to Analog Reservoir Learning
Reservoir computation is a recurrent framework for learning and predicting time series data, that benefits from extremely simple training and interpretability, often as the the dynamics of a physical system. In this paper, we will study the impact of noise on the learning capabilities of analog reservoir computers. Recent work on reservoir computation has shown that the information processing capacity (IPC) is a useful metric for quantifying the degradation of the performance due to noise. We further this analysis and demonstrate that this degradation of the IPC limits the possible features that can be meaningfully constructed in an analog reservoir computing setting. We borrow a result from quantum complexity theory that relates the circuit model of computation to a continuous time model, and demonstrate an exponential reduction in the accessible volume of reservoir configurations. We conclude by relating this degradation in the IPC to the fat-shattering dimension of a family of functions describing the reservoir dynamics, which allows us to express our result in terms of a classification task. We conclude that any physical, analog reservoir computer that is exposed to noise can only be used to perform a polynomial amount of learning, despite the exponentially large latent space, even with an exponential amount of post-processing.
comment: 10 pages, 1 figure
♻ ☆ Differentially Private Kernel Density Estimation
We introduce a refined differentially private (DP) data structure for kernel density estimation (KDE), offering not only improved privacy-utility tradeoff but also better efficiency over prior results. Specifically, we study the mathematical problem: given a similarity function $f$ (or DP KDE) and a private dataset $X \subset \mathbb{R}^d$, our goal is to preprocess $X$ so that for any query $y\in\mathbb{R}^d$, we approximate $\sum_{x \in X} f(x, y)$ in a differentially private fashion. The best previous algorithm for $f(x,y) =\| x - y \|_1$ is the node-contaminated balanced binary tree by [Backurs, Lin, Mahabadi, Silwal, and Tarnawski, ICLR 2024]. Their algorithm requires $O(nd)$ space and time for preprocessing with $n=|X|$. For any query point, the query time is $d \log n$, with an error guarantee of $(1+\alpha)$-approximation and $\epsilon^{-1} \alpha^{-0.5} d^{1.5} R \log^{1.5} n$. In this paper, we improve the best previous result [Backurs, Lin, Mahabadi, Silwal, and Tarnawski, ICLR 2024] in three aspects: - We reduce query time by a factor of $\alpha^{-1} \log n$. - We improve the approximation ratio from $\alpha$ to 1. - We reduce the error dependence by a factor of $\alpha^{-0.5}$. From a technical perspective, our method of constructing the search tree differs from previous work [Backurs, Lin, Mahabadi, Silwal, and Tarnawski, ICLR 2024]. In prior work, for each query, the answer is split into $\alpha^{-1} \log n$ numbers, each derived from the summation of $\log n$ values in interval tree countings. In contrast, we construct the tree differently, splitting the answer into $\log n$ numbers, where each is a smart combination of two distance values, two counting values, and $y$ itself. We believe our tree structure may be of independent interest.
comment: v2: Appendix added. v3: Numerical validations added
♻ ☆ Training and Evaluating with Human Label Variation: An Empirical Study
Human label variation (HLV) challenges the standard assumption that a labelled instance has a single ground truth, instead embracing the natural variation in human annotation to train and evaluate models. While various training methods and metrics for HLV have been proposed, it is still unclear which methods and metrics perform best in what settings. We propose new evaluation metrics for HLV leveraging fuzzy set theory. Since these new proposed metrics are differentiable, we then in turn experiment with employing these metrics as training objectives. We conduct an extensive study over 6 HLV datasets testing 14 training methods and 6 evaluation metrics. We find that training on either disaggregated annotations or soft labels performs best across metrics, outperforming training using the proposed training objectives with differentiable metrics. We also show that our proposed soft metric is more interpretable and correlates best with human preference.
comment: 25 pages
Multimedia 5
☆ Merge Mode for Template-based Intra Mode Derivation (TIMD) in ECM
This paper presents an intra coding tool, named Merge mode for Template-based Intra Mode Derivation (TIMD). TIMD-Merge has been adopted in the 15\textsuperscript{th} version of the Enhanced Compression Model (ECM) software that explores video coding technologies beyond Versatile Video Coding (VVC) standard. This proposed tool operates on top of the regular TIMD mode that applies a template-based search on a causal adjacent template at top and at left of the current block, in order to find the best Intra Prediction Modes (IPMs) that matches the template. The proposed TIMD-Merge in this paper addresses a shortcoming in the regular TIMD method where due to texture discrepancy, the adjacent template information around the block is not reliable. To do so, the proposed TIMD-Merge constructs a list of all TIMD-coded blocks in relatively larger template area than the template of the regular TIMD, which also includes non-adjacent neighboring blocks. This list, called the merge list, is then sorted on the template to give one best set of TIMD modes. The use of TIMD-Merge mode is signalled at the block level and the implementation in the ECM-14.0 demonstrates -0.08\% performance improvement in terms of luma BDR gain, with negligible encoding and decoding runtime increase of 100.6% and 100.2%, respectively.
☆ EvAnimate: Event-conditioned Image-to-Video Generation for Human Animation
Conditional human animation transforms a static reference image into a dynamic sequence by applying motion cues such as poses. These motion cues are typically derived from video data but are susceptible to limitations including low temporal resolution, motion blur, overexposure, and inaccuracies under low-light conditions. In contrast, event cameras provide data streams with exceptionally high temporal resolution, a wide dynamic range, and inherent resistance to motion blur and exposure issues. In this work, we propose EvAnimate, a framework that leverages event streams as motion cues to animate static human images. Our approach employs a specialized event representation that transforms asynchronous event streams into 3-channel slices with controllable slicing rates and appropriate slice density, ensuring compatibility with diffusion models. Subsequently, a dual-branch architecture generates high-quality videos by harnessing the inherent motion dynamics of the event streams, thereby enhancing both video quality and temporal consistency. Specialized data augmentation strategies further enhance cross-person generalization. Finally, we establish a new benchmarking, including simulated event data for training and validation, and a real-world event dataset capturing human actions under normal and extreme scenarios. The experiment results demonstrate that EvAnimate achieves high temporal fidelity and robust performance in scenarios where traditional video-derived cues fall short.
♻ ☆ Grounded Chain-of-Thought for Multimodal Large Language Models
Despite great progress, existing multimodal large language models (MLLMs) are prone to visual hallucination, greatly impeding their trustworthy applications. In this paper, we study this problem from the perspective of visual-spatial reasoning, and propose a new learning task for MLLMs, termed Grounded Chain-of-Thought (GCoT). Different from recent visual CoT studies, which focus more on visual knowledge reasoning, GCoT is keen to helping MLLMs to recognize and ground the relevant visual cues step by step, thereby predicting the correct answer with grounding coordinates as the intuitive basis. To facilitate this task, we also carefully design and construct a dataset called multimodal grounded chain-of-thought (MM-GCoT) consisting of 24,022 GCoT examples for 5,033 images. Besides, a comprehensive consistency evaluation system is also introduced, including the metrics of answer accuracy, grounding accuracy and answer-grounding consistency. We further design and conduct a bunch of experiments on 12 advanced MLLMs, and reveal some notable findings: i. most MLLMs performs poorly on the consistency evaluation, indicating obvious visual hallucination; ii. visual hallucination is not directly related to the parameter size and general multimodal performance, i.e., a larger and stronger MLLM is not less affected by this issue. Lastly, we also demonstrate that the proposed dataset can help existing MLLMs to well cultivate their GCoT capability and reduce the inconsistent answering significantly. Moreover, their GCoT can be also generalized to exiting multimodal tasks, such as open-world QA and REC.
♻ ☆ STA-V2A: Video-to-Audio Generation with Semantic and Temporal Alignment ICASSP 2025
Visual and auditory perception are two crucial ways humans experience the world. Text-to-video generation has made remarkable progress over the past year, but the absence of harmonious audio in generated video limits its broader applications. In this paper, we propose Semantic and Temporal Aligned Video-to-Audio (STA-V2A), an approach that enhances audio generation from videos by extracting both local temporal and global semantic video features and combining these refined video features with text as cross-modal guidance. To address the issue of information redundancy in videos, we propose an onset prediction pretext task for local temporal feature extraction and an attentive pooling module for global semantic feature extraction. To supplement the insufficient semantic information in videos, we propose a Latent Diffusion Model with Text-to-Audio priors initialization and cross-modal guidance. We also introduce Audio-Audio Align, a new metric to assess audio-temporal alignment. Subjective and objective metrics demonstrate that our method surpasses existing Video-to-Audio models in generating audio with better quality, semantic consistency, and temporal alignment. The ablation experiment validated the effectiveness of each module. Audio samples are available at https://y-ren16.github.io/STAV2A.
comment: ICASSP 2025-2025 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)
♻ ☆ ReTaKe: Reducing Temporal and Knowledge Redundancy for Long Video Understanding
Video Large Language Models (VideoLLMs) have made significant strides in video understanding but struggle with long videos due to the limitations of their backbone LLMs. Existing solutions rely on length extrapolation, which is memory-constrained, or visual token compression, which primarily leverages low-level temporal redundancy while overlooking the more effective high-level knowledge redundancy. To address this, we propose $\textbf{ReTaKe}$, a training-free method with two novel modules DPSelect and PivotKV, to jointly reduce both temporal visual redundancy and knowledge redundancy for video compression. To align with the way of human temporal perception, DPSelect identifies keyframes based on inter-frame distance peaks. To leverage LLMs' learned prior knowledge, PivotKV marks the keyframes as pivots and compress non-pivot frames by pruning low-attention tokens in their KV cache. ReTaKe enables VideoLLMs to process 8 times longer frames (up to 2048), outperforming similar-sized models by 3-5% and even rivaling much larger ones on VideoMME, MLVU, LongVideoBench, and LVBench. Moreover, by overlapping compression operations with prefilling, ReTaKe introduces only ~10% prefilling latency overhead while reducing decoding latency by ~20%. Our code is available at https://github.com/SCZwangxiao/video-ReTaKe.
comment: Rewrite the methods section. Add more ablation studies and results in LongVideoBench. Update metadata
Robotics 55
☆ Aether: Geometric-Aware Unified World Modeling
The integration of geometric reconstruction and generative modeling remains a critical challenge in developing AI systems capable of human-like spatial reasoning. This paper proposes Aether, a unified framework that enables geometry-aware reasoning in world models by jointly optimizing three core capabilities: (1) 4D dynamic reconstruction, (2) action-conditioned video prediction, and (3) goal-conditioned visual planning. Through task-interleaved feature learning, Aether achieves synergistic knowledge sharing across reconstruction, prediction, and planning objectives. Building upon video generation models, our framework demonstrates unprecedented synthetic-to-real generalization despite never observing real-world data during training. Furthermore, our approach achieves zero-shot generalization in both action following and reconstruction tasks, thanks to its intrinsic geometric modeling. Remarkably, even without real-world data, its reconstruction performance far exceeds that of domain-specific models. Additionally, Aether leverages a geometry-informed action space to seamlessly translate predictions into actions, enabling effective autonomous trajectory planning. We hope our work inspires the community to explore new frontiers in physically-reasonable world modeling and its applications.
comment: Project Page: https://aether-world.github.io/
AdaWorld: Learning Adaptable World Models with Latent Actions
World models aim to learn action-controlled prediction models and have proven essential for the development of intelligent agents. However, most existing world models rely heavily on substantial action-labeled data and costly training, making it challenging to adapt to novel environments with heterogeneous actions through limited interactions. This limitation can hinder their applicability across broader domains. To overcome this challenge, we propose AdaWorld, an innovative world model learning approach that enables efficient adaptation. The key idea is to incorporate action information during the pretraining of world models. This is achieved by extracting latent actions from videos in a self-supervised manner, capturing the most critical transitions between frames. We then develop an autoregressive world model that conditions on these latent actions. This learning paradigm enables highly adaptable world models, facilitating efficient transfer and learning of new actions even with limited interactions and finetuning. Our comprehensive experiments across multiple environments demonstrate that AdaWorld achieves superior performance in both simulation quality and visual planning.
comment: Project page: https://adaptable-world-model.github.io/
☆ Autonomous Generation of Sub-goals for Lifelong Learning in Robots
One of the challenges of open-ended learning in robots is the need to autonomously discover goals and learn skills to achieve them. However, when in lifelong learning settings, it is always desirable to generate sub-goals with their associated skills, without relying on explicit reward, as steppingstones to a goal. This allows sub-goals and skills to be reused to facilitate achieving other goals. This work proposes a two-pronged approach for sub-goal generation to address this challenge: a top-down approach, where sub-goals are hierarchically derived from general goals using intrinsic motivations to discover them, and a bottom-up approach, where sub-goal chains emerge from making latent relationships between goals and perceptual classes that were previously learned in different domains explicit. These methods help the robot to autonomously generate and chain sub-goals as a way to achieve more general goals. Additionally, they create more abstract representations of goals, helping to reduce sub-goal duplication and make the learning of skills more efficient. Implemented within an existing cognitive architecture for lifelong open-ended learning and tested with a real robot, our approach enhances the robot's ability to discover and achieve goals, generate sub-goals in an efficient manner, generalize learned skills, and operate in dynamic and unknown environments without explicit intermediate rewards.
comment: 16 pages
☆ Online 3D Scene Reconstruction Using Neural Object Priors 3DV 2025
This paper addresses the problem of reconstructing a scene online at the level of objects given an RGB-D video sequence. While current object-aware neural implicit representations hold promise, they are limited in online reconstruction efficiency and shape completion. Our main contributions to alleviate the above limitations are twofold. First, we propose a feature grid interpolation mechanism to continuously update grid-based object-centric neural implicit representations as new object parts are revealed. Second, we construct an object library with previously mapped objects in advance and leverage the corresponding shape priors to initialize geometric object models in new videos, subsequently completing them with novel views as well as synthesized past views to avoid losing original object details. Extensive experiments on synthetic environments from the Replica dataset, real-world ScanNet sequences and videos captured in our laboratory demonstrate that our approach outperforms state-of-the-art neural implicit models for this task in terms of reconstruction accuracy and completeness.
comment: 3DV 2025. Project page: https://www.di.ens.fr/willow/research/online-scene-reconstruction/
☆ Bootstrapped Model Predictive Control ICLR 2025
Model Predictive Control (MPC) has been demonstrated to be effective in continuous control tasks. When a world model and a value function are available, planning a sequence of actions ahead of time leads to a better policy. Existing methods typically obtain the value function and the corresponding policy in a model-free manner. However, we find that such an approach struggles with complex tasks, resulting in poor policy learning and inaccurate value estimation. To address this problem, we leverage the strengths of MPC itself. In this work, we introduce Bootstrapped Model Predictive Control (BMPC), a novel algorithm that performs policy learning in a bootstrapped manner. BMPC learns a network policy by imitating an MPC expert, and in turn, uses this policy to guide the MPC process. Combined with model-based TD-learning, our policy learning yields better value estimation and further boosts the efficiency of MPC. We also introduce a lazy reanalyze mechanism, which enables computationally efficient imitation learning. Our method achieves superior performance over prior works on diverse continuous control tasks. In particular, on challenging high-dimensional locomotion tasks, BMPC significantly improves data efficiency while also enhancing asymptotic performance and training stability, with comparable training time and smaller network sizes. Code is available at https://github.com/wertyuilife2/bmpc.
comment: Published as a conference paper at ICLR 2025
☆ Learning Multi-Robot Coordination through Locality-Based Factorized Multi-Agent Actor-Critic Algorithm
In this work, we present a novel cooperative multi-agent reinforcement learning method called \textbf{Loc}ality based \textbf{Fac}torized \textbf{M}ulti-Agent \textbf{A}ctor-\textbf{C}ritic (Loc-FACMAC). Existing state-of-the-art algorithms, such as FACMAC, rely on global reward information, which may not accurately reflect the quality of individual robots' actions in decentralized systems. We integrate the concept of locality into critic learning, where strongly related robots form partitions during training. Robots within the same partition have a greater impact on each other, leading to more precise policy evaluation. Additionally, we construct a dependency graph to capture the relationships between robots, facilitating the partitioning process. This approach mitigates the curse of dimensionality and prevents robots from using irrelevant information. Our method improves existing algorithms by focusing on local rewards and leveraging partition-based learning to enhance training efficiency and performance. We evaluate the performance of Loc-FACMAC in three environments: Hallway, Multi-cartpole, and Bounded-Cooperative-Navigation. We explore the impact of partition sizes on the performance and compare the result with baseline MARL algorithms such as LOMAQ, FACMAC, and QMIX. The experiments reveal that, if the locality structure is defined properly, Loc-FACMAC outperforms these baseline algorithms up to 108\%, indicating that exploiting the locality structure in the actor-critic framework improves the MARL performance.
☆ AlphaSpace: Enabling Robotic Actions through Semantic Tokenization and Symbolic Reasoning
This paper presents AlphaSpace, a novel methodology designed to enhance the spatial reasoning capabilities of large language models (LLMs) for 3D Cartesian space navigation. AlphaSpace employs a semantics-based tokenization strategy, encoding height information through specialized semantic tokens, and integrates primarily symbolic synthetic reasoning data. This approach enables LLMs to accurately manipulate objects by positioning them at specific [x, y, z] coordinates. Experimental results demonstrate that AlphaSpace significantly outperforms existing models on manipulation subtasks, achieving a total accuracy of 66.67%, compared to 37.5% for GPT-4o and 29.17% for Claude 3.5 Sonnet.
☆ Robust Tube-based Control Strategy for Vision-guided Autonomous Vehicles
A robust control strategy for autonomous vehicles can improve system stability, enhance riding comfort, and prevent driving accidents. This paper presents a novel interpolation tube-based constrained iterative linear quadratic regulator (itube-CILQR) algorithm for autonomous computer-vision-based vehicle lane-keeping. The goal of the algorithm is to enhance robustness during high-speed cornering on tight turns. The advantages of itube-CILQR over the standard tube-approach include reduced system conservatism and increased computational speed. Numerical and vision-based experiments were conducted to examine the feasibility of the proposed algorithm. The proposed itube-CILQR algorithm is better suited to vehicle lane-keeping than variational CILQR-based methods and model predictive control (MPC) approaches using a classical interior-point solver. Specifically, in evaluation experiments, itube-CILQR achieved an average runtime of 3.16 ms to generate a control signal to guide a self-driving vehicle; itube-MPC typically required a 4.67-times longer computation time to complete the same task. Moreover, the influence of conservatism on system behavior was investigated by exploring the interpolation variable trajectories derived from the proposed itube-CILQR algorithm during lane-keeping maneuvers.
comment: 13 pages, 14 figures
☆ RoboEngine: Plug-and-Play Robot Data Augmentation with Semantic Robot Segmentation and Background Generation
Visual augmentation has become a crucial technique for enhancing the visual robustness of imitation learning. However, existing methods are often limited by prerequisites such as camera calibration or the need for controlled environments (e.g., green screen setups). In this work, we introduce RoboEngine, the first plug-and-play visual robot data augmentation toolkit. For the first time, users can effortlessly generate physics- and task-aware robot scenes with just a few lines of code. To achieve this, we present a novel robot scene segmentation dataset, a generalizable high-quality robot segmentation model, and a fine-tuned background generation model, which together form the core components of the out-of-the-box toolkit. Using RoboEngine, we demonstrate the ability to generalize robot manipulation tasks across six entirely new scenes, based solely on demonstrations collected from a single scene, achieving a more than 200% performance improvement compared to the no-augmentation baseline. All datasets, model weights, and the toolkit will be publicly released.
comment: Project Page: https://roboengine.github.io/
☆ Efficient Continual Adaptation of Pretrained Robotic Policy with Online Meta-Learned Adapters
Continual adaptation is essential for general autonomous agents. For example, a household robot pretrained with a repertoire of skills must still adapt to unseen tasks specific to each household. Motivated by this, building upon parameter-efficient fine-tuning in language models, prior works have explored lightweight adapters to adapt pretrained policies, which can preserve learned features from the pretraining phase and demonstrate good adaptation performances. However, these approaches treat task learning separately, limiting knowledge transfer between tasks. In this paper, we propose Online Meta-Learned adapters (OMLA). Instead of applying adapters directly, OMLA can facilitate knowledge transfer from previously learned tasks to current learning tasks through a novel meta-learning objective. Extensive experiments in both simulated and real-world environments demonstrate that OMLA can lead to better adaptation performances compared to the baseline methods. The project link: https://ricky-zhu.github.io/OMLA/.
comment: Project link: https://ricky-zhu.github.io/OMLA/
☆ Any6D: Model-free 6D Pose Estimation of Novel Objects CVPR 2025
We introduce Any6D, a model-free framework for 6D object pose estimation that requires only a single RGB-D anchor image to estimate both the 6D pose and size of unknown objects in novel scenes. Unlike existing methods that rely on textured 3D models or multiple viewpoints, Any6D leverages a joint object alignment process to enhance 2D-3D alignment and metric scale estimation for improved pose accuracy. Our approach integrates a render-and-compare strategy to generate and refine pose hypotheses, enabling robust performance in scenarios with occlusions, non-overlapping views, diverse lighting conditions, and large cross-environment variations. We evaluate our method on five challenging datasets: REAL275, Toyota-Light, HO3D, YCBINEOAT, and LM-O, demonstrating its effectiveness in significantly outperforming state-of-the-art methods for novel object pose estimation. Project page: https://taeyeop.com/any6d
comment: CVPR 2025, Project Page: https://taeyeop.com/any6d
☆ FF-SRL: High Performance GPU-Based Surgical Simulation For Robot Learning
Robotic surgery is a rapidly developing field that can greatly benefit from the automation of surgical tasks. However, training techniques such as Reinforcement Learning (RL) require a high number of task repetitions, which are generally unsafe and impractical to perform on real surgical systems. This stresses the need for simulated surgical environments, which are not only realistic, but also computationally efficient and scalable. We introduce FF-SRL (Fast and Flexible Surgical Reinforcement Learning), a high-performance learning environment for robotic surgery. In FF-SRL both physics simulation and RL policy training reside entirely on a single GPU. This avoids typical bottlenecks associated with data transfer between the CPU and GPU, leading to accelerated learning rates. Our results show that FF-SRL reduces the training time of a complex tissue manipulation task by an order of magnitude, down to a couple of minutes, compared to a common CPU/GPU simulator. Such speed-up may facilitate the experimentation with RL techniques and contribute to the development of new generation of surgical systems. To this end, we make our code publicly available to the community.
☆ Multi-agent coordination for data gathering with periodic requests and deliveries
In this demo work we develop a method to plan and coordinate a multi-agent team to gather information on demand. The data is periodically requested by a static Operation Center (OC) from changeable goals locations. The mission of the team is to reach these locations, taking measurements and delivering the data to the OC. Due to the limited communication range as well as signal attenuation because of the obstacles, the agents must travel to the OC, to upload the data. The agents can play two roles: ones as workers gathering data, the others as collectors traveling invariant paths for collecting the data of the workers to re-transmit it to the OC. The refreshing time of the delivered information depends on the number of available agents as well as of the scenario. The proposed algorithm finds out the best balance between the number of collectors-workers and the partition of the scenario into working areas in the planning phase, which provides the minimum refreshing time and will be the one executed by the agents.
☆ Communication-aware planning for robot teams deployment
In the present work we address the problem of deploying a team of robots in a scenario where some locations of interest must be reached. Thus, a planning for a deployment is required, before sending the robots. The obstacles, the limited communication range, and the need of communicating to a base station, constrain the connectivity of the team and the deployment planning. We propose a method consisting of three algorithms: a distributed path planner to obtain communication-aware trajectories; a deployment planner providing dual-use of the robots, visiting primary goals and performing connectivity tasks; and a clustering algorithm to allocate the tasks to robots, and obtain the best goal visit order for the mission.
☆ Parental Guidance: Efficient Lifelong Learning through Evolutionary Distillation
Developing robotic agents that can perform well in diverse environments while showing a variety of behaviors is a key challenge in AI and robotics. Traditional reinforcement learning (RL) methods often create agents that specialize in narrow tasks, limiting their adaptability and diversity. To overcome this, we propose a preliminary, evolution-inspired framework that includes a reproduction module, similar to natural species reproduction, balancing diversity and specialization. By integrating RL, imitation learning (IL), and a coevolutionary agent-terrain curriculum, our system evolves agents continuously through complex tasks. This approach promotes adaptability, inheritance of useful traits, and continual learning. Agents not only refine inherited skills but also surpass their predecessors. Our initial experiments show that this method improves exploration efficiency and supports open-ended learning, offering a scalable solution where sparse reward coupled with diverse terrain environments induces a multi-task setting.
comment: 4 pages, 3 figures, CoRL 2024 Workshop MAPoDeL
☆ P3Nav: A Unified Framework for Embodied Navigation Integrating Perception, Planning, and Prediction
In language-guided visual navigation, agents locate target objects in unseen environments using natural language instructions. For reliable navigation in unfamiliar scenes, agents must possess strong perception, planning, and prediction capabilities. Additionally, when agents revisit previously explored areas during long-term navigation, they may retain irrelevant and redundant historical perceptions, leading to suboptimal results. In this work, we introduce \textbf{P3Nav}, a unified framework that integrates \textbf{P}erception, \textbf{P}lanning, and \textbf{P}rediction capabilities through \textbf{Multitask Collaboration} on navigation and embodied question answering (EQA) tasks, thereby enhancing navigation performance. Furthermore, P3Nav employs an \textbf{Adaptive 3D-aware History Sampling} strategy to effectively and efficiently utilize historical observations. By leveraging the large language models (LLM), P3Nav comprehends diverse commands and complex visual scenes, resulting in appropriate navigation actions. P3Nav achieves a 75\% success rate in object goal navigation on the $\mathrm{CHORES}$-$\mathbb{S}$ benchmark, setting a new state-of-the-art performance.
comment: 14 pages, 7 figures
☆ Analysis of Forces Exerted by Shoulder and Elbow Fabric-based Pneumatic Actuators for Pediatric Exosuits
To enhance pediatric exosuit design, it is crucial to assess the actuator-generated forces. This work evaluates the contact forces exerted by soft fabric-based pneumatic actuators in an upper extremity pediatric exosuit. Two actuators were examined: a single-cell bidirectional actuator for shoulder abduction/adduction and a bellow-type actuator for elbow extension/flexion. Experiments assessed the impact of actuator anchoring points and the adjacent joint's angle on exerted forces and actuated joint range of motion (ROM). These were measured via load cells and encoders integrated into a custom infant-scale engineered apparatus with two degrees of freedom (two revolute joints). For the shoulder actuator, results show that anchoring it further from the shoulder joint center while the elbow is flexed at $90^\circ$ yields the highest ROM while minimizing the peak force exerted on the body. For the elbow actuator, anchoring it symmetrically while the shoulder joint is at $0^\circ$ optimizes actuator performance. These findings contribute a key step toward co-optimizing the considered exosuit design for functionality and wearability.
☆ Innovative Automated Stretch Elastic Waistband Sewing Machine for Garment Manufacturing
There is applied research for the development of the Automated Stretch Elastic Waistband Sewing Machine represents a significant advancement in garment manufacturing, addressing the industry's need for increased efficiency, precision, and adaptability. This machine integrates innovative features such as a sensor-based automatic waistband expansion system, synchronized sewing speed and rolling wheel speed, and a differential feed top-loading mechanism. These enhancements streamline the sewing process, reduce manual intervention, and ensure consistent product quality. The machine's design incorporates both 3-wheel and 2-wheel rolling systems, each optimized for different elastic band dimensions and elongation factors. The 3-wheel rolling system accommodates a larger maximum boundary, while the 2-wheel rolling system offers a tighter operational range, providing flexibility to meet diverse manufacturing requirements. The Automated Stretch Elastic Waistband Sewing Machine has a design that controls the pulling apart force so as not to break the elastic waistband. It sets a new standard for quality and innovation, empowering manufacturers to meet the demands of a competitive market with precision and ease.
comment: 13 pages, 10 Figures
☆ Reinforcement Learning for Adaptive Planner Parameter Tuning: A Perspective on Hierarchical Architecture
Automatic parameter tuning methods for planning algorithms, which integrate pipeline approaches with learning-based techniques, are regarded as promising due to their stability and capability to handle highly constrained environments. While existing parameter tuning methods have demonstrated considerable success, further performance improvements require a more structured approach. In this paper, we propose a hierarchical architecture for reinforcement learning-based parameter tuning. The architecture introduces a hierarchical structure with low-frequency parameter tuning, mid-frequency planning, and high-frequency control, enabling concurrent enhancement of both upper-layer parameter tuning and lower-layer control through iterative training. Experimental evaluations in both simulated and real-world environments show that our method surpasses existing parameter tuning approaches. Furthermore, our approach achieves first place in the Benchmark for Autonomous Robot Navigation (BARN) Challenge.
☆ Latent Embedding Adaptation for Human Preference Alignment in Diffusion Planners
This work addresses the challenge of personalizing trajectories generated in automated decision-making systems by introducing a resource-efficient approach that enables rapid adaptation to individual users' preferences. Our method leverages a pretrained conditional diffusion model with Preference Latent Embeddings (PLE), trained on a large, reward-free offline dataset. The PLE serves as a compact representation for capturing specific user preferences. By adapting the pretrained model using our proposed preference inversion method, which directly optimizes the learnable PLE, we achieve superior alignment with human preferences compared to existing solutions like Reinforcement Learning from Human Feedback (RLHF) and Low-Rank Adaptation (LoRA). To better reflect practical applications, we create a benchmark experiment using real human preferences on diverse, high-reward trajectories.
comment: 8 pages
☆ Vision-Guided Loco-Manipulation with a Snake Robot
This paper presents the development and integration of a vision-guided loco-manipulation pipeline for Northeastern University's snake robot, COBRA. The system leverages a YOLOv8-based object detection model and depth data from an onboard stereo camera to estimate the 6-DOF pose of target objects in real time. We introduce a framework for autonomous detection and control, enabling closed-loop loco-manipulation for transporting objects to specified goal locations. Additionally, we demonstrate open-loop experiments in which COBRA successfully performs real-time object detection and loco-manipulation tasks.
☆ NMPC-based Unified Posture Manipulation and Thrust Vectoring for Fault Recovery
Multi-rotors face significant risks, as actuator failures at high altitudes can easily result in a crash and the robot's destruction. Therefore, rapid fault recovery in the event of an actuator failure is necessary for the fault-tolerant and safe operation of unmanned aerial robots. In this work, we present a fault recovery approach based on the unification of posture manipulation and thrust vectoring. The key contributions of this work are: 1) Derivation of two flight dynamics models (high-fidelity and reduced-order) that capture posture control and thrust vectoring. 2) Design of a controller based on Nonlinear Model Predictive Control (NMPC) and demonstration of fault recovery in simulation using a high-fidelity model of the Multi-Modal Mobility Morphobot (M4) in Simscape.
☆ Ground Penetrating Radar-Assisted Multimodal Robot Odometry Using Subsurface Feature Matrix
Localization of robots using subsurface features observed by ground-penetrating radar (GPR) enhances and adds robustness to common sensor modalities, as subsurface features are less affected by weather, seasons, and surface changes. We introduce an innovative multimodal odometry approach using inputs from GPR, an inertial measurement unit (IMU), and a wheel encoder. To efficiently address GPR signal noise, we introduce an advanced feature representation called the subsurface feature matrix (SFM). The SFM leverages frequency domain data and identifies peaks within radar scans. Additionally, we propose a novel feature matching method that estimates GPR displacement by aligning SFMs. The integrations from these three input sources are consolidated using a factor graph approach to achieve multimodal robot odometry. Our method has been developed and evaluated with the CMU-GPR public dataset, demonstrating improvements in accuracy and robustness with real-time performance in robotic odometry tasks.
☆ Learning Orientation Field for OSM-Guided Autonomous Navigation
OpenStreetMap (OSM) has gained popularity recently in autonomous navigation due to its public accessibility, lower maintenance costs, and broader geographical coverage. However, existing methods often struggle with noisy OSM data and incomplete sensor observations, leading to inaccuracies in trajectory planning. These challenges are particularly evident in complex driving scenarios, such as at intersections or facing occlusions. To address these challenges, we propose a robust and explainable two-stage framework to learn an Orientation Field (OrField) for robot navigation by integrating LiDAR scans and OSM routes. In the first stage, we introduce the novel representation, OrField, which can provide orientations for each grid on the map, reasoning jointly from noisy LiDAR scans and OSM routes. To generate a robust OrField, we train a deep neural network by encoding a versatile initial OrField and output an optimized OrField. Based on OrField, we propose two trajectory planners for OSM-guided robot navigation, called Field-RRT* and Field-Bezier, respectively, in the second stage by improving the Rapidly Exploring Random Tree (RRT) algorithm and Bezier curve to estimate the trajectories. Thanks to the robustness of OrField which captures both global and local information, Field-RRT* and Field-Bezier can generate accurate and reliable trajectories even in challenging conditions. We validate our approach through experiments on the SemanticKITTI dataset and our own campus dataset. The results demonstrate the effectiveness of our method, achieving superior performance in complex and noisy conditions. Our code for network training and real-world deployment is available at https://github.com/IMRL/OriField.
comment: 14 pages, 12 figures, and 5 tables
☆ GI-SLAM: Gaussian-Inertial SLAM
3D Gaussian Splatting (3DGS) has recently emerged as a powerful representation of geometry and appearance for dense Simultaneous Localization and Mapping (SLAM). Through rapid, differentiable rasterization of 3D Gaussians, many 3DGS SLAM methods achieve near real-time rendering and accelerated training. However, these methods largely overlook inertial data, witch is a critical piece of information collected from the inertial measurement unit (IMU). In this paper, we present GI-SLAM, a novel gaussian-inertial SLAM system which consists of an IMU-enhanced camera tracking module and a realistic 3D Gaussian-based scene representation for mapping. Our method introduces an IMU loss that seamlessly integrates into the deep learning framework underpinning 3D Gaussian Splatting SLAM, effectively enhancing the accuracy, robustness and efficiency of camera tracking. Moreover, our SLAM system supports a wide range of sensor configurations, including monocular, stereo, and RGBD cameras, both with and without IMU integration. Our method achieves competitive performance compared with existing state-of-the-art real-time methods on the EuRoC and TUM-RGBD datasets.
comment: 10 pages, 2 figures, 5 tables
☆ Optimal Modified Feedback Strategies in LQ Games under Control Imperfections
Game-theoretic approaches and Nash equilibrium have been widely applied across various engineering domains. However, practical challenges such as disturbances, delays, and actuator limitations can hinder the precise execution of Nash equilibrium strategies. This work explores the impact of such implementation imperfections on game trajectories and players' costs within the context of a two-player linear quadratic (LQ) nonzero-sum game. Specifically, we analyze how small deviations by one player affect the state and cost function of the other player. To address these deviations, we propose an adjusted control policy that not only mitigates adverse effects optimally but can also exploit the deviations to enhance performance. Rigorous mathematical analysis and proofs are presented, demonstrating through a representative example that the proposed policy modification achieves up to $61\%$ improvement compared to the unadjusted feedback policy and up to $0.59\%$ compared to the feedback Nash strategy.
comment: 6 pages, 2 figures, Preprint version of a paper submitted to L-CSS and CDC
☆ Contact-based Grasp Control and Inverse Kinematics for a Five-fingered Robotic Hand
This paper presents an implementation and analysis of a five-fingered robotic grasping system that combines contact-based control with inverse kinematics solutions. Using the PyBullet simulation environment and the DexHand v2 model, we demonstrate a comprehensive approach to achieving stable grasps through contact point optimization with force closure validation. Our method achieves movement efficiency ratings between 0.966-0.996 for non-thumb fingers and 0.879 for the thumb, while maintaining positional accuracy within 0.0267-0.0283m for non-thumb digits and 0.0519m for the thumb. The system demonstrates rapid position stabilization at 240Hz simulation frequency and maintains stable contact configurations throughout the grasp execution. Experimental results validate the effectiveness of our approach, while also identifying areas for future enhancement in thumb opposition movements and horizontal plane control.
comment: 10 Pages, 5 Figures, 1 Table
☆ Dom, cars don't fly! -- Or do they? In-Air Vehicle Maneuver for High-Speed Off-Road Navigation
When pushing the speed limit for aggressive off-road navigation on uneven terrain, it is inevitable that vehicles may become airborne from time to time. During time-sensitive tasks, being able to fly over challenging terrain can also save time, instead of cautiously circumventing or slowly negotiating through. However, most off-road autonomy systems operate under the assumption that the vehicles are always on the ground and therefore limit operational speed. In this paper, we present a novel approach for in-air vehicle maneuver during high-speed off-road navigation. Based on a hybrid forward kinodynamic model using both physics principles and machine learning, our fixed-horizon, sampling-based motion planner ensures accurate vehicle landing poses and their derivatives within a short airborne time window using vehicle throttle and steering commands. We test our approach in extensive in-air experiments both indoors and outdoors, compare it against an error-driven control method, and demonstrate that precise and timely in-air vehicle maneuver is possible through existing ground vehicle controls.
comment: 8 Pages, 4 Figures
☆ Cooperative Control of Multi-Quadrotors for Transporting Cable-Suspended Payloads: Obstacle-Aware Planning and Event-Based Nonlinear Model Predictive Control
This paper introduces a novel methodology for the cooperative control of multiple quadrotors transporting cablesuspended payloads, emphasizing obstacle-aware planning and event-based Nonlinear Model Predictive Control (NMPC). Our approach integrates trajectory planning with real-time control through a combination of the A* algorithm for global path planning and NMPC for local control, enhancing trajectory adaptability and obstacle avoidance. We propose an advanced event-triggered control system that updates based on events identified through dynamically generated environmental maps. These maps are constructed using a dual-camera setup, which includes multi-camera systems for static obstacle detection and event cameras for high-resolution, low-latency detection of dynamic obstacles. This design is crucial for addressing fast-moving and transient obstacles that conventional cameras may overlook, particularly in environments with rapid motion and variable lighting conditions. When new obstacles are detected, the A* algorithm recalculates waypoints based on the updated map, ensuring safe and efficient navigation. This real-time obstacle detection and map updating integration allows the system to adaptively respond to environmental changes, markedly improving safety and navigation efficiency. The system employs SLAM and object detection techniques utilizing data from multi-cameras, event cameras, and IMUs for accurate localization and comprehensive environmental mapping. The NMPC framework adeptly manages the complex dynamics of multiple quadrotors and suspended payloads, incorporating safety constraints to maintain dynamic feasibility and stability. Extensive simulations validate the proposed approach, demonstrating significant enhancements in energy efficiency, computational resource management, and responsiveness.
☆ Evolutionary Policy Optimization
Despite its extreme sample inefficiency, on-policy reinforcement learning has become a fundamental tool in real-world applications. With recent advances in GPU-driven simulation, the ability to collect vast amounts of data for RL training has scaled exponentially. However, studies show that current on-policy methods, such as PPO, fail to fully leverage the benefits of parallelized environments, leading to performance saturation beyond a certain scale. In contrast, Evolutionary Algorithms (EAs) excel at increasing diversity through randomization, making them a natural complement to RL. However, existing EvoRL methods have struggled to gain widespread adoption due to their extreme sample inefficiency. To address these challenges, we introduce Evolutionary Policy Optimization (EPO), a novel policy gradient algorithm that combines the strengths of EA and policy gradients. We show that EPO significantly improves performance across diverse and challenging environments, demonstrating superior scalability with parallelized simulations.
comment: Website at https://sites.google.com/view/epo-rl
♻ ☆ Double Oracle Algorithm for Game-Theoretic Robot Allocation on Graphs
We study the problem of game-theoretic robot allocation where two players strategically allocate robots to compete for multiple sites of interest. Robots possess offensive or defensive capabilities to interfere and weaken their opponents to take over a competing site. This problem belongs to the conventional Colonel Blotto Game. Considering the robots' heterogeneous capabilities and environmental factors, we generalize the conventional Blotto game by incorporating heterogeneous robot types and graph constraints that capture the robot transitions between sites. Then we employ the Double Oracle Algorithm (DOA) to solve for the Nash equilibrium of the generalized Blotto game. Particularly, for cyclic-dominance-heterogeneous (CDH) robots that inhibit each other, we define a new transformation rule between any two robot types. Building on the transformation, we design a novel utility function to measure the game's outcome quantitatively. Moreover, we rigorously prove the correctness of the designed utility function. Finally, we conduct extensive simulations to demonstrate the effectiveness of DOA on computing Nash equilibrium for homogeneous, linear heterogeneous, and CDH robot allocation on graphs.
♻ ☆ Physics-Informed Multi-Agent Reinforcement Learning for Distributed Multi-Robot Problems
The networked nature of multi-robot systems presents challenges in the context of multi-agent reinforcement learning. Centralized control policies do not scale with increasing numbers of robots, whereas independent control policies do not exploit the information provided by other robots, exhibiting poor performance in cooperative-competitive tasks. In this work we propose a physics-informed reinforcement learning approach able to learn distributed multi-robot control policies that are both scalable and make use of all the available information to each robot. Our approach has three key characteristics. First, it imposes a port-Hamiltonian structure on the policy representation, respecting energy conservation properties of physical robot systems and the networked nature of robot team interactions. Second, it uses self-attention to ensure a sparse policy representation able to handle time-varying information at each robot from the interaction graph. Third, we present a soft actor-critic reinforcement learning algorithm parameterized by our self-attention port-Hamiltonian control policy, which accounts for the correlation among robots during training while overcoming the need of value function factorization. Extensive simulations in different multi-robot scenarios demonstrate the success of the proposed approach, surpassing previous multi-robot reinforcement learning solutions in scalability, while achieving similar or superior performance (with averaged cumulative reward up to x2 greater than the state-of-the-art with robot teams x6 larger than the number of robots at training time). We also validate our approach on multiple real robots in the Georgia Tech Robotarium under imperfect communication, demonstrating zero-shot sim-to-real transfer and scalability across number of robots.
comment: This paper is under review at IEEE T-RO
♻ ☆ Kalib: Easy Hand-Eye Calibration with Reference Point Tracking
Hand-eye calibration aims to estimate the transformation between a camera and a robot. Traditional methods rely on fiducial markers, which require considerable manual effort and precise setup. Recent advances in deep learning have introduced markerless techniques but come with more prerequisites, such as retraining networks for each robot, and accessing accurate mesh models for data generation. In this paper, we propose Kalib, an automatic and easy-to-setup hand-eye calibration method that leverages the generalizability of visual foundation models to overcome these challenges. It features only two basic prerequisites, the robot's kinematic chain and a predefined reference point on the robot. During calibration, the reference point is tracked in the camera space. Its corresponding 3D coordinates in the robot coordinate can be inferred by forward kinematics. Then, a PnP solver directly estimates the transformation between the camera and the robot without training new networks or accessing mesh models. Evaluations in simulated and real-world benchmarks show that Kalib achieves good accuracy with a lower manual workload compared with recent baseline methods. We also demonstrate its application in multiple real-world settings with various robot arms and grippers. Kalib's user-friendly design and minimal setup requirements make it a possible solution for continuous operation in unstructured environments.
comment: The code, data, and supplementary materials are available at https://sites.google.com/view/hand-eye-kalib
♻ ☆ ETAP: Event-based Tracking of Any Point
Tracking any point (TAP) recently shifted the motion estimation paradigm from focusing on individual salient points with local templates to tracking arbitrary points with global image contexts. However, while research has mostly focused on driving the accuracy of models in nominal settings, addressing scenarios with difficult lighting conditions and high-speed motions remains out of reach due to the limitations of the sensor. This work addresses this challenge with the first event camera-based TAP method. It leverages the high temporal resolution and high dynamic range of event cameras for robust high-speed tracking, and the global contexts in TAP methods to handle asynchronous and sparse event measurements. We further extend the TAP framework to handle event feature variations induced by motion -- thereby addressing an open challenge in purely event-based tracking -- with a novel feature-alignment loss which ensures the learning of motion-robust features. Our method is trained with data from a new data generation pipeline and systematically ablated across all design decisions. Our method shows strong cross-dataset generalization and performs 136% better on the average Jaccard metric than the baselines. Moreover, on an established feature tracking benchmark, it achieves a 20% improvement over the previous best event-only method and even surpasses the previous best events-and-frames method by 4.1%. Our code is available at https://github.com/tub-rip/ETAP
comment: 17 pages, 15 figures, 8 tables. Project page: https://github.com/tub-rip/ETAP
♻ ☆ CarPlanner: Consistent Auto-regressive Trajectory Planning for Large-scale Reinforcement Learning in Autonomous Driving CVPR 2025
Trajectory planning is vital for autonomous driving, ensuring safe and efficient navigation in complex environments. While recent learning-based methods, particularly reinforcement learning (RL), have shown promise in specific scenarios, RL planners struggle with training inefficiencies and managing large-scale, real-world driving scenarios. In this paper, we introduce \textbf{CarPlanner}, a \textbf{C}onsistent \textbf{a}uto-\textbf{r}egressive \textbf{Planner} that uses RL to generate multi-modal trajectories. The auto-regressive structure enables efficient large-scale RL training, while the incorporation of consistency ensures stable policy learning by maintaining coherent temporal consistency across time steps. Moreover, CarPlanner employs a generation-selection framework with an expert-guided reward function and an invariant-view module, simplifying RL training and enhancing policy performance. Extensive analysis demonstrates that our proposed RL framework effectively addresses the challenges of training efficiency and performance enhancement, positioning CarPlanner as a promising solution for trajectory planning in autonomous driving. To the best of our knowledge, we are the first to demonstrate that the RL-based planner can surpass both IL- and rule-based state-of-the-arts (SOTAs) on the challenging large-scale real-world dataset nuPlan. Our proposed CarPlanner surpasses RL-, IL-, and rule-based SOTA approaches within this demanding dataset.
comment: CVPR 2025
♻ ☆ Believing is Seeing: Unobserved Object Detection using Generative Models
Can objects that are not visible in an image -- but are in the vicinity of the camera -- be detected? This study introduces the novel tasks of 2D, 2.5D and 3D unobserved object detection for predicting the location of nearby objects that are occluded or lie outside the image frame. We adapt several state-of-the-art pre-trained generative models to address this task, including 2D and 3D diffusion models and vision-language models, and show that they can be used to infer the presence of objects that are not directly observed. To benchmark this task, we propose a suite of metrics that capture different aspects of performance. Our empirical evaluation on indoor scenes from the RealEstate10k and NYU Depth v2 datasets demonstrate results that motivate the use of generative models for the unobserved object detection task.
comment: IEEE/CVF Computer Vision and Pattern Recognition 2025; 22 pages
♻ ☆ Evolution 6.0: Evolving Robotic Capabilities Through Generative Design IROS
We propose a new concept, Evolution 6.0, which represents the evolution of robotics driven by Generative AI. When a robot lacks the necessary tools to accomplish a task requested by a human, it autonomously designs the required instruments and learns how to use them to achieve the goal. Evolution 6.0 is an autonomous robotic system powered by Vision-Language Models (VLMs), Vision-Language Action (VLA) models, and Text-to-3D generative models for tool design and task execution. The system comprises two key modules: the Tool Generation Module, which fabricates task-specific tools from visual and textual data, and the Action Generation Module, which converts natural language instructions into robotic actions. It integrates QwenVLM for environmental understanding, OpenVLA for task execution, and Llama-Mesh for 3D tool generation. Evaluation results demonstrate a 90% success rate for tool generation with a 10-second inference time, and action generation achieving 83.5% in physical and visual generalization, 70% in motion generalization, and 37% in semantic generalization. Future improvements will focus on bimanual manipulation, expanded task capabilities, and enhanced environmental interpretation to improve real-world adaptability.
comment: Submitted to IROS
♻ ☆ Improving robot navigation in crowded environments using intrinsic rewards ICRA
Autonomous navigation in crowded environments is an open problem with many applications, essential for the coexistence of robots and humans in the smart cities of the future. In recent years, deep reinforcement learning approaches have proven to outperform model-based algorithms. Nevertheless, even though the results provided are promising, the works are not able to take advantage of the capabilities that their models offer. They usually get trapped in local optima in the training process, that prevent them from learning the optimal policy. They are not able to visit and interact with every possible state appropriately, such as with the states near the goal or near the dynamic obstacles. In this work, we propose using intrinsic rewards to balance between exploration and exploitation and explore depending on the uncertainty of the states instead of on the time the agent has been trained, encouraging the agent to get more curious about unknown states. We explain the benefits of the approach and compare it with other exploration algorithms that may be used for crowd navigation. Many simulation experiments are performed modifying several algorithms of the state-of-the-art, showing that the use of intrinsic rewards makes the robot learn faster and reach higher rewards and success rates (fewer collisions) in shorter navigation times, outperforming the state-of-the-art.
comment: Paper accepted in 2023 IEEE International Conference on Robotics and Automation (ICRA)
♻ ☆ AVOCADO: Adaptive Optimal Collision Avoidance driven by Opinion
We present AVOCADO (AdaptiVe Optimal Collision Avoidance Driven by Opinion), a novel navigation approach to address holonomic robot collision avoidance when the robot does not know how cooperative the other agents in the environment are. AVOCADO departs from a Velocity Obstacle's (VO) formulation akin to the Optimal Reciprocal Collision Avoidance method. However, instead of assuming reciprocity, it poses an adaptive control problem to adapt to the cooperation level of other robots and agents in real time. This is achieved through a novel nonlinear opinion dynamics design that relies solely on sensor observations. As a by-product, we leverage tools from the opinion dynamics formulation to naturally avoid the deadlocks in geometrically symmetric scenarios that typically suffer VO-based planners. Extensive numerical simulations show that AVOCADO surpasses existing motion planners in mixed cooperative/non-cooperative navigation environments in terms of success rate, time to goal and computational time. In addition, we conduct multiple real experiments that verify that AVOCADO is able to avoid collisions in environments crowded with other robots and humans.
comment: This paper is published at IEEE Transactions on Robotics under DOI 10.1109/TRO.2025.3552350
♻ ☆ Humanoid Policy ~ Human Policy
Training manipulation policies for humanoid robots with diverse data enhances their robustness and generalization across tasks and platforms. However, learning solely from robot demonstrations is labor-intensive, requiring expensive tele-operated data collection which is difficult to scale. This paper investigates a more scalable data source, egocentric human demonstrations, to serve as cross-embodiment training data for robot learning. We mitigate the embodiment gap between humanoids and humans from both the data and modeling perspectives. We collect an egocentric task-oriented dataset (PH2D) that is directly aligned with humanoid manipulation demonstrations. We then train a human-humanoid behavior policy, which we term Human Action Transformer (HAT). The state-action space of HAT is unified for both humans and humanoid robots and can be differentiably retargeted to robot actions. Co-trained with smaller-scale robot data, HAT directly models humanoid robots and humans as different embodiments without additional supervision. We show that human data improves both generalization and robustness of HAT with significantly better data collection efficiency. Code and data: https://human-as-robot.github.io/
comment: Code and data: https://human-as-robot.github.io/
♻ ☆ Emergency-Brake Simplex: Toward A Verifiably Safe Control-CPS Architecture for Abrupt Runtime Reachability Constraint Changes
When a system's constraints change abruptly, the system's reachability safety does no longer sustain. Thus, the system can reach a forbidden/dangerous value. Conventional remedy practically involves online controller redesign (OCR) to re-establish the reachability's compliance with the new constraints, which, however, is usually too slow. There is a need for an online strategy capable of managing runtime changes in reachability constraints. However, to the best of the authors' knowledge, this topic has not been addressed in the existing literature. In this paper, we propose a fast fault tolerance strategy to recover the system's reachability safety in runtime. Instead of redesigning the system's controller, we propose to change the system's reference state to modify the system's reachability to comply with the new constraints. We frame the reference state search as an optimization problem and employ the Karush-Kuhn-Tucker (KKT) method as well as the Interior Point Method (IPM) based Newton's method (as a fallback for the KKT method) for fast solution derivation. The optimization also allows more future fault tolerance. Numerical simulations demonstrate that our method outperforms the conventional OCR method in terms of computational efficiency and success rate. Specifically, the results show that the proposed method finds a solution $10^{2}$ (with the IPM based Newton's method) $\sim 10^{4}$ (with the KKT method) times faster than the OCR method. Additionally, the improvement rate of the success rate of our method over the OCR method is $40.81\%$ without considering the deadline of run time. The success rate remains at $49.44\%$ for the proposed method, while it becomes $0\%$ for the OCR method when a deadline of $1.5 \; seconds$ is imposed.
comment: 12 pages, 2 figures,
♻ ☆ Uni-Gaussians: Unifying Camera and Lidar Simulation with Gaussians for Dynamic Driving Scenarios
Ensuring the safety of autonomous vehicles necessitates comprehensive simulation of multi-sensor data, encompassing inputs from both cameras and LiDAR sensors, across various dynamic driving scenarios. Neural rendering techniques, which utilize collected raw sensor data to simulate these dynamic environments, have emerged as a leading methodology. While NeRF-based approaches can uniformly represent scenes for rendering data from both camera and LiDAR, they are hindered by slow rendering speeds due to dense sampling. Conversely, Gaussian Splatting-based methods employ Gaussian primitives for scene representation and achieve rapid rendering through rasterization. However, these rasterization-based techniques struggle to accurately model non-linear optical sensors. This limitation restricts their applicability to sensors beyond pinhole cameras. To address these challenges and enable unified representation of dynamic driving scenarios using Gaussian primitives, this study proposes a novel hybrid approach. Our method utilizes rasterization for rendering image data while employing Gaussian ray-tracing for LiDAR data rendering. Experimental results on public datasets demonstrate that our approach outperforms current state-of-the-art methods. This work presents a unified and efficient solution for realistic simulation of camera and LiDAR data in autonomous driving scenarios using Gaussian primitives, offering significant advancements in both rendering quality and computational efficiency.
comment: 10 pages
♻ ☆ LaMOuR: Leveraging Language Models for Out-of-Distribution Recovery in Reinforcement Learning
Deep Reinforcement Learning (DRL) has demonstrated strong performance in robotic control but remains susceptible to out-of-distribution (OOD) states, often resulting in unreliable actions and task failure. While previous methods have focused on minimizing or preventing OOD occurrences, they largely neglect recovery once an agent encounters such states. Although the latest research has attempted to address this by guiding agents back to in-distribution states, their reliance on uncertainty estimation hinders scalability in complex environments. To overcome this limitation, we introduce Language Models for Out-of-Distribution Recovery (LaMOuR), which enables recovery learning without relying on uncertainty estimation. LaMOuR generates dense reward codes that guide the agent back to a state where it can successfully perform its original task, leveraging the capabilities of LVLMs in image description, logical reasoning, and code generation. Experimental results show that LaMOuR substantially enhances recovery efficiency across diverse locomotion tasks and even generalizes effectively to complex environments, including humanoid locomotion and mobile manipulation, where existing methods struggle. The code and supplementary materials are available at https://lamour-rl.github.io/.
comment: 14 pages, 17 figures
♻ ☆ Diff-DAgger: Uncertainty Estimation with Diffusion Policy for Robotic Manipulation ICRA
Recently, diffusion policy has shown impressive results in handling multi-modal tasks in robotic manipulation. However, it has fundamental limitations in out-of-distribution failures that persist due to compounding errors and its limited capability to extrapolate. One way to address these limitations is robot-gated DAgger, an interactive imitation learning with a robot query system to actively seek expert help during policy rollout. While robot-gated DAgger has high potential for learning at scale, existing methods like Ensemble-DAgger struggle with highly expressive policies: They often misinterpret policy disagreements as uncertainty at multi-modal decision points. To address this problem, we introduce Diff-DAgger, an efficient robot-gated DAgger algorithm that leverages the training objective of diffusion policy. We evaluate Diff-DAgger across different robot tasks including stacking, pushing, and plugging, and show that Diff-DAgger improves the task failure prediction by 39.0%, the task completion rate by 20.6%, and reduces the wall-clock time by a factor of 7.8. We hope that this work opens up a path for efficiently incorporating expressive yet data-hungry policies into interactive robot learning settings. The project website is available at: https://diffdagger.github.io.
comment: Project website: diffdagger.github.io 8 pages, 6 figures, accepted by International Conference on Robotics and Automation (ICRA) 2025
♻ ☆ Control Strategies for Pursuit-Evasion Under Occlusion Using Visibility and Safety Barrier Functions
This paper develops a control strategy for pursuit-evasion problems in environments with occlusions. We address the challenge of a mobile pursuer keeping a mobile evader within its field of view (FoV) despite line-of-sight obstructions. The signed distance function (SDF) of the FoV is used to formulate visibility as a control barrier function (CBF) constraint on the pursuer's control inputs. Similarly, obstacle avoidance is formulated as a CBF constraint based on the SDF of the obstacle set. While the visibility and safety CBFs are Lipschitz continuous, they are not differentiable everywhere, necessitating the use of generalized gradients. To achieve non-myopic pursuit, we generate reference control trajectories leading to evader visibility using a sampling-based kinodynamic planner. The pursuer then tracks this reference via convex optimization under the CBF constraints. We validate our approach in CARLA simulations and real-world robot experiments, demonstrating successful visibility maintenance using only onboard sensing, even under severe occlusions and dynamic evader movements.
comment: 7 pages, 7 figures
♻ ☆ DiffusionDrive: Truncated Diffusion Model for End-to-End Autonomous Driving CVPR 2025
Recently, the diffusion model has emerged as a powerful generative technique for robotic policy learning, capable of modeling multi-mode action distributions. Leveraging its capability for end-to-end autonomous driving is a promising direction. However, the numerous denoising steps in the robotic diffusion policy and the more dynamic, open-world nature of traffic scenes pose substantial challenges for generating diverse driving actions at a real-time speed. To address these challenges, we propose a novel truncated diffusion policy that incorporates prior multi-mode anchors and truncates the diffusion schedule, enabling the model to learn denoising from anchored Gaussian distribution to the multi-mode driving action distribution. Additionally, we design an efficient cascade diffusion decoder for enhanced interaction with conditional scene context. The proposed model, DiffusionDrive, demonstrates 10$\times$ reduction in denoising steps compared to vanilla diffusion policy, delivering superior diversity and quality in just 2 steps. On the planning-oriented NAVSIM dataset, with the aligned ResNet-34 backbone, DiffusionDrive achieves 88.1 PDMS without bells and whistles, setting a new record, while running at a real-time speed of 45 FPS on an NVIDIA 4090. Qualitative results on challenging scenarios further confirm that DiffusionDrive can robustly generate diverse plausible driving actions. Code and model will be available at https://github.com/hustvl/DiffusionDrive.
comment: Accepted to CVPR 2025. Code & demo & model are available at https://github.com/hustvl/DiffusionDrive
♻ ☆ Applications of Spiking Neural Networks in Visual Place Recognition
In robotics, Spiking Neural Networks (SNNs) are increasingly recognized for their largely-unrealized potential energy efficiency and low latency particularly when implemented on neuromorphic hardware. Our paper highlights three advancements for SNNs in Visual Place Recognition (VPR). Firstly, we propose Modular SNNs, where each SNN represents a set of non-overlapping geographically distinct places, enabling scalable networks for large environments. Secondly, we present Ensembles of Modular SNNs, where multiple networks represent the same place, significantly enhancing accuracy compared to single-network models. Each of our Modular SNN modules is compact, comprising only 1500 neurons and 474k synapses, making them ideally suited for ensembling due to their small size. Lastly, we investigate the role of sequence matching in SNN-based VPR, a technique where consecutive images are used to refine place recognition. We demonstrate competitive performance of our method on a range of datasets, including higher responsiveness to ensembling compared to conventional VPR techniques and higher R@1 improvements with sequence matching than VPR techniques with comparable baseline performance. Our contributions highlight the viability of SNNs for VPR, offering scalable and robust solutions, and paving the way for their application in various energy-sensitive robotic tasks.
comment: 20 pages, 10 figures, IEEE Transactions on Robotics (TRO)
♻ ☆ AVR: Active Vision-Driven Robotic Precision Manipulation with Viewpoint and Focal Length Optimization
Robotic manipulation within dynamic environments presents challenges to precise control and adaptability. Traditional fixed-view camera systems face challenges adapting to change viewpoints and scale variations, limiting perception and manipulation precision. To tackle these issues, we propose the Active Vision-driven Robotic (AVR) framework, a teleoperation hardware solution that supports dynamic viewpoint and dynamic focal length adjustments to continuously center targets and maintain optimal scale, accompanied by a corresponding algorithm that effectively enhances the success rates of various operational tasks. Using the RoboTwin platform with a real-time image processing plugin, AVR framework improves task success rates by 5%-16% on five manipulation tasks. Physical deployment on a dual-arm system demonstrates in collaborative tasks and 36% precision in screwdriver insertion, outperforming baselines by over 25%. Experimental results confirm that AVR framework enhances environmental perception, manipulation repeatability (40% $\le $1 cm error), and robustness in complex scenarios, paving the way for future robotic precision manipulation methods in the pursuit of human-level robot dexterity and precision.
comment: Previously, there were some problems with our experimental data, and the conclusions need to be further verified. Now that we have completed a full-scale experiment and analysis, and added supporting materials to our website, we hope to be able to resubmit it
♻ ☆ 3D-MVP: 3D Multiview Pretraining for Robotic Manipulation CVPR 2025
Recent works have shown that visual pretraining on egocentric datasets using masked autoencoders (MAE) can improve generalization for downstream robotics tasks. However, these approaches pretrain only on 2D images, while many robotics applications require 3D scene understanding. In this work, we propose 3D-MVP, a novel approach for 3D Multi-View Pretraining using masked autoencoders. We leverage Robotic View Transformer (RVT), which uses a multi-view transformer to understand the 3D scene and predict gripper pose actions. We split RVT's multi-view transformer into visual encoder and action decoder, and pretrain its visual encoder using masked autoencoding on large-scale 3D datasets such as Objaverse. We evaluate 3D-MVP on a suite of virtual robot manipulation tasks and demonstrate improved performance over baselines. Our results suggest that 3D-aware pretraining is a promising approach to improve generalization of vision-based robotic manipulation policies. Project site: https://jasonqsy.github.io/3DMVP
comment: CVPR 2025
♻ ☆ MVCTrack: Boosting 3D Point Cloud Tracking via Multimodal-Guided Virtual Cues ICRA 2025
3D single object tracking is essential in autonomous driving and robotics. Existing methods often struggle with sparse and incomplete point cloud scenarios. To address these limitations, we propose a Multimodal-guided Virtual Cues Projection (MVCP) scheme that generates virtual cues to enrich sparse point clouds. Additionally, we introduce an enhanced tracker MVCTrack based on the generated virtual cues. Specifically, the MVCP scheme seamlessly integrates RGB sensors into LiDAR-based systems, leveraging a set of 2D detections to create dense 3D virtual cues that significantly improve the sparsity of point clouds. These virtual cues can naturally integrate with existing LiDAR-based 3D trackers, yielding substantial performance gains. Extensive experiments demonstrate that our method achieves competitive performance on the NuScenes dataset.
comment: Accepted by ICRA 2025
♻ ☆ NextStop: An Improved Tracker For Panoptic LIDAR Segmentation Data
4D panoptic LiDAR segmentation is essential for scene understanding in autonomous driving and robotics, combining semantic and instance segmentation with temporal consistency. Current methods, like 4D-PLS and 4D-STOP, use a tracking-by-detection methodology, employing deep learning networks to perform semantic and instance segmentation on each frame. To maintain temporal consistency, large-size instances detected in the current frame are compared and associated with instances within a temporal window that includes the current and preceding frames. However, their reliance on short-term instance detection, lack of motion estimation, and exclusion of small-sized instances lead to frequent identity switches and reduced tracking performance. We address these issues with the NextStop1 tracker, which integrates Kalman filter-based motion estimation, data association, and lifespan management, along with a tracklet state concept to improve prioritization. Evaluated using the LiDAR Segmentation and Tracking Quality (LSTQ) metric on the SemanticKITTI validation set, NextStop demonstrated enhanced tracking performance, particularly for small-sized objects like people and bicyclists, with fewer ID switches, earlier tracking initiation, and improved reliability in complex environments. The source code is available at https://github.com/AIROTAU/NextStop
♻ ☆ MGSO: Monocular Real-time Photometric SLAM with Efficient 3D Gaussian Splatting
Real-time SLAM with dense 3D mapping is computationally challenging, especially on resource-limited devices. The recent development of 3D Gaussian Splatting (3DGS) offers a promising approach for real-time dense 3D reconstruction. However, existing 3DGS-based SLAM systems struggle to balance hardware simplicity, speed, and map quality. Most systems excel in one or two of the aforementioned aspects but rarely achieve all. A key issue is the difficulty of initializing 3D Gaussians while concurrently conducting SLAM. To address these challenges, we present Monocular GSO (MGSO), a novel real-time SLAM system that integrates photometric SLAM with 3DGS. Photometric SLAM provides dense structured point clouds for 3DGS initialization, accelerating optimization and producing more efficient maps with fewer Gaussians. As a result, experiments show that our system generates reconstructions with a balance of quality, memory efficiency, and speed that outperforms the state-of-the-art. Furthermore, our system achieves all results using RGB inputs. We evaluate the Replica, TUM-RGBD, and EuRoC datasets against current live dense reconstruction systems. Not only do we surpass contemporary systems, but experiments also show that we maintain our performance on laptop hardware, making it a practical solution for robotics, A/R, and other real-time applications.
comment: The final version of this work has been approved by the IEEE for publication. This version may no longer be accessible without notice. Copyright 2025 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses
♻ ☆ CAHSOR: Competence-Aware High-Speed Off-Road Ground Navigation in SE(3)
While the workspace of traditional ground vehicles is usually assumed to be in a 2D plane, i.e., SE(2), such an assumption may not hold when they drive at high speeds on unstructured off-road terrain: High-speed sharp turns on high-friction surfaces may lead to vehicle rollover; Turning aggressively on loose gravel or grass may violate the non-holonomic constraint and cause significant lateral sliding; Driving quickly on rugged terrain will produce extensive vibration along the vertical axis. Therefore, most offroad vehicles are currently limited to drive only at low speeds to assure vehicle stability and safety. In this work, we aim at empowering high-speed off-road vehicles with competence awareness in SE(3) so that they can reason about the consequences of taking aggressive maneuvers on different terrain with a 6-DoF forward kinodynamic model. The model is learned from visual and inertial Terrain Representation for Off-road Navigation (TRON) using multimodal, self-supervised vehicle-terrain interactions. We demonstrate the efficacy of our Competence-Aware High-Speed Off-Road (CAHSOR) navigation approach on a physical ground robot in both an autonomous navigation and a human shared-control setup and show that CAHSOR can efficiently reduce vehicle instability by 62% while only compromising 8.6% average speed with the help of TRON.
♻ ☆ A Rapid Trajectory Optimization and Control Framework for Resource-Constrained Applications
This paper presents a computationally efficient model predictive control formulation that uses an integral Chebyshev collocation method to enable rapid operations of autonomous agents. By posing the finite-horizon optimal control problem and recursive re-evaluation of the optimal trajectories, minimization of the L2 norms of the state and control errors are transcribed into a quadratic program. Control and state variable constraints are parameterized using Chebyshev polynomials and are accommodated in the optimal trajectory generation programs to incorporate the actuator limits and keep-out constraints. Differentiable collision detection of polytopes is leveraged for optimal collision avoidance. Results obtained from the collocation methods are benchmarked against the existing approaches on an edge computer to outline the performance improvements. Finally, collaborative control scenarios involving multi-agent space systems are considered to demonstrate the technical merits of the proposed work.
comment: This work has been accepted for publication at the IEEE ACC 2025
♻ ☆ Morphological Symmetries in Robotics
We present a comprehensive framework for studying and leveraging morphological symmetries in robotic systems. These are intrinsic properties of the robot's morphology, frequently observed in animal biology and robotics, which stem from the replication of kinematic structures and the symmetrical distribution of mass. We illustrate how these symmetries extend to the robot's state space and both proprioceptive and exteroceptive sensor measurements, resulting in the equivariance of the robot's equations of motion and optimal control policies. Thus, we recognize morphological symmetries as a relevant and previously unexplored physics-informed geometric prior, with significant implications for both data-driven and analytical methods used in modeling, control, estimation and design in robotics. For data-driven methods, we demonstrate that morphological symmetries can enhance the sample efficiency and generalization of machine learning models through data augmentation, or by applying equivariant/invariant constraints on the model's architecture. In the context of analytical methods, we employ abstract harmonic analysis to decompose the robot's dynamics into a superposition of lower-dimensional, independent dynamics. We substantiate our claims with both synthetic and real-world experiments conducted on bipedal and quadrupedal robots. Lastly, we introduce the repository MorphoSymm to facilitate the practical use of the theory and applications outlined in this work.
comment: 18 pages, 11 figures
Artificial Intelligence 188
☆ Aether: Geometric-Aware Unified World Modeling
The integration of geometric reconstruction and generative modeling remains a critical challenge in developing AI systems capable of human-like spatial reasoning. This paper proposes Aether, a unified framework that enables geometry-aware reasoning in world models by jointly optimizing three core capabilities: (1) 4D dynamic reconstruction, (2) action-conditioned video prediction, and (3) goal-conditioned visual planning. Through task-interleaved feature learning, Aether achieves synergistic knowledge sharing across reconstruction, prediction, and planning objectives. Building upon video generation models, our framework demonstrates unprecedented synthetic-to-real generalization despite never observing real-world data during training. Furthermore, our approach achieves zero-shot generalization in both action following and reconstruction tasks, thanks to its intrinsic geometric modeling. Remarkably, even without real-world data, its reconstruction performance far exceeds that of domain-specific models. Additionally, Aether leverages a geometry-informed action space to seamlessly translate predictions into actions, enabling effective autonomous trajectory planning. We hope our work inspires the community to explore new frontiers in physically-reasonable world modeling and its applications.
comment: Project Page: https://aether-world.github.io/
☆ Video-T1: Test-Time Scaling for Video Generation
With the scale capability of increasing training data, model size, and computational cost, video generation has achieved impressive results in digital creation, enabling users to express creativity across various domains. Recently, researchers in Large Language Models (LLMs) have expanded the scaling to test-time, which can significantly improve LLM performance by using more inference-time computation. Instead of scaling up video foundation models through expensive training costs, we explore the power of Test-Time Scaling (TTS) in video generation, aiming to answer the question: if a video generation model is allowed to use non-trivial amount of inference-time compute, how much can it improve generation quality given a challenging text prompt. In this work, we reinterpret the test-time scaling of video generation as a searching problem to sample better trajectories from Gaussian noise space to the target video distribution. Specifically, we build the search space with test-time verifiers to provide feedback and heuristic algorithms to guide searching process. Given a text prompt, we first explore an intuitive linear search strategy by increasing noise candidates at inference time. As full-step denoising all frames simultaneously requires heavy test-time computation costs, we further design a more efficient TTS method for video generation called Tree-of-Frames (ToF) that adaptively expands and prunes video branches in an autoregressive manner. Extensive experiments on text-conditioned video generation benchmarks demonstrate that increasing test-time compute consistently leads to significant improvements in the quality of videos. Project page: https://liuff19.github.io/Video-T1
comment: Project page: https://liuff19.github.io/Video-T1
AdaWorld: Learning Adaptable World Models with Latent Actions
World models aim to learn action-controlled prediction models and have proven essential for the development of intelligent agents. However, most existing world models rely heavily on substantial action-labeled data and costly training, making it challenging to adapt to novel environments with heterogeneous actions through limited interactions. This limitation can hinder their applicability across broader domains. To overcome this challenge, we propose AdaWorld, an innovative world model learning approach that enables efficient adaptation. The key idea is to incorporate action information during the pretraining of world models. This is achieved by extracting latent actions from videos in a self-supervised manner, capturing the most critical transitions between frames. We then develop an autoregressive world model that conditions on these latent actions. This learning paradigm enables highly adaptable world models, facilitating efficient transfer and learning of new actions even with limited interactions and finetuning. Our comprehensive experiments across multiple environments demonstrate that AdaWorld achieves superior performance in both simulation quality and visual planning.
comment: Project page: https://adaptable-world-model.github.io/
☆ Statistical Proof of Execution (SPEX)
Many real-world applications are increasingly incorporating automated decision-making, driven by the widespread adoption of ML/AI inference for planning and guidance. This study examines the growing need for verifiable computing in autonomous decision-making. We formalize the problem of verifiable computing and introduce a sampling-based protocol that is significantly faster, more cost-effective, and simpler than existing methods. Furthermore, we tackle the challenges posed by non-determinism, proposing a set of strategies to effectively manage common scenarios.
☆ SimpleRL-Zoo: Investigating and Taming Zero Reinforcement Learning for Open Base Models in the Wild
DeepSeek-R1 has shown that long chain-of-thought (CoT) reasoning can naturally emerge through a simple reinforcement learning (RL) framework with rule-based rewards, where the training may directly start from the base models-a paradigm referred to as zero RL training. Most recent efforts to reproduce zero RL training have primarily focused on the Qwen2.5 model series, which may not be representative as we find the base models already exhibit strong instruction-following and self-reflection abilities. In this work, we investigate zero RL training across 10 diverse base models, spanning different families and sizes including LLama3-8B, Mistral-7B/24B, DeepSeek-Math-7B, Qwen2.5-math-7B, and all Qwen2.5 models from 0.5B to 32B. Leveraging several key design strategies-such as adjusting format reward and controlling query difficulty-we achieve substantial improvements in both reasoning accuracy and response length across most settings. However, by carefully monitoring the training dynamics, we observe that different base models exhibit distinct patterns during training. For instance, the increased response length does not always correlate with the emergence of certain cognitive behaviors such as verification (i.e., the "aha moment"). Notably, we observe the "aha moment" for the first time in small models not from the Qwen family. We share the key designs that enable successful zero RL training, along with our findings and practices. To facilitate further research, we open-source the code, models, and analysis tools.
☆ AgentDropout: Dynamic Agent Elimination for Token-Efficient and High-Performance LLM-Based Multi-Agent Collaboration
Multi-agent systems (MAS) based on large language models (LLMs) have demonstrated significant potential in collaborative problem-solving. However, they still face substantial challenges of low communication efficiency and suboptimal task performance, making the careful design of the agents' communication topologies particularly important. Inspired by the management theory that roles in an efficient team are often dynamically adjusted, we propose AgentDropout, which identifies redundant agents and communication across different communication rounds by optimizing the adjacency matrices of the communication graphs and eliminates them to enhance both token efficiency and task performance. Compared to state-of-the-art methods, AgentDropout achieves an average reduction of 21.6% in prompt token consumption and 18.4% in completion token consumption, along with a performance improvement of 1.14 on the tasks. Furthermore, the extended experiments demonstrate that AgentDropout achieves notable domain transferability and structure robustness, revealing its reliability and effectiveness. We release our code at https://github.com/wangzx1219/AgentDropout.
☆ Bootstrapped Model Predictive Control ICLR 2025
Model Predictive Control (MPC) has been demonstrated to be effective in continuous control tasks. When a world model and a value function are available, planning a sequence of actions ahead of time leads to a better policy. Existing methods typically obtain the value function and the corresponding policy in a model-free manner. However, we find that such an approach struggles with complex tasks, resulting in poor policy learning and inaccurate value estimation. To address this problem, we leverage the strengths of MPC itself. In this work, we introduce Bootstrapped Model Predictive Control (BMPC), a novel algorithm that performs policy learning in a bootstrapped manner. BMPC learns a network policy by imitating an MPC expert, and in turn, uses this policy to guide the MPC process. Combined with model-based TD-learning, our policy learning yields better value estimation and further boosts the efficiency of MPC. We also introduce a lazy reanalyze mechanism, which enables computationally efficient imitation learning. Our method achieves superior performance over prior works on diverse continuous control tasks. In particular, on challenging high-dimensional locomotion tasks, BMPC significantly improves data efficiency while also enhancing asymptotic performance and training stability, with comparable training time and smaller network sizes. Code is available at https://github.com/wertyuilife2/bmpc.
comment: Published as a conference paper at ICLR 2025
☆ Reasoning to Learn from Latent Thoughts
Compute scaling for language model (LM) pretraining has outpaced the growth of human-written texts, leading to concerns that data will become the bottleneck to LM scaling. To continue scaling pretraining in this data-constrained regime, we propose that explicitly modeling and inferring the latent thoughts that underlie the text generation process can significantly improve pretraining data efficiency. Intuitively, our approach views web text as the compressed final outcome of a verbose human thought process and that the latent thoughts contain important contextual knowledge and reasoning steps that are critical to data-efficient learning. We empirically demonstrate the effectiveness of our approach through data-constrained continued pretraining for math. We first show that synthetic data approaches to inferring latent thoughts significantly improve data efficiency, outperforming training on the same amount of raw data (5.7\% $\rightarrow$ 25.4\% on MATH). Furthermore, we demonstrate latent thought inference without a strong teacher, where an LM bootstraps its own performance by using an EM algorithm to iteratively improve the capability of the trained LM and the quality of thought-augmented pretraining data. We show that a 1B LM can bootstrap its performance across at least three iterations and significantly outperform baselines trained on raw data, with increasing gains from additional inference compute when performing the E-step. The gains from inference scaling and EM iterations suggest new opportunities for scaling data-constrained pretraining.
☆ Structuring Scientific Innovation: A Framework for Modeling and Discovering Impactful Knowledge Combinations
The emergence of large language models offers new possibilities for structured exploration of scientific knowledge. Rather than viewing scientific discovery as isolated ideas or content, we propose a structured approach that emphasizes the role of method combinations in shaping disruptive insights. Specifically, we investigate how knowledge unit--especially those tied to methodological design--can be modeled and recombined to yield research breakthroughs.Our proposed framework addresses two key challenges. First, we introduce a contrastive learning-based mechanism to identify distinguishing features of historically disruptive method combinations within problem-driven contexts.Second, we propose a reasoning-guided Monte Carlo search algorithm that leverages the chain-of-thought capability of LLMs to identify promising knowledge recombinations for new problem statements.Empirical studies across multiple domains show that the framework is capable of modeling the structural dynamics of innovation and successfully highlights combinations with high disruptive potential.This research provides a new path for computationally guided scientific ideation grounded in structured reasoning and historical data modeling.
☆ Exploring the Integration of Key-Value Attention Into Pure and Hybrid Transformers for Semantic Segmentation
While CNNs were long considered state of the art for image processing, the introduction of Transformer architectures has challenged this position. While achieving excellent results in image classification and segmentation, Transformers remain inherently reliant on large training datasets and remain computationally expensive. A newly introduced Transformer derivative named KV Transformer shows promising results in synthetic, NLP, and image classification tasks, while reducing complexity and memory usage. This is especially conducive to use cases where local inference is required, such as medical screening applications. We endeavoured to further evaluate the merit of KV Transformers on semantic segmentation tasks, specifically in the domain of medical imaging. By directly comparing traditional and KV variants of the same base architectures, we provide further insight into the practical tradeoffs of reduced model complexity. We observe a notable reduction in parameter count and multiply accumulate operations, while achieving similar performance from most of the KV variant models when directly compared to their QKV implementation.
comment: 6 pages, 3 figures, Preprint. Final version published in: Bildverarbeitung f\"ur die Medizin 2025, Springer. DOI: https://doi.org/10.1007/978-3-658-47422-5_71
☆ MC-LLaVA: Multi-Concept Personalized Vision-Language Model
Current vision-language models (VLMs) show exceptional abilities across diverse tasks, such as visual question answering. To enhance user experience, recent studies investigate VLM personalization to understand user-provided concepts. However, they mainly focus on single-concept personalization, neglecting the existence and interplay of multiple concepts, which limits real-world applicability. This paper proposes the first multi-concept personalization paradigm, MC-LLaVA. Specifically, MC-LLaVA employs a multi-concept instruction tuning strategy, effectively integrating multiple concepts in a single training step. To reduce the costs related to joint training, we propose a personalized textual prompt that uses visual token information to initialize concept tokens. Additionally, we introduce a personalized visual prompt during inference, aggregating location confidence maps for enhanced recognition and grounding capabilities. To advance multi-concept personalization research, we further contribute a high-quality instruction tuning dataset. We carefully collect images with multiple characters and objects from movies and manually generate question-answer samples for multi-concept scenarios, featuring superior diversity. Comprehensive qualitative and quantitative experiments demonstrate that MC-LLaVA can achieve impressive multi-concept personalized responses, paving the way for VLMs to become better user-specific assistants. The code and dataset will be publicly available at $\href{https://github.com/arctanxarc/MC-LLaVA}{https://github.com/arctanxarc/MC-LLaVA}$.
comment: The code and dataset will be publicly available at $\href{https://github.com/arctanxarc/MC-LLaVA}{https://github.com/arctanxarc/MC-LLaVA}$
☆ Self-Organizing Graph Reasoning Evolves into a Critical State for Continuous Discovery Through Structural-Semantic Dynamics
We report fundamental insights into how agentic graph reasoning systems spontaneously evolve toward a critical state that sustains continuous semantic discovery. By rigorously analyzing structural (Von Neumann graph entropy) and semantic (embedding) entropy, we identify a subtle yet robust regime in which semantic entropy persistently dominates over structural entropy. This interplay is quantified by a dimensionless Critical Discovery Parameter that stabilizes at a small negative value, indicating a consistent excess of semantic entropy. Empirically, we observe a stable fraction (12%) of "surprising" edges, links between semantically distant concepts, providing evidence of long-range or cross-domain connections that drive continuous innovation. Concomitantly, the system exhibits scale-free and small-world topological features, alongside a negative cross-correlation between structural and semantic measures, reinforcing the analogy to self-organized criticality. These results establish clear parallels with critical phenomena in physical, biological, and cognitive complex systems, revealing an entropy-based principle governing adaptability and continuous innovation. Crucially, semantic richness emerges as the underlying driver of sustained exploration, despite not being explicitly used by the reasoning process. Our findings provide interdisciplinary insights and practical strategies for engineering intelligent systems with intrinsic capacities for long-term discovery and adaptation, and offer insights into how model training strategies can be developed that reinforce critical discovery.
☆ Three Kinds of AI Ethics
There is an overwhelmingly abundance of works in AI Ethics. This growth is chaotic because of how sudden it is, its volume, and its multidisciplinary nature. This makes difficult to keep track of debates, and to systematically characterize goals, research questions, methods, and expertise required by AI ethicists. In this article, I show that the relation between AI and ethics can be characterized in at least three ways, which correspond to three well-represented kinds of AI ethics: ethics and AI; ethics in AI; ethics of AI. I elucidate the features of these three kinds of AI Ethics, characterize their research questions, and identify the kind of expertise that each kind needs. I also show how certain criticisms to AI ethics are misplaced, as being done from the point of view of one kind of AI ethics, to another kind with different goals. All in all, this work sheds light on the nature of AI ethics, and set the grounds for more informed discussions about scope, methods, and trainings of AI ethicists.
comment: 16 pages, two figures
☆ Dual-domain Multi-path Self-supervised Diffusion Model for Accelerated MRI Reconstruction
Magnetic resonance imaging (MRI) is a vital diagnostic tool, but its inherently long acquisition times reduce clinical efficiency and patient comfort. Recent advancements in deep learning, particularly diffusion models, have improved accelerated MRI reconstruction. However, existing diffusion models' training often relies on fully sampled data, models incur high computational costs, and often lack uncertainty estimation, limiting their clinical applicability. To overcome these challenges, we propose a novel framework, called Dual-domain Multi-path Self-supervised Diffusion Model (DMSM), that integrates a self-supervised dual-domain diffusion model training scheme, a lightweight hybrid attention network for the reconstruction diffusion model, and a multi-path inference strategy, to enhance reconstruction accuracy, efficiency, and explainability. Unlike traditional diffusion-based models, DMSM eliminates the dependency on training from fully sampled data, making it more practical for real-world clinical settings. We evaluated DMSM on two human MRI datasets, demonstrating that it achieves favorable performance over several supervised and self-supervised baselines, particularly in preserving fine anatomical structures and suppressing artifacts under high acceleration factors. Additionally, our model generates uncertainty maps that correlate reasonably well with reconstruction errors, offering valuable clinically interpretable guidance and potentially enhancing diagnostic confidence.
comment: 10 pages, 8 figures, 5 tables
☆ Interpretable and Fair Mechanisms for Abstaining Classifiers ECML
Abstaining classifiers have the option to refrain from providing a prediction for instances that are difficult to classify. The abstention mechanism is designed to trade off the classifier's performance on the accepted data while ensuring a minimum number of predictions. In this setting, often fairness concerns arise when the abstention mechanism solely reduces errors for the majority groups of the data, resulting in increased performance differences across demographic groups. While there exist a bunch of methods that aim to reduce discrimination when abstaining, there is no mechanism that can do so in an explainable way. In this paper, we fill this gap by introducing Interpretable and Fair Abstaining Classifier IFAC, an algorithm that can reject predictions both based on their uncertainty and their unfairness. By rejecting possibly unfair predictions, our method reduces error and positive decision rate differences across demographic groups of the non-rejected data. Since the unfairness-based rejections are based on an interpretable-by-design method, i.e., rule-based fairness checks and situation testing, we create a transparent process that can empower human decision-makers to review the unfair predictions and make more just decisions for them. This explainable aspect is especially important in light of recent AI regulations, mandating that any high-risk decision task should be overseen by human experts to reduce discrimination risks.
comment: 25 pages, 8 figures. In: Machine Learning and Knowledge Discovery in Databases. Research Track. ECML PKDD 2024
☆ EconEvals: Benchmarks and Litmus Tests for LLM Agents in Unknown Environments
We develop benchmarks for LLM agents that act in, learn from, and strategize in unknown environments, the specifications of which the LLM agent must learn over time from deliberate exploration. Our benchmarks consist of decision-making tasks derived from key problems in economics. To forestall saturation, the benchmark tasks are synthetically generated with scalable difficulty levels. Additionally, we propose litmus tests, a new kind of quantitative measure for LLMs and LLM agents. Unlike benchmarks, litmus tests quantify differences in character, values, and tendencies of LLMs and LLM agents, by considering their behavior when faced with tradeoffs (e.g., efficiency versus equality) where there is no objectively right or wrong behavior. Overall, our benchmarks and litmus tests assess the abilities and tendencies of LLM agents in tackling complex economic problems in diverse settings spanning procurement, scheduling, task allocation, and pricing -- applications that should grow in importance as such agents are further integrated into the economy.
☆ Enhanced OoD Detection through Cross-Modal Alignment of Multi-Modal Representations CVPR 2025
Prior research on out-of-distribution detection (OoDD) has primarily focused on single-modality models. Recently, with the advent of large-scale pretrained vision-language models such as CLIP, OoDD methods utilizing such multi-modal representations through zero-shot and prompt learning strategies have emerged. However, these methods typically involve either freezing the pretrained weights or only partially tuning them, which can be suboptimal for downstream datasets. In this paper, we highlight that multi-modal fine-tuning (MMFT) can achieve notable OoDD performance. Despite some recent works demonstrating the impact of fine-tuning methods for OoDD, there remains significant potential for performance improvement. We investigate the limitation of na\"ive fine-tuning methods, examining why they fail to fully leverage the pretrained knowledge. Our empirical analysis suggests that this issue could stem from the modality gap within in-distribution (ID) embeddings. To address this, we propose a training objective that enhances cross-modal alignment by regularizing the distances between image and text embeddings of ID data. This adjustment helps in better utilizing pretrained textual information by aligning similar semantics from different modalities (i.e., text and image) more closely in the hyperspherical representation space. We theoretically demonstrate that the proposed regularization corresponds to the maximum likelihood estimation of an energy-based model on a hypersphere. Utilizing ImageNet-1k OoD benchmark datasets, we show that our method, combined with post-hoc OoDD approaches leveraging pretrained knowledge (e.g., NegLabel), significantly outperforms existing methods, achieving state-of-the-art OoDD performance and leading ID accuracy.
comment: CVPR 2025
☆ Learning Multi-Robot Coordination through Locality-Based Factorized Multi-Agent Actor-Critic Algorithm
In this work, we present a novel cooperative multi-agent reinforcement learning method called \textbf{Loc}ality based \textbf{Fac}torized \textbf{M}ulti-Agent \textbf{A}ctor-\textbf{C}ritic (Loc-FACMAC). Existing state-of-the-art algorithms, such as FACMAC, rely on global reward information, which may not accurately reflect the quality of individual robots' actions in decentralized systems. We integrate the concept of locality into critic learning, where strongly related robots form partitions during training. Robots within the same partition have a greater impact on each other, leading to more precise policy evaluation. Additionally, we construct a dependency graph to capture the relationships between robots, facilitating the partitioning process. This approach mitigates the curse of dimensionality and prevents robots from using irrelevant information. Our method improves existing algorithms by focusing on local rewards and leveraging partition-based learning to enhance training efficiency and performance. We evaluate the performance of Loc-FACMAC in three environments: Hallway, Multi-cartpole, and Bounded-Cooperative-Navigation. We explore the impact of partition sizes on the performance and compare the result with baseline MARL algorithms such as LOMAQ, FACMAC, and QMIX. The experiments reveal that, if the locality structure is defined properly, Loc-FACMAC outperforms these baseline algorithms up to 108\%, indicating that exploiting the locality structure in the actor-critic framework improves the MARL performance.
☆ Towards Responsible AI Music: an Investigation of Trustworthy Features for Creative Systems
Generative AI is radically changing the creative arts, by fundamentally transforming the way we create and interact with cultural artefacts. While offering unprecedented opportunities for artistic expression and commercialisation, this technology also raises ethical, societal, and legal concerns. Key among these are the potential displacement of human creativity, copyright infringement stemming from vast training datasets, and the lack of transparency, explainability, and fairness mechanisms. As generative systems become pervasive in this domain, responsible design is crucial. Whilst previous work has tackled isolated aspects of generative systems (e.g., transparency, evaluation, data), we take a comprehensive approach, grounding these efforts within the Ethics Guidelines for Trustworthy Artificial Intelligence produced by the High-Level Expert Group on AI appointed by the European Commission - a framework for designing responsible AI systems across seven macro requirements. Focusing on generative music AI, we illustrate how these requirements can be contextualised for the field, addressing trustworthiness across multiple dimensions and integrating insights from the existing literature. We further propose a roadmap for operationalising these contextualised requirements, emphasising interdisciplinary collaboration and stakeholder engagement. Our work provides a foundation for designing and evaluating responsible music generation systems, calling for collaboration among AI experts, ethicists, legal scholars, and artists. This manuscript is accompanied by a website: https://amresearchlab.github.io/raim-framework/.
☆ Defeating Prompt Injections by Design
Large Language Models (LLMs) are increasingly deployed in agentic systems that interact with an external environment. However, LLM agents are vulnerable to prompt injection attacks when handling untrusted data. In this paper we propose CaMeL, a robust defense that creates a protective system layer around the LLM, securing it even when underlying models may be susceptible to attacks. To operate, CaMeL explicitly extracts the control and data flows from the (trusted) query; therefore, the untrusted data retrieved by the LLM can never impact the program flow. To further improve security, CaMeL relies on a notion of a capability to prevent the exfiltration of private data over unauthorized data flows. We demonstrate effectiveness of CaMeL by solving $67\%$ of tasks with provable security in AgentDojo [NeurIPS 2024], a recent agentic security benchmark.
☆ Classical Planning with LLM-Generated Heuristics: Challenging the State of the Art with Python Code
In recent years, large language models (LLMs) have shown remarkable capabilities in various artificial intelligence problems. However, they fail to plan reliably, even when prompted with a detailed definition of the planning task. Attempts to improve their planning capabilities, such as chain-of-thought prompting, fine-tuning, and explicit "reasoning" still yield incorrect plans and usually fail to generalize to larger tasks. In this paper, we show how to use LLMs to generate correct plans, even for out-of-distribution tasks of increasing size. For a given planning domain, we ask an LLM to generate several domain-dependent heuristic functions in the form of Python code, evaluate them on a set of training tasks within a greedy best-first search, and choose the strongest one. The resulting LLM-generated heuristics solve many more unseen test tasks than state-of-the-art domain-independent heuristics for classical planning. They are even competitive with the strongest learning algorithm for domain-dependent planning. These findings are especially remarkable given that our proof-of-concept implementation is based on an unoptimized Python planner and the baselines all build upon highly optimized C++ code. In some domains, the LLM-generated heuristics expand fewer states than the baselines, revealing that they are not only efficiently computable, but sometimes even more informative than the state-of-the-art heuristics. Overall, our results show that sampling a set of planning heuristic function programs can significantly improve the planning capabilities of LLMs.
☆ REALM: A Dataset of Real-World LLM Use Cases
Large Language Models, such as the GPT series, have driven significant industrial applications, leading to economic and societal transformations. However, a comprehensive understanding of their real-world applications remains limited. To address this, we introduce REALM, a dataset of over 94,000 LLM use cases collected from Reddit and news articles. REALM captures two key dimensions: the diverse applications of LLMs and the demographics of their users. It categorizes LLM applications and explores how users' occupations relate to the types of applications they use. By integrating real-world data, REALM offers insights into LLM adoption across different domains, providing a foundation for future research on their evolving societal roles. A dedicated dashboard https://realm-e7682.web.app/ presents the data.
comment: 9 pages, 5 figures
☆ Frequency Dynamic Convolution for Dense Image Prediction CVPR 2025
While Dynamic Convolution (DY-Conv) has shown promising performance by enabling adaptive weight selection through multiple parallel weights combined with an attention mechanism, the frequency response of these weights tends to exhibit high similarity, resulting in high parameter costs but limited adaptability. In this work, we introduce Frequency Dynamic Convolution (FDConv), a novel approach that mitigates these limitations by learning a fixed parameter budget in the Fourier domain. FDConv divides this budget into frequency-based groups with disjoint Fourier indices, enabling the construction of frequency-diverse weights without increasing the parameter cost. To further enhance adaptability, we propose Kernel Spatial Modulation (KSM) and Frequency Band Modulation (FBM). KSM dynamically adjusts the frequency response of each filter at the spatial level, while FBM decomposes weights into distinct frequency bands in the frequency domain and modulates them dynamically based on local content. Extensive experiments on object detection, segmentation, and classification validate the effectiveness of FDConv. We demonstrate that when applied to ResNet-50, FDConv achieves superior performance with a modest increase of +3.6M parameters, outperforming previous methods that require substantial increases in parameter budgets (e.g., CondConv +90M, KW +76.5M). Moreover, FDConv seamlessly integrates into a variety of architectures, including ConvNeXt, Swin-Transformer, offering a flexible and efficient solution for modern vision tasks. The code is made publicly available at https://github.com/Linwei-Chen/FDConv.
comment: Accepted by CVPR 2025
☆ The case for delegated AI autonomy for Human AI teaming in healthcare
In this paper we propose an advanced approach to integrating artificial intelligence (AI) into healthcare: autonomous decision support. This approach allows the AI algorithm to act autonomously for a subset of patient cases whilst serving a supportive role in other subsets of patient cases based on defined delegation criteria. By leveraging the complementary strengths of both humans and AI, it aims to deliver greater overall performance than existing human-AI teaming models. It ensures safe handling of patient cases and potentially reduces clinician review time, whilst being mindful of AI tool limitations. After setting the approach within the context of current human-AI teaming models, we outline the delegation criteria and apply them to a specific AI-based tool used in histopathology. The potential impact of the approach and the regulatory requirements for its successful implementation are then discussed.
☆ BitDecoding: Unlocking Tensor Cores for Long-Context LLMs Decoding with Low-Bit KV Cache
The growing adoption of long-context Large Language Models (LLMs) has introduced significant memory and computational challenges in autoregressive decoding due to the expanding Key-Value (KV) cache. KV cache quantization has emerged as a promising solution, with prior work showing that 4-bit or even 2-bit quantization can maintain model accuracy while reducing memory costs. However, despite these benefits, preliminary implementations for the low-bit KV cache struggle to deliver the expected speedup due to quantization and dequantization overheads and the lack of Tensor Cores utilization. In this work, we propose BitDecoding, a GPU-optimized framework that unlocks Tensor Cores for efficient decoding with low-bit KV cache. Efficiently leveraging Tensor Cores for low-bit KV cache is challenging due to the dynamic nature of KV cache generation at each decoding step. BitDecoding addresses these challenges with a Tensor Cores-Centric BitFusion Scheme that ensures data layout compatibility to enable high utilization of Tensor Cores. Additionally, BitDecoding incorporates a warp-efficient parallel decoding kernel and a fine-grained asynchronous pipeline, minimizing dequantization overhead and improving computational efficiency. Experiments show that BitDecoding achieves up to 7.5x speedup on RTX 4090, 4.8x on A100, and 8.9x on H100, compared to FP16 FlashDecoding-v2. It also outperforms the state-of-the-art low-bit KV cache implementation (QServe) by up to 4.3x. On LLaMA-3.1-8B with a 128K sequence length, BitDecoding reduces single-batch decoding latency by 3x, demonstrating its effectiveness in long-context generation scenarios. The code is available at https://github.com/DD-DuDa/BitDecoding.
☆ Mechanistic Interpretability of Fine-Tuned Vision Transformers on Distorted Images: Decoding Attention Head Behavior for Transparent and Trustworthy AI
Mechanistic interpretability improves the safety, reliability, and robustness of large AI models. This study examined individual attention heads in vision transformers (ViTs) fine tuned on distorted 2D spectrogram images containing non relevant content (axis labels, titles, color bars). By introducing extraneous features, the study analyzed how transformer components processed unrelated information, using mechanistic interpretability to debug issues and reveal insights into transformer architectures. Attention maps assessed head contributions across layers. Heads in early layers (1 to 3) showed minimal task impact with ablation increased MSE loss slightly ({\mu}=0.11%, {\sigma}=0.09%), indicating focus on less critical low level features. In contrast, deeper heads (e.g., layer 6) caused a threefold higher loss increase ({\mu}=0.34%, {\sigma}=0.02%), demonstrating greater task importance. Intermediate layers (6 to 11) exhibited monosemantic behavior, attending exclusively to chirp regions. Some early heads (1 to 4) were monosemantic but non task relevant (e.g. text detectors, edge or corner detectors). Attention maps distinguished monosemantic heads (precise chirp localization) from polysemantic heads (multiple irrelevant regions). These findings revealed functional specialization in ViTs, showing how heads processed relevant vs. extraneous information. By decomposing transformers into interpretable components, this work enhanced model understanding, identified vulnerabilities, and advanced safer, more transparent AI.
comment: 15 pages, 8 figures
☆ EgoSurgery-HTS: A Dataset for Egocentric Hand-Tool Segmentation in Open Surgery Videos
Egocentric open-surgery videos capture rich, fine-grained details essential for accurately modeling surgical procedures and human behavior in the operating room. A detailed, pixel-level understanding of hands and surgical tools is crucial for interpreting a surgeon's actions and intentions. We introduce EgoSurgery-HTS, a new dataset with pixel-wise annotations and a benchmark suite for segmenting surgical tools, hands, and interacting tools in egocentric open-surgery videos. Specifically, we provide a labeled dataset for (1) tool instance segmentation of 14 distinct surgical tools, (2) hand instance segmentation, and (3) hand-tool segmentation to label hands and the tools they manipulate. Using EgoSurgery-HTS, we conduct extensive evaluations of state-of-the-art segmentation methods and demonstrate significant improvements in the accuracy of hand and hand-tool segmentation in egocentric open-surgery videos compared to existing datasets. The dataset will be released at https://github.com/Fujiry0/EgoSurgery.
☆ Construction Identification and Disambiguation Using BERT: A Case Study of NPN ACL
Construction Grammar hypothesizes that knowledge of a language consists chiefly of knowledge of form-meaning pairs (''constructions'') that include vocabulary, general grammar rules, and even idiosyncratic patterns. Recent work has shown that transformer language models represent at least some constructional patterns, including ones where the construction is rare overall. In this work, we probe BERT's representation of the form and meaning of a minor construction of English, the NPN (noun-preposition-noun) construction -- exhibited in such expressions as face to face and day to day -- which is known to be polysemous. We construct a benchmark dataset of semantically annotated corpus instances (including distractors that superficially resemble the construction). With this dataset, we train and evaluate probing classifiers. They achieve decent discrimination of the construction from distractors, as well as sense disambiguation among true instances of the construction, revealing that BERT embeddings carry indications of the construction's semantics. Moreover, artificially permuting the word order of true construction instances causes them to be rejected, indicating sensitivity to matters of form. We conclude that BERT does latently encode at least some knowledge of the NPN construction going beyond a surface syntactic pattern and lexical cues.
comment: 8 pages, ACL long-paper format (preprint)
☆ Energy-Efficient Dynamic Training and Inference for GNN-Based Network Modeling
Efficient network modeling is essential for resource optimization and network planning in next-generation large-scale complex networks. Traditional approaches, such as queuing theory-based modeling and packet-based simulators, can be inefficient due to the assumption made and the computational expense, respectively. To address these challenges, we propose an innovative energy-efficient dynamic orchestration of Graph Neural Networks (GNN) based model training and inference framework for context-aware network modeling and predictions. We have developed a low-complexity solution framework, QAG, that is a Quantum approximation optimization (QAO) algorithm for Adaptive orchestration of GNN-based network modeling. We leverage the tripartite graph model to represent a multi-application system with many compute nodes. Thereafter, we apply the constrained graph-cutting using QAO to find the feasible energy-efficient configurations of the GNN-based model and deploying them on the available compute nodes to meet the network modeling application requirements. The proposed QAG scheme closely matches the optimum and offers atleast a 50% energy saving while meeting the application requirements with 60% lower churn-rate.
comment: Accepted in IEEE WCNC 2025
☆ Efficient Continual Adaptation of Pretrained Robotic Policy with Online Meta-Learned Adapters
Continual adaptation is essential for general autonomous agents. For example, a household robot pretrained with a repertoire of skills must still adapt to unseen tasks specific to each household. Motivated by this, building upon parameter-efficient fine-tuning in language models, prior works have explored lightweight adapters to adapt pretrained policies, which can preserve learned features from the pretraining phase and demonstrate good adaptation performances. However, these approaches treat task learning separately, limiting knowledge transfer between tasks. In this paper, we propose Online Meta-Learned adapters (OMLA). Instead of applying adapters directly, OMLA can facilitate knowledge transfer from previously learned tasks to current learning tasks through a novel meta-learning objective. Extensive experiments in both simulated and real-world environments demonstrate that OMLA can lead to better adaptation performances compared to the baseline methods. The project link: https://ricky-zhu.github.io/OMLA/.
comment: Project link: https://ricky-zhu.github.io/OMLA/
☆ Commander-GPT: Fully Unleashing the Sarcasm Detection Capability of Multi-Modal Large Language Models
Sarcasm detection, as a crucial research direction in the field of Natural Language Processing (NLP), has attracted widespread attention. Traditional sarcasm detection tasks have typically focused on single-modal approaches (e.g., text), but due to the implicit and subtle nature of sarcasm, such methods often fail to yield satisfactory results. In recent years, researchers have shifted the focus of sarcasm detection to multi-modal approaches. However, effectively leveraging multi-modal information to accurately identify sarcastic content remains a challenge that warrants further exploration. Leveraging the powerful integrated processing capabilities of Multi-Modal Large Language Models (MLLMs) for various information sources, we propose an innovative multi-modal Commander-GPT framework. Inspired by military strategy, we first decompose the sarcasm detection task into six distinct sub-tasks. A central commander (decision-maker) then assigns the best-suited large language model to address each specific sub-task. Ultimately, the detection results from each model are aggregated to identify sarcasm. We conducted extensive experiments on MMSD and MMSD 2.0, utilizing four multi-modal large language models and six prompting strategies. Our experiments demonstrate that our approach achieves state-of-the-art performance, with a 19.3% improvement in F1 score, without necessitating fine-tuning or ground-truth rationales.
☆ Any6D: Model-free 6D Pose Estimation of Novel Objects CVPR 2025
We introduce Any6D, a model-free framework for 6D object pose estimation that requires only a single RGB-D anchor image to estimate both the 6D pose and size of unknown objects in novel scenes. Unlike existing methods that rely on textured 3D models or multiple viewpoints, Any6D leverages a joint object alignment process to enhance 2D-3D alignment and metric scale estimation for improved pose accuracy. Our approach integrates a render-and-compare strategy to generate and refine pose hypotheses, enabling robust performance in scenarios with occlusions, non-overlapping views, diverse lighting conditions, and large cross-environment variations. We evaluate our method on five challenging datasets: REAL275, Toyota-Light, HO3D, YCBINEOAT, and LM-O, demonstrating its effectiveness in significantly outperforming state-of-the-art methods for novel object pose estimation. Project page: https://taeyeop.com/any6d
comment: CVPR 2025, Project Page: https://taeyeop.com/any6d
☆ AgentSpec: Customizable Runtime Enforcement for Safe and Reliable LLM Agents
Agents built on LLMs are increasingly deployed across diverse domains, automating complex decision-making and task execution. However, their autonomy introduces safety risks, including security vulnerabilities, legal violations, and unintended harmful actions. Existing mitigation methods, such as model-based safeguards and early enforcement strategies, fall short in robustness, interpretability, and adaptability. To address these challenges, we propose AgentSpec, a lightweight domain-specific language for specifying and enforcing runtime constraints on LLM agents. With AgentSpec, users define structured rules that incorporate triggers, predicates, and enforcement mechanisms, ensuring agents operate within predefined safety boundaries. We implement AgentSpec across multiple domains, including code execution, embodied agents, and autonomous driving, demonstrating its adaptability and effectiveness. Our evaluation shows that AgentSpec successfully prevents unsafe executions in over 90% of code agent cases, eliminates all hazardous actions in embodied agent tasks, and enforces 100% compliance by autonomous vehicles (AVs). Despite its strong safety guarantees, AgentSpec remains computationally lightweight, with overheads in milliseconds. By combining interpretability, modularity, and efficiency, AgentSpec provides a practical and scalable solution for enforcing LLM agent safety across diverse applications. We also automate the generation of rules using LLMs and assess their effectiveness. Our evaluation shows that the rules generated by OpenAI o1 achieve a precision of 95.56% and recall of 70.96% for embodied agents, successfully identifying 87.26% of the risky code, and prevent AVs from breaking laws in 5 out of 8 scenarios.
☆ From Fragment to One Piece: A Survey on AI-Driven Graphic Design
This survey provides a comprehensive overview of the advancements in Artificial Intelligence in Graphic Design (AIGD), focusing on integrating AI techniques to support design interpretation and enhance the creative process. We categorize the field into two primary directions: perception tasks, which involve understanding and analyzing design elements, and generation tasks, which focus on creating new design elements and layouts. The survey covers various subtasks, including visual element perception and generation, aesthetic and semantic understanding, layout analysis, and generation. We highlight the role of large language models and multimodal approaches in bridging the gap between localized visual features and global design intent. Despite significant progress, challenges remain to understanding human intent, ensuring interpretability, and maintaining control over multilayered compositions. This survey serves as a guide for researchers, providing information on the current state of AIGD and potential future directions\footnote{https://github.com/zhangtianer521/excellent\_Intelligent\_graphic\_design}.
☆ Towards Human-Understandable Multi-Dimensional Concept Discovery
Concept-based eXplainable AI (C-XAI) aims to overcome the limitations of traditional saliency maps by converting pixels into human-understandable concepts that are consistent across an entire dataset. A crucial aspect of C-XAI is completeness, which measures how well a set of concepts explains a model's decisions. Among C-XAI methods, Multi-Dimensional Concept Discovery (MCD) effectively improves completeness by breaking down the CNN latent space into distinct and interpretable concept subspaces. However, MCD's explanations can be difficult for humans to understand, raising concerns about their practical utility. To address this, we propose Human-Understandable Multi-dimensional Concept Discovery (HU-MCD). HU-MCD uses the Segment Anything Model for concept identification and implements a CNN-specific input masking technique to reduce noise introduced by traditional masking methods. These changes to MCD, paired with the completeness relation, enable HU-MCD to enhance concept understandability while maintaining explanation faithfulness. Our experiments, including human subject studies, show that HU-MCD provides more precise and reliable explanations than existing C-XAI methods. The code is available at https://github.com/grobruegge/hu-mcd.
☆ Dig2DIG: Dig into Diffusion Information Gains for Image Fusion
Image fusion integrates complementary information from multi-source images to generate more informative results. Recently, the diffusion model, which demonstrates unprecedented generative potential, has been explored in image fusion. However, these approaches typically incorporate predefined multimodal guidance into diffusion, failing to capture the dynamically changing significance of each modality, while lacking theoretical guarantees. To address this issue, we reveal a significant spatio-temporal imbalance in image denoising; specifically, the diffusion model produces dynamic information gains in different image regions with denoising steps. Based on this observation, we Dig into the Diffusion Information Gains (Dig2DIG) and theoretically derive a diffusion-based dynamic image fusion framework that provably reduces the upper bound of the generalization error. Accordingly, we introduce diffusion information gains (DIG) to quantify the information contribution of each modality at different denoising steps, thereby providing dynamic guidance during the fusion process. Extensive experiments on multiple fusion scenarios confirm that our method outperforms existing diffusion-based approaches in terms of both fusion quality and inference efficiency.
☆ Adventurer: Exploration with BiGAN for Deep Reinforcement Learning
Recent developments in deep reinforcement learning have been very successful in learning complex, previously intractable problems. Sample efficiency and local optimality, however, remain significant challenges. To address these challenges, novelty-driven exploration strategies have emerged and shown promising potential. Unfortunately, no single algorithm outperforms all others in all tasks and most of them struggle with tasks with high-dimensional and complex observations. In this work, we propose Adventurer, a novelty-driven exploration algorithm that is based on Bidirectional Generative Adversarial Networks (BiGAN), where BiGAN is trained to estimate state novelty. Intuitively, a generator that has been trained on the distribution of visited states should only be able to generate a state coming from the distribution of visited states. As a result, novel states using the generator to reconstruct input states from certain latent representations would lead to larger reconstruction errors. We show that BiGAN performs well in estimating state novelty for complex observations. This novelty estimation method can be combined with intrinsic-reward-based exploration. Our empirical results show that Adventurer produces competitive results on a range of popular benchmark tasks, including continuous robotic manipulation tasks (e.g. Mujoco robotics) and high-dimensional image-based tasks (e.g. Atari games).
comment: Accepted at Applied Intelligence
☆ Reinforcement Learning in Switching Non-Stationary Markov Decision Processes: Algorithms and Convergence Analysis
Reinforcement learning in non-stationary environments is challenging due to abrupt and unpredictable changes in dynamics, often causing traditional algorithms to fail to converge. However, in many real-world cases, non-stationarity has some structure that can be exploited to develop algorithms and facilitate theoretical analysis. We introduce one such structure, Switching Non-Stationary Markov Decision Processes (SNS-MDP), where environments switch over time based on an underlying Markov chain. Under a fixed policy, the value function of an SNS-MDP admits a closed-form solution determined by the Markov chain's statistical properties, and despite the inherent non-stationarity, Temporal Difference (TD) learning methods still converge to the correct value function. Furthermore, policy improvement can be performed, and it is shown that policy iteration converges to the optimal policy. Moreover, since Q-learning converges to the optimal Q-function, it likewise yields the corresponding optimal policy. To illustrate the practical advantages of SNS-MDPs, we present an example in communication networks where channel noise follows a Markovian pattern, demonstrating how this framework can effectively guide decision-making in complex, time-varying contexts.
☆ Adaptive Unimodal Regulation for Balanced Multimodal Information Acquisition CVPR2025
Sensory training during the early ages is vital for human development. Inspired by this cognitive phenomenon, we observe that the early training stage is also important for the multimodal learning process, where dataset information is rapidly acquired. We refer to this stage as the prime learning window. However, based on our observation, this prime learning window in multimodal learning is often dominated by information-sufficient modalities, which in turn suppresses the information acquisition of information-insufficient modalities. To address this issue, we propose Information Acquisition Regulation (InfoReg), a method designed to balance information acquisition among modalities. Specifically, InfoReg slows down the information acquisition process of information-sufficient modalities during the prime learning window, which could promote information acquisition of information-insufficient modalities. This regulation enables a more balanced learning process and improves the overall performance of the multimodal network. Experiments show that InfoReg outperforms related multimodal imbalanced methods across various datasets, achieving superior model performance. The code is available at https://github.com/GeWu-Lab/InfoReg_CVPR2025.
comment: 10pages, 16 figures, CVPR2025
☆ ClinText-SP and RigoBERTa Clinical: a new set of open resources for Spanish Clinical NLP
We present a novel contribution to Spanish clinical natural language processing by introducing the largest publicly available clinical corpus, ClinText-SP, along with a state-of-the-art clinical encoder language model, RigoBERTa Clinical. Our corpus was meticulously curated from diverse open sources, including clinical cases from medical journals and annotated corpora from shared tasks, providing a rich and diverse dataset that was previously difficult to access. RigoBERTa Clinical, developed through domain-adaptive pretraining on this comprehensive dataset, significantly outperforms existing models on multiple clinical NLP benchmarks. By publicly releasing both the dataset and the model, we aim to empower the research community with robust resources that can drive further advancements in clinical NLP and ultimately contribute to improved healthcare applications.
☆ The Role of Artificial Intelligence in Enhancing Insulin Recommendations and Therapy Outcomes
The growing worldwide incidence of diabetes requires more effective approaches for managing blood glucose levels. Insulin delivery systems have advanced significantly, with artificial intelligence (AI) playing a key role in improving their precision and adaptability. AI algorithms, particularly those based on reinforcement learning, allow for personalised insulin dosing by continuously adapting to an individual's responses. Despite these advancements, challenges such as data privacy, algorithm transparency, and accessibility still need to be addressed. Continued progress and validation in AI-driven insulin delivery systems promise to improve therapy outcomes further, offering people more effective and individualised management of their diabetes. This paper presents an overview of current strategies, key challenges, and future directions.
☆ Galaxy Walker: Geometry-aware VLMs For Galaxy-scale Understanding
Modern vision-language models (VLMs) develop patch embedding and convolution backbone within vector space, especially Euclidean ones, at the very founding. When expanding VLMs to a galaxy scale for understanding astronomical phenomena, the integration of spherical space for planetary orbits and hyperbolic spaces for black holes raises two formidable challenges. a) The current pre-training model is confined to Euclidean space rather than a comprehensive geometric embedding. b) The predominant architecture lacks suitable backbones for anisotropic physical geometries. In this paper, we introduced Galaxy-Walker, a geometry-aware VLM, for the universe-level vision understanding tasks. We proposed the geometry prompt that generates geometry tokens by random walks across diverse spaces on a multi-scale physical graph, along with a geometry adapter that compresses and reshapes the space anisotropy in a mixture-of-experts manner. Extensive experiments demonstrate the effectiveness of our approach, with Galaxy-Walker achieving state-of-the-art performance in both galaxy property estimation ($R^2$ scores up to $0.91$) and morphology classification tasks (up to $+0.17$ F1 improvement in challenging features), significantly outperforming both domain-specific models and general-purpose VLMs.
☆ Identifying and Characterising Higher Order Interactions in Mobility Networks Using Hypergraphs
Understanding human mobility is essential for applications ranging from urban planning to public health. Traditional mobility models such as flow networks and colocation matrices capture only pairwise interactions between discrete locations, overlooking higher-order relationships among locations (i.e., mobility flow among two or more locations). To address this, we propose co-visitation hypergraphs, a model that leverages temporal observation windows to extract group interactions between locations from individual mobility trajectory data. Using frequent pattern mining, our approach constructs hypergraphs that capture dynamic mobility behaviors across different spatial and temporal scales. We validate our method on a publicly available mobility dataset and demonstrate its effectiveness in analyzing city-scale mobility patterns, detecting shifts during external disruptions such as extreme weather events, and examining how a location's connectivity (degree) relates to the number of points of interest (POIs) within it. Our results demonstrate that our hypergraph-based mobility analysis framework is a valuable tool with potential applications in diverse fields such as public health, disaster resilience, and urban planning.
☆ Anchor-based oversampling for imbalanced tabular data via contrastive and adversarial learning
Imbalanced data represent a distribution with more frequencies of one class (majority) than the other (minority). This phenomenon occurs across various domains, such as security, medical care and human activity. In imbalanced learning, classification algorithms are typically inclined to classify the majority class accurately, resulting in artificially high accuracy rates. As a result, many minority samples are mistakenly labelled as majority-class instances, resulting in a bias that benefits the majority class. This study presents a framework based on boundary anchor samples to tackle the imbalance learning challenge. First, we select and use anchor samples to train a multilayer perceptron (MLP) classifier, which acts as a prior knowledge model and aids the adversarial and contrastive learning procedures. Then, we designed a novel deep generative model called Anchor Stabilized Conditional Generative Adversarial Network or Anch-SCGAN in short. Anch-SCGAN is supported with two generators for the minority and majority classes and a discriminator incorporating additional class-specific information from the pre-trained feature extractor MLP. In addition, we facilitate the generator's training procedure in two ways. First, we define a new generator loss function based on reprocessed anchor samples and contrastive learning. Second, we apply a scoring strategy to stabilize the adversarial training part in generators. We train Anch-SCGAN and further finetune it with anchor samples to improve the precision of the generated samples. Our experiments on 16 real-world imbalanced datasets illustrate that Anch-SCGAN outperforms the renowned methods in imbalanced learning.
☆ Distil-xLSTM: Learning Attention Mechanisms through Recurrent Structures
The current era of Natural Language Processing (NLP) is dominated by Transformer models. However, novel architectures relying on recurrent mechanisms, such as xLSTM and Mamba, have been proposed as alternatives to attention-based models. Although computation is done differently than with the attention mechanism mechanism, these recurrent models yield good results and sometimes even outperform state-of-the-art attention-based models. In this work, we propose Distil-xLSTM, an xLSTM-based Small Language Model (SLM) trained by distilling knowledge from a Large Language Model (LLM) that shows promising results while being compute and scale efficient. Our Distil-xLSTM focuses on approximating a transformer-based model attention parametrization using its recurrent sequence mixing components and shows good results with minimal training.
☆ Self-Reported Confidence of Large Language Models in Gastroenterology: Analysis of Commercial, Open-Source, and Quantized Models
This study evaluated self-reported response certainty across several large language models (GPT, Claude, Llama, Phi, Mistral, Gemini, Gemma, and Qwen) using 300 gastroenterology board-style questions. The highest-performing models (GPT-o1 preview, GPT-4o, and Claude-3.5-Sonnet) achieved Brier scores of 0.15-0.2 and AUROC of 0.6. Although newer models demonstrated improved performance, all exhibited a consistent tendency towards overconfidence. Uncertainty estimation presents a significant challenge to the safe use of LLMs in healthcare. Keywords: Large Language Models; Confidence Elicitation; Artificial Intelligence; Gastroenterology; Uncertainty Quantification
comment: 35 pages, 5 figures, 1 table, 7 supplementary figures
☆ EvAnimate: Event-conditioned Image-to-Video Generation for Human Animation
Conditional human animation transforms a static reference image into a dynamic sequence by applying motion cues such as poses. These motion cues are typically derived from video data but are susceptible to limitations including low temporal resolution, motion blur, overexposure, and inaccuracies under low-light conditions. In contrast, event cameras provide data streams with exceptionally high temporal resolution, a wide dynamic range, and inherent resistance to motion blur and exposure issues. In this work, we propose EvAnimate, a framework that leverages event streams as motion cues to animate static human images. Our approach employs a specialized event representation that transforms asynchronous event streams into 3-channel slices with controllable slicing rates and appropriate slice density, ensuring compatibility with diffusion models. Subsequently, a dual-branch architecture generates high-quality videos by harnessing the inherent motion dynamics of the event streams, thereby enhancing both video quality and temporal consistency. Specialized data augmentation strategies further enhance cross-person generalization. Finally, we establish a new benchmarking, including simulated event data for training and validation, and a real-world event dataset capturing human actions under normal and extreme scenarios. The experiment results demonstrate that EvAnimate achieves high temporal fidelity and robust performance in scenarios where traditional video-derived cues fall short.
☆ Discriminative protein sequence modelling with Latent Space Diffusion
We explore a framework for protein sequence representation learning that decomposes the task between manifold learning and distributional modelling. Specifically we present a Latent Space Diffusion architecture which combines a protein sequence autoencoder with a denoising diffusion model operating on its latent space. We obtain a one-parameter family of learned representations from the diffusion model, along with the autoencoder's latent representation. We propose and evaluate two autoencoder architectures: a homogeneous model forcing amino acids of the same type to be identically distributed in the latent space, and an inhomogeneous model employing a noise-based variant of masking. As a baseline we take a latent space learned by masked language modelling, and evaluate discriminative capability on a range of protein property prediction tasks. Our finding is twofold: the diffusion models trained on both our proposed variants display higher discriminative power than the one trained on the masked language model baseline, none of the diffusion representations achieve the performance of the masked language model embeddings themselves.
☆ RLCAD: Reinforcement Learning Training Gym for Revolution Involved CAD Command Sequence Generation
A CAD command sequence is a typical parametric design paradigm in 3D CAD systems where a model is constructed by overlaying 2D sketches with operations such as extrusion, revolution, and Boolean operations. Although there is growing academic interest in the automatic generation of command sequences, existing methods and datasets only support operations such as 2D sketching, extrusion,and Boolean operations. This limitation makes it challenging to represent more complex geometries. In this paper, we present a reinforcement learning (RL) training environment (gym) built on a CAD geometric engine. Given an input boundary representation (B-Rep) geometry, the policy network in the RL algorithm generates an action. This action, along with previously generated actions, is processed within the gym to produce the corresponding CAD geometry, which is then fed back into the policy network. The rewards, determined by the difference between the generated and target geometries within the gym, are used to update the RL network. Our method supports operations beyond sketches, Boolean, and extrusion, including revolution operations. With this training gym, we achieve state-of-the-art (SOTA) quality in generating command sequences from B-Rep geometries. In addition, our method can significantly improve the efficiency of command sequence generation by a factor of 39X compared with the previous training gym.
☆ An Identity and Interaction Based Network Forensic Analysis
In todays landscape of increasing electronic crime, network forensics plays a pivotal role in digital investigations. It aids in understanding which systems to analyse and as a supplement to support evidence found through more traditional computer based investigations. However, the nature and functionality of the existing Network Forensic Analysis Tools (NFATs) fall short compared to File System Forensic Analysis Tools (FS FATs) in providing usable data. The analysis tends to focus upon IP addresses, which are not synonymous with user identities, a point of significant interest to investigators. This paper presents several experiments designed to create a novel NFAT approach that can identify users and understand how they are using network based applications whilst the traffic remains encrypted. The experiments build upon the prior art and investigate how effective this approach is in classifying users and their actions. Utilising an in-house dataset composed of 50 million packers, the experiments are formed of three incremental developments that assist in improving performance. Building upon the successful experiments, a proposed NFAT interface is presented to illustrate the ease at which investigators would be able to ask relevant questions of user interactions. The experiments profiled across 27 users, has yielded an average 93.3% True Positive Identification Rate (TPIR), with 41% of users experiencing 100% TPIR. Skype, Wikipedia and Hotmail services achieved a notably high level of recognition performance. The study has developed and evaluated an approach to analyse encrypted network traffic more effectively through the modelling of network traffic and to subsequently visualise these interactions through a novel network forensic analysis tool.
☆ UniPCGC: Towards Practical Point Cloud Geometry Compression via an Efficient Unified Approach AAAI 2025
Learning-based point cloud compression methods have made significant progress in terms of performance. However, these methods still encounter challenges including high complexity, limited compression modes, and a lack of support for variable rate, which restrict the practical application of these methods. In order to promote the development of practical point cloud compression, we propose an efficient unified point cloud geometry compression framework, dubbed as UniPCGC. It is a lightweight framework that supports lossy compression, lossless compression, variable rate and variable complexity. First, we introduce the Uneven 8-Stage Lossless Coder (UELC) in the lossless mode, which allocates more computational complexity to groups with higher coding difficulty, and merges groups with lower coding difficulty. Second, Variable Rate and Complexity Module (VRCM) is achieved in the lossy mode through joint adoption of a rate modulation module and dynamic sparse convolution. Finally, through the dynamic combination of UELC and VRCM, we achieve lossy compression, lossless compression, variable rate and complexity within a unified framework. Compared to the previous state-of-the-art method, our method achieves a compression ratio (CR) gain of 8.1\% on lossless compression, and a Bjontegaard Delta Rate (BD-Rate) gain of 14.02\% on lossy compression, while also supporting variable rate and variable complexity.
comment: Accepted to AAAI 2025
☆ HiRes-FusedMIM: A High-Resolution RGB-DSM Pre-trained Model for Building-Level Remote Sensing Applications
Recent advances in self-supervised learning have led to the development of foundation models that have significantly advanced performance in various computer vision tasks. However, despite their potential, these models often overlook the crucial role of high-resolution digital surface models (DSMs) in understanding urban environments, particularly for building-level analysis, which is essential for applications like digital twins. To address this gap, we introduce HiRes-FusedMIM, a novel pre-trained model specifically designed to leverage the rich information contained within high-resolution RGB and DSM data. HiRes-FusedMIM utilizes a dual-encoder simple masked image modeling (SimMIM) architecture with a multi-objective loss function that combines reconstruction and contrastive objectives, enabling it to learn powerful, joint representations from both modalities. We conducted a comprehensive evaluation of HiRes-FusedMIM on a diverse set of downstream tasks, including classification, semantic segmentation, and instance segmentation. Our results demonstrate that: 1) HiRes-FusedMIM outperforms previous state-of-the-art geospatial methods on several building-related datasets, including WHU Aerial and LoveDA, demonstrating its effectiveness in capturing and leveraging fine-grained building information; 2) Incorporating DSMs during pre-training consistently improves performance compared to using RGB data alone, highlighting the value of elevation information for building-level analysis; 3) The dual-encoder architecture of HiRes-FusedMIM, with separate encoders for RGB and DSM data, significantly outperforms a single-encoder model on the Vaihingen segmentation task, indicating the benefits of learning specialized representations for each modality. To facilitate further research and applications in this direction, we will publicly release the trained model weights.
☆ Natural Language Processing for Electronic Health Records in Scandinavian Languages: Norwegian, Swedish, and Danish
Background: Clinical natural language processing (NLP) refers to the use of computational methods for extracting, processing, and analyzing unstructured clinical text data, and holds a huge potential to transform healthcare in various clinical tasks. Objective: The study aims to perform a systematic review to comprehensively assess and analyze the state-of-the-art NLP methods for the mainland Scandinavian clinical text. Method: A literature search was conducted in various online databases including PubMed, ScienceDirect, Google Scholar, ACM digital library, and IEEE Xplore between December 2022 and February 2024. Further, relevant references to the included articles were also used to solidify our search. The final pool includes articles that conducted clinical NLP in the mainland Scandinavian languages and were published in English between 2010 and 2024. Results: Out of the 113 articles, 18% (n=21) focus on Norwegian clinical text, 64% (n=72) on Swedish, 10% (n=11) on Danish, and 8% (n=9) focus on more than one language. Generally, the review identified positive developments across the region despite some observable gaps and disparities between the languages. There are substantial disparities in the level of adoption of transformer-based models. In essential tasks such as de-identification, there is significantly less research activity focusing on Norwegian and Danish compared to Swedish text. Further, the review identified a low level of sharing resources such as data, experimentation code, pre-trained models, and rate of adaptation and transfer learning in the region. Conclusion: The review presented a comprehensive assessment of the state-of-the-art Clinical NLP for electronic health records (EHR) text in mainland Scandinavian languages and, highlighted the potential barriers and challenges that hinder the rapid advancement of the field in the region.
comment: 45 pages including the appendix, 9 figures in the main manuscript and 11 figures in the Appendix
☆ MMCR: Advancing Visual Language Model in Multimodal Multi-Turn Contextual Reasoning
Compared to single-turn dialogue, multi-turn dialogue involving multiple images better aligns with the needs of real-world human-AI interactions. Additionally, as training data, it provides richer contextual reasoning information, thereby guiding the model to achieve better performance. However, existing vision-language models (VLMs) primarily rely on single-turn dialogue training and evaluation benchmarks. In this paper, following the characteristics of human dialogue, such as focused topics and concise, clear content, we present MMCR (Multimodal Multi-turn Contextual Reasoning), a novel dataset comprising: (1) MMCR-310k -- the largest multi-image multi-turn instruction tuning dataset with 310K contextual dialogues, each covering 1-4 images and 4 or 8 dialogue turns; and (2) MMCR-Bench -- a diagnostic benchmark featuring dialogues, spanning 8 domains (Humanities, Natural, Science, Education, etc.) and 40 sub-topics. Extensive evaluations demonstrate that models fine-tuned with MMCR-310k achieve 5.2\% higher contextual accuracy on MMCR-Bench, while showing consistent improvements on existing benchmarks (+1.1\% on AI2D, +1.2\% on MMMU and MMVet). MMCR and prompt engineering will be released publicly.
☆ SciClaims: An End-to-End Generative System for Biomedical Claim Analysis
Validating key claims in scientific literature, particularly in biomedical research, is essential for ensuring accuracy and advancing knowledge. This process is critical in sectors like the pharmaceutical industry, where rapid scientific progress requires automation and deep domain expertise. However, current solutions have significant limitations. They lack end-to-end pipelines encompassing all claim extraction, evidence retrieval, and verification steps; rely on complex NLP and information retrieval pipelines prone to multiple failure points; and often fail to provide clear, user-friendly justifications for claim verification outcomes. To address these challenges, we introduce SciClaims, an advanced system powered by state-of-the-art large language models (LLMs) that seamlessly integrates the entire scientific claim analysis process. SciClaims outperforms previous approaches in both claim extraction and verification without requiring additional fine-tuning, setting a new benchmark for automated scientific claim analysis.
comment: Pre-print version
☆ Neuro-symbolic Weak Supervision: Theory and Semantics
Weak supervision allows machine learning models to learn from limited or noisy labels, but it introduces challenges in interpretability and reliability - particularly in multi-instance partial label learning (MI-PLL), where models must resolve both ambiguous labels and uncertain instance-label mappings. We propose a semantics for neuro-symbolic framework that integrates Inductive Logic Programming (ILP) to improve MI-PLL by providing structured relational constraints that guide learning. Within our semantic characterization, ILP defines a logical hypothesis space for label transitions, clarifies classifier semantics, and establishes interpretable performance standards. This hybrid approach improves robustness, transparency, and accountability in weakly supervised settings, ensuring neural predictions align with domain knowledge. By embedding weak supervision into a logical framework, we enhance both interpretability and learning, making weak supervision more suitable for real-world, high-stakes applications.
☆ Statistically Testing Training Data for Unwanted Error Patterns using Rule-Oriented Regression
Artificial intelligence models trained from data can only be as good as the underlying data is. Biases in training data propagating through to the output of a machine learning model are a well-documented and well-understood phenomenon, but the machinery to prevent these undesired effects is much less developed. Efforts to ensure data is clean during collection, such as using bias-aware sampling, are most effective when the entity controlling data collection also trains the AI. In cases where the data is already available, how do we find out if the data was already manipulated, i.e., ``poisoned'', so that an undesired behavior would be trained into a machine learning model? This is a challenge fundamentally different to (just) improving approximation accuracy or efficiency, and we provide a method to test training data for flaws, to establish a trustworthy ground-truth for a subsequent training of machine learning models (of any kind). Unlike the well-studied problem of approximating data using fuzzy rules that are generated from the data, our method hinges on a prior definition of rules to happen before seeing the data to be tested. Therefore, the proposed method can also discover hidden error patterns, which may also have substantial influence. Our approach extends the abilities of conventional statistical testing by letting the ``test-condition'' be any Boolean condition to describe a pattern in the data, whose presence we wish to determine. The method puts fuzzy inference into a regression model, to get the best of the two: explainability from fuzzy logic with statistical properties and diagnostics from the regression, and finally also being applicable to ``small data'', hence not requiring large datasets as deep learning methods do. We provide an open source implementation for demonstration and experiments.
☆ Verbal Process Supervision Elicits Better Coding Agents
The emergence of large language models and their applications as AI agents have significantly advanced state-of-the-art code generation benchmarks, transforming modern software engineering tasks. However, even with test-time computed reasoning models, these systems still struggle with complex software engineering challenges. This work introduces CURA, a code understanding and reasoning agent system enhanced with verbal process supervision (VPS), achieving a 3.65\% improvement over baseline models on challenging benchmarks like BigCodeBench. Furthermore, CURA, when paired with the o3-mini model and VPS techniques, attains state-of-the-art performance. This work represents a step forward in integrating reasoning-driven architectures with LLM-based code generation, enabling agentic reasoning for language models to solve complex software engineering tasks.
☆ Safeguarding Mobile GUI Agent via Logic-based Action Verification
Large Foundation Models (LFMs) have unlocked new possibilities in human-computer interaction, particularly with the rise of mobile Graphical User Interface (GUI) Agents capable of interpreting GUIs. These agents promise to revolutionize mobile computing by allowing users to automate complex mobile tasks through simple natural language instructions. However, the inherent probabilistic nature of LFMs, coupled with the ambiguity and context-dependence of mobile tasks, makes LFM-based automation unreliable and prone to errors. To address this critical challenge, we introduce VeriSafe Agent (VSA): a formal verification system that serves as a logically grounded safeguard for Mobile GUI Agents. VSA is designed to deterministically ensure that an agent's actions strictly align with user intent before conducting an action. At its core, VSA introduces a novel autoformalization technique that translates natural language user instructions into a formally verifiable specification, expressed in our domain-specific language (DSL). This enables runtime, rule-based verification, allowing VSA to detect and prevent erroneous actions executing an action, either by providing corrective feedback or halting unsafe behavior. To the best of our knowledge, VSA is the first attempt to bring the rigor of formal verification to GUI agent. effectively bridging the gap between LFM-driven automation and formal software verification. We implement VSA using off-the-shelf LLM services (GPT-4o) and evaluate its performance on 300 user instructions across 18 widely used mobile apps. The results demonstrate that VSA achieves 94.3%-98.33% accuracy in verifying agent actions, representing a significant 20.4%-25.6% improvement over existing LLM-based verification methods, and consequently increases the GUI agent's task completion rate by 90%-130%.
☆ Large Language Models powered Network Attack Detection: Architecture, Opportunities and Case Study
Network attack detection is a pivotal technology to identify network anomaly and classify malicious traffic. Large Language Models (LLMs) are trained on a vast corpus of text, have amassed remarkable capabilities of context-understanding and commonsense knowledge. This has opened up a new door for network threat detection. Researchers have already initiated discussions regarding the application of LLMs on specific cyber-security tasks. Unfortunately, there is still a lack of comprehensive elaboration how to mine LLMs' potentials in network threat detections, as well as the opportunities and challenges. In this paper, we mainly focus on the classification of malicious traffic from the perspective of LLMs' capability. We present a holistic view of the architecture of LLM-powered network attack detection, including Pre-training, Fine-tuning, and Detection. Especially, by exploring the knowledge and capabilities of LLM, we identify three distinct roles LLM can act in network attack detection: \textit{Classifier, Encoder, and Predictor}. For each of them, the modeling paradigm, opportunities and challenges are elaborated. Finally, we present our design on LLM-powered DDoS detection as a case study. The proposed framework attains accurate detection on carpet bombing DDoS by exploiting LLMs' capabilities in contextual mining. The evaluation shows its efficacy, exhibiting a nearly $35$\% improvement compared to existing systems.
comment: submitted for peer-review
☆ Words as Bridges: Exploring Computational Support for Cross-Disciplinary Translation Work
Scholars often explore literature outside of their home community of study. This exploration process is frequently hampered by field-specific jargon. Past computational work often focuses on supporting translation work by removing jargon through simplification and summarization; here, we explore a different approach that preserves jargon as useful bridges to new conceptual spaces. Specifically, we cast different scholarly domains as different language-using communities, and explore how to adapt techniques from unsupervised cross-lingual alignment of word embeddings to explore conceptual alignments between domain-specific word embedding spaces.We developed a prototype cross-domain search engine that uses aligned domain-specific embeddings to support conceptual exploration, and tested this prototype in two case studies. We discuss qualitative insights into the promises and pitfalls of this approach to translation work, and suggest design insights for future interfaces that provide computational support for cross-domain information seeking.
comment: 26 pages, 8 tables, 6 figures
☆ MetaSpatial: Reinforcing 3D Spatial Reasoning in VLMs for the Metaverse
We present MetaSpatial, the first reinforcement learning (RL)-based framework designed to enhance 3D spatial reasoning in vision-language models (VLMs), enabling real-time 3D scene generation without the need for hard-coded optimizations. MetaSpatial addresses two core challenges: (i) the lack of internalized 3D spatial reasoning in VLMs, which limits their ability to generate realistic layouts, and (ii) the inefficiency of traditional supervised fine-tuning (SFT) for layout generation tasks, as perfect ground truth annotations are unavailable. Our key innovation is a multi-turn RL-based optimization mechanism that integrates physics-aware constraints and rendered image evaluations, ensuring generated 3D layouts are coherent, physically plausible, and aesthetically consistent. Methodologically, MetaSpatial introduces an adaptive, iterative reasoning process, where the VLM refines spatial arrangements over multiple turns by analyzing rendered outputs, improving scene coherence progressively. Empirical evaluations demonstrate that MetaSpatial significantly enhances the spatial consistency and formatting stability of various scale models. Post-training, object placements are more realistic, aligned, and functionally coherent, validating the effectiveness of RL for 3D spatial reasoning in metaverse, AR/VR, digital twins, and game development applications. Our code, data, and training pipeline are publicly available at https://github.com/PzySeere/MetaSpatial.
comment: Working Paper
☆ PALATE: Peculiar Application of the Law of Total Expectation to Enhance the Evaluation of Deep Generative Models
Deep generative models (DGMs) have caused a paradigm shift in the field of machine learning, yielding noteworthy advancements in domains such as image synthesis, natural language processing, and other related areas. However, a comprehensive evaluation of these models that accounts for the trichotomy between fidelity, diversity, and novelty in generated samples remains a formidable challenge. A recently introduced solution that has emerged as a promising approach in this regard is the Feature Likelihood Divergence (FLD), a method that offers a theoretically motivated practical tool, yet also exhibits some computational challenges. In this paper, we propose PALATE, a novel enhancement to the evaluation of DGMs that addresses limitations of existing metrics. Our approach is based on a peculiar application of the law of total expectation to random variables representing accessible real data. When combined with the MMD baseline metric and DINOv2 feature extractor, PALATE offers a holistic evaluation framework that matches or surpasses state-of-the-art solutions while providing superior computational efficiency and scalability to large-scale datasets. Through a series of experiments, we demonstrate the effectiveness of the PALATE enhancement, contributing a computationally efficient, holistic evaluation approach that advances the field of DGMs assessment, especially in detecting sample memorization and evaluating generalization capabilities.
☆ ModiGen: A Large Language Model-Based Workflow for Multi-Task Modelica Code Generation
Modelica is a widely adopted language for simulating complex physical systems, yet effective model creation and optimization require substantial domain expertise. Although large language models (LLMs) have demonstrated promising capabilities in code generation, their application to modeling remains largely unexplored. To address this gap, we have developed benchmark datasets specifically designed to evaluate the performance of LLMs in generating Modelica component models and test cases. Our evaluation reveals substantial limitations in current LLMs, as the generated code often fails to simulate successfully. To overcome these challenges, we propose a specialized workflow that integrates supervised fine-tuning, graph retrieval-augmented generation, and feedback optimization to improve the accuracy and reliability of Modelica code generation. The evaluation results demonstrate significant performance gains: the maximum improvement in pass@1 reached 0.3349 for the component generation task and 0.2457 for the test case generation task. This research underscores the potential of LLMs to advance intelligent modeling tools and offers valuable insights for future developments in system modeling and engineering applications.
☆ Teaching LLMs for Step-Level Automatic Math Correction via Reinforcement Learning
Automatic math correction aims to check students' solutions to mathematical problems via artificial intelligence technologies. Most existing studies focus on judging the final answer at the problem level, while they ignore detailed feedback on each step in a math problem-solving process, which requires abilities of semantic understanding and reasoning. In this paper, we propose a reinforcement learning (RL)-based method to boost large language model (LLM) for step-level automatic math correction, named StepAMC. Particularly, we convert the step-level automatic math correction within the text classification task into an RL problem to enhance the reasoning capabilities of LLMs. Then, we design a space-constrained policy network to improve the stability of RL. Then, we introduce a fine-grained reward network to convert the binary human feedback into a continuous value. We conduct extensive experiments over two benchmark datasets and the results show that our model outperforms the eleven strong baselines.
☆ Generative AI in Knowledge Work: Design Implications for Data Navigation and Decision-Making
Our study of 20 knowledge workers revealed a common challenge: the difficulty of synthesizing unstructured information scattered across multiple platforms to make informed decisions. Drawing on their vision of an ideal knowledge synthesis tool, we developed Yodeai, an AI-enabled system, to explore both the opportunities and limitations of AI in knowledge work. Through a user study with 16 product managers, we identified three key requirements for Generative AI in knowledge work: adaptable user control, transparent collaboration mechanisms, and the ability to integrate background knowledge with external information. However, we also found significant limitations, including overreliance on AI, user isolation, and contextual factors outside the AI's reach. As AI tools become increasingly prevalent in professional settings, we propose design principles that emphasize adaptability to diverse workflows, accountability in personal and collaborative contexts, and context-aware interoperability to guide the development of human-centered AI systems for product managers and knowledge workers.
comment: Accepted to CHI '25 (Conference on Human Factors in Computing Systems), to appear April 26-May 1, 2025, Yokohama, Japan
☆ Knowledge Graph Enhanced Generative Multi-modal Models for Class-Incremental Learning
Continual learning in computer vision faces the critical challenge of catastrophic forgetting, where models struggle to retain prior knowledge while adapting to new tasks. Although recent studies have attempted to leverage the generalization capabilities of pre-trained models to mitigate overfitting on current tasks, models still tend to forget details of previously learned categories as tasks progress, leading to misclassification. To address these limitations, we introduce a novel Knowledge Graph Enhanced Generative Multi-modal model (KG-GMM) that builds an evolving knowledge graph throughout the learning process. Our approach utilizes relationships within the knowledge graph to augment the class labels and assigns different relations to similar categories to enhance model differentiation. During testing, we propose a Knowledge Graph Augmented Inference method that locates specific categories by analyzing relationships within the generated text, thereby reducing the loss of detailed information about old classes when learning new knowledge and alleviating forgetting. Experiments demonstrate that our method effectively leverages relational information to help the model correct mispredictions, achieving state-of-the-art results in both conventional CIL and few-shot CIL settings, confirming the efficacy of knowledge graphs at preserving knowledge in the continual learning scenarios.
☆ PRECTR: A Synergistic Framework for Integrating Personalized Search Relevance Matching and CTR Prediction
The two primary tasks in the search recommendation system are search relevance matching and click-through rate (CTR) prediction -- the former focuses on seeking relevant items for user queries whereas the latter forecasts which item may better match user interest. Prior research typically develops two models to predict the CTR and search relevance separately, then ranking candidate items based on the fusion of the two outputs. However, such a divide-and-conquer paradigm creates the inconsistency between different models. Meanwhile, the search relevance model mainly concentrates on the degree of objective text matching while neglecting personalized differences among different users, leading to restricted model performance. To tackle these issues, we propose a unified \textbf{P}ersonalized Search RElevance Matching and CTR Prediction Fusion Model(PRECTR). Specifically, based on the conditional probability fusion mechanism, PRECTR integrates the CTR prediction and search relevance matching into one framework to enhance the interaction and consistency of the two modules. However, directly optimizing CTR binary classification loss may bring challenges to the fusion model's convergence and indefinitely promote the exposure of items with high CTR, regardless of their search relevance. Hence, we further introduce two-stage training and semantic consistency regularization to accelerate the model's convergence and restrain the recommendation of irrelevant items. Finally, acknowledging that different users may have varied relevance preferences, we assessed current users' relevance preferences by analyzing past users' preferences for similar queries and tailored incentives for different candidate items accordingly. Extensive experimental results on our production dataset and online A/B testing demonstrate the effectiveness and superiority of our proposed PRECTR method.
☆ Manipulation and the AI Act: Large Language Model Chatbots and the Danger of Mirrors
Large Language Model chatbots are increasingly taking the form and visage of human beings, adapting human faces, names, voices, personalities, and quirks, including those of celebrities and well-known political figures. Personifying AI chatbots could foreseeably increase their trust with users. However, it could also make them more capable of manipulation, by creating the illusion of a close and intimate relationship with an artificial entity. The European Commission has finalized the AI Act, with the EU Parliament making amendments banning manipulative and deceptive AI systems that cause significant harm to users. Although the AI Act covers harms that accumulate over time, it is unlikely to prevent harms associated with prolonged discussions with AI chatbots. Specifically, a chatbot could reinforce a person's negative emotional state over weeks, months, or years through negative feedback loops, prolonged conversations, or harmful recommendations, contributing to a user's deteriorating mental health.
☆ Resource-Efficient Motion Control for Video Generation via Dynamic Mask Guidance
Recent advances in diffusion models bring new vitality to visual content creation. However, current text-to-video generation models still face significant challenges such as high training costs, substantial data requirements, and difficulties in maintaining consistency between given text and motion of the foreground object. To address these challenges, we propose mask-guided video generation, which can control video generation through mask motion sequences, while requiring limited training data. Our model enhances existing architectures by incorporating foreground masks for precise text-position matching and motion trajectory control. Through mask motion sequences, we guide the video generation process to maintain consistent foreground objects throughout the sequence. Additionally, through a first-frame sharing strategy and autoregressive extension approach, we achieve more stable and longer video generation. Extensive qualitative and quantitative experiments demonstrate that this approach excels in various video generation tasks, such as video editing and generating artistic videos, outperforming previous methods in terms of consistency and quality. Our generated results can be viewed in the supplementary materials.
☆ RoCA: Robust Contrastive One-class Time Series Anomaly Detection with Contaminated Data
The accumulation of time-series signals and the absence of labels make time-series Anomaly Detection (AD) a self-supervised task of deep learning. Methods based on normality assumptions face the following three limitations: (1) A single assumption could hardly characterize the whole normality or lead to some deviation. (2) Some assumptions may go against the principle of AD. (3) Their basic assumption is that the training data is uncontaminated (free of anomalies), which is unrealistic in practice, leading to a decline in robustness. This paper proposes a novel robust approach, RoCA, which is the first to address all of the above three challenges, as far as we are aware. It fuses the separated assumptions of one-class classification and contrastive learning in a single training process to characterize a more complete so-called normality. Additionally, it monitors the training data and computes a carefully designed anomaly score throughout the training process. This score helps identify latent anomalies, which are then used to define the classification boundary, inspired by the concept of outlier exposure. The performance on AIOps datasets improved by 6% compared to when contamination was not considered (COCA). On two large and high-dimensional multivariate datasets, the performance increased by 5% to 10%. RoCA achieves the highest average performance on both univariate and multivariate datasets. The source code is available at https://github.com/ruiking04/RoCA.
☆ PP-FormulaNet: Bridging Accuracy and Efficiency in Advanced Formula Recognition
Formula recognition is an important task in document intelligence. It involves converting mathematical expressions from document images into structured symbolic formats that computers can easily work with. LaTeX is the most common format used for this purpose. In this work, we present PP-FormulaNet, a state-of-the-art formula recognition model that excels in both accuracy and efficiency. To meet the diverse needs of applications, we have developed two specialized models: PP-FormulaNet-L, tailored for high-accuracy scenarios, and PP-FormulaNet-S, optimized for high-efficiency contexts. Our extensive evaluations reveal that PP-FormulaNet-L attains accuracy levels that surpass those of prominent models such as UniMERNet by a significant 6%. Conversely, PP-FormulaNet-S operates at speeds that are over 16 times faster. These advancements facilitate seamless integration of PP-FormulaNet into a broad spectrum of document processing environments that involve intricate mathematical formulas. Furthermore, we introduce a Formula Mining System, which is capable of extracting a vast amount of high-quality formula data. This system further enhances the robustness and applicability of our formula recognition model. Code and models are publicly available at PaddleOCR(https://github.com/PaddlePaddle/PaddleOCR) and PaddleX(https://github.com/PaddlePaddle/PaddleX).
☆ Maximum Redundancy Pruning: A Principle-Driven Layerwise Sparsity Allocation for LLMs
Large language models (LLMs) have demonstrated impressive capabilities, but their enormous size poses significant challenges for deployment in real-world applications. To address this issue, researchers have sought to apply network pruning techniques to LLMs. A critical challenge in pruning is allocation the sparsity for each layer. Recent sparsity allocation methods is often based on heuristics or search that can easily lead to suboptimal performance. In this paper, we conducted an extensive investigation into various LLMs and revealed three significant discoveries: (1) the layerwise pruning sensitivity (LPS) of LLMs is highly non-uniform, (2) the choice of pruning metric affects LPS, and (3) the performance of a sparse model is related to the uniformity of its layerwise redundancy level. Based on these observations, we propose that the layerwise sparsity of LLMs should adhere to three principles: \emph{non-uniformity}, \emph{pruning metric dependency}, and \emph{uniform layerwise redundancy level} in the pruned model. To this end, we proposed Maximum Redundancy Pruning (MRP), an iterative pruning algorithm that prunes in the most redundant layers (\emph{i.e.}, those with the highest non-outlier ratio) at each iteration. The achieved layerwise sparsity aligns with the outlined principles. We conducted extensive experiments on publicly available LLMs, including the LLaMA2 and OPT, across various benchmarks. Experimental results validate the effectiveness of MRP, demonstrating its superiority over previous methods.
☆ Latent Embedding Adaptation for Human Preference Alignment in Diffusion Planners
This work addresses the challenge of personalizing trajectories generated in automated decision-making systems by introducing a resource-efficient approach that enables rapid adaptation to individual users' preferences. Our method leverages a pretrained conditional diffusion model with Preference Latent Embeddings (PLE), trained on a large, reward-free offline dataset. The PLE serves as a compact representation for capturing specific user preferences. By adapting the pretrained model using our proposed preference inversion method, which directly optimizes the learnable PLE, we achieve superior alignment with human preferences compared to existing solutions like Reinforcement Learning from Human Feedback (RLHF) and Low-Rank Adaptation (LoRA). To better reflect practical applications, we create a benchmark experiment using real human preferences on diverse, high-reward trajectories.
comment: 8 pages
☆ Optimizing Influence Campaigns: Nudging under Bounded Confidence
Influence campaigns in online social networks are often run by organizations, political parties, and nation states to influence large audiences. These campaigns are employed through the use of agents in the network that share persuasive content. Yet, their impact might be minimal if the audiences remain unswayed, often due to the bounded confidence phenomenon, where only a narrow spectrum of viewpoints can influence them. Here we show that to persuade under bounded confidence, an agent must nudge its targets to gradually shift their opinions. Using a control theory approach, we show how to construct an agent's nudging policy under the bounded confidence opinion dynamics model and also how to select targets for multiple agents in an influence campaign on a social network. Simulations on real Twitter networks show that a multi-agent nudging policy can shift the mean opinion, decrease opinion polarization, or even increase it. We find that our nudging based policies outperform other common techniques that do not consider the bounded confidence effect. Finally, we show how to craft prompts for large language models, such as ChatGPT, to generate text-based content for real nudging policies. This illustrates the practical feasibility of our approach, allowing one to go from mathematical nudging policies to real social media content.
☆ Plug-and-Play Interpretable Responsible Text-to-Image Generation via Dual-Space Multi-facet Concept Control
Ethical issues around text-to-image (T2I) models demand a comprehensive control over the generative content. Existing techniques addressing these issues for responsible T2I models aim for the generated content to be fair and safe (non-violent/explicit). However, these methods remain bounded to handling the facets of responsibility concepts individually, while also lacking in interpretability. Moreover, they often require alteration to the original model, which compromises the model performance. In this work, we propose a unique technique to enable responsible T2I generation by simultaneously accounting for an extensive range of concepts for fair and safe content generation in a scalable manner. The key idea is to distill the target T2I pipeline with an external plug-and-play mechanism that learns an interpretable composite responsible space for the desired concepts, conditioned on the target T2I pipeline. We use knowledge distillation and concept whitening to enable this. At inference, the learned space is utilized to modulate the generative content. A typical T2I pipeline presents two plug-in points for our approach, namely; the text embedding space and the diffusion model latent space. We develop modules for both points and show the effectiveness of our approach with a range of strong results.
☆ Bridging Writing Manner Gap in Visual Instruction Tuning by Creating LLM-aligned Instructions
In the realm of Large Multi-modal Models (LMMs), the instruction quality during the visual instruction tuning stage significantly influences the performance of modality alignment. In this paper, we assess the instruction quality from a unique perspective termed \textbf{Writing Manner}, which encompasses the selection of vocabulary, grammar and sentence structure to convey specific semantics. We argue that there exists a substantial writing manner gap between the visual instructions and the base Large Language Models (LLMs) within LMMs. This gap forces the pre-trained base LLMs to deviate from their original writing styles, leading to capability degradation of both base LLMs and LMMs. To bridge the writing manner gap while preserving the original semantics, we propose directly leveraging the base LLM to align the writing manner of soft-format visual instructions with that of the base LLM itself, resulting in novel LLM-aligned instructions. The manual writing manner evaluation results demonstrate that our approach successfully minimizes the writing manner gap. By utilizing LLM-aligned instructions, the baseline models LLaVA-7B and QwenVL demonstrate enhanced resistance to hallucinations and non-trivial comprehensive improvements across all $15$ visual and language benchmarks.
☆ LoTUS: Large-Scale Machine Unlearning with a Taste of Uncertainty CVPR 2025
We present LoTUS, a novel Machine Unlearning (MU) method that eliminates the influence of training samples from pre-trained models, avoiding retraining from scratch. LoTUS smooths the prediction probabilities of the model -- up to an information theoretic bound -- mitigating its over-confidence that stems from data memorization. We evaluate LoTUS on the Transformer and ResNet18 models, against eight baseline methods, on five public datasets. Beyond established MU benchmarks, we evaluate unlearning on a large-scale dataset (ImageNet1k) which deters retraining, simulating real-world conditions. Moreover, we introduce the novel Retrain-Free Jensen-Shannon Divergence (RF-JSD) metric to enable evaluation under real-world conditions. Experimental results show that LoTUS outperforms state-of-the-art methods in terms of both efficiency and effectiveness. Code: https://github.com/cspartalis/LoTUS.
comment: Accepted as a main conference paper at CVPR 2025
☆ DeepFund: Will LLM be Professional at Fund Investment? A Live Arena Perspective
Large Language Models (LLMs) have demonstrated impressive capabilities across various domains, but their effectiveness in financial decision making, particularly in fund investment, remains inadequately evaluated. Current benchmarks primarily assess LLMs understanding of financial documents rather than their ability to manage assets or analyze trading opportunities in dynamic market conditions. A critical limitation in existing evaluation methodologies is the backtesting approach, which suffers from information leakage when LLMs are evaluated on historical data they may have encountered during pretraining. This paper introduces DeepFund, a comprehensive platform for evaluating LLM based trading strategies in a simulated live environment. Our approach implements a multi agent framework where LLMs serve as both analysts and managers, creating a realistic simulation of investment decision making. The platform employs a forward testing methodology that mitigates information leakage by evaluating models on market data released after their training cutoff dates. We provide a web interface that visualizes model performance across different market conditions and investment parameters, enabling detailed comparative analysis. Through DeepFund, we aim to provide a more accurate and fair assessment of LLMs capabilities in fund investment, offering insights into their potential real world applications in financial markets.
comment: Work in progress
☆ How to Capture and Study Conversations Between Research Participants and ChatGPT: GPT for Researchers (g4r.org)
As large language models (LLMs) like ChatGPT become increasingly integrated into our everyday lives--from customer service and education to creative work and personal productivity--understanding how people interact with these AI systems has become a pressing issue. Despite the widespread use of LLMs, researchers lack standardized tools for systematically studying people's interactions with LLMs. To address this issue, we introduce GPT for Researchers (G4R), or g4r.org, a free website that researchers can use to easily create and integrate a GPT Interface into their studies. At g4r.org, researchers can (1) enable their study participants to interact with GPT (such as ChatGPT), (2) customize GPT Interfaces to guide participants' interactions with GPT (e.g., set constraints on topics or adjust GPT's tone or response style), and (3) capture participants' interactions with GPT by downloading data on messages exchanged between participants and GPT. By facilitating study participants' interactions with GPT and providing detailed data on these interactions, G4R can support research on topics such as consumer interactions with AI agents or LLMs, AI-assisted decision-making, and linguistic patterns in human-AI communication. With this goal in mind, we provide a step-by-step guide to using G4R at g4r.org.
☆ DiffMove: Group Mobility Tendency Enhanced Trajectory Recovery via Diffusion Model
In the real world, trajectory data is often sparse and incomplete due to low collection frequencies or limited device coverage. Trajectory recovery aims to recover these missing trajectory points, making the trajectories denser and more complete. However, this task faces two key challenges: 1) The excessive sparsity of individual trajectories makes it difficult to effectively leverage historical information for recovery; 2) Sparse trajectories make it harder to capture complex individual mobility preferences. To address these challenges, we propose a novel method called DiffMove. Firstly, we harness crowd wisdom for trajectory recovery. Specifically, we construct a group tendency graph using the collective trajectories of all users and then integrate the group mobility trends into the location representations via graph embedding. This solves the challenge of sparse trajectories being unable to rely on individual historical trajectories for recovery. Secondly, we capture individual mobility preferences from both historical and current perspectives. Finally, we integrate group mobility tendencies and individual preferences into the spatiotemporal distribution of the trajectory to recover high-quality trajectories. Extensive experiments on two real-world datasets demonstrate that DiffMove outperforms existing state-of-the-art methods. Further analysis validates the robustness of our method.
☆ When is dataset cartography ineffective? Using training dynamics does not improve robustness against Adversarial SQuAD
In this paper, I investigate the effectiveness of dataset cartography for extractive question answering on the SQuAD dataset. I begin by analyzing annotation artifacts in SQuAD and evaluate the impact of two adversarial datasets, AddSent and AddOneSent, on an ELECTRA-small model. Using training dynamics, I partition SQuAD into easy-to-learn, ambiguous, and hard-to-learn subsets. I then compare the performance of models trained on these subsets to those trained on randomly selected samples of equal size. Results show that training on cartography-based subsets does not improve generalization to the SQuAD validation set or the AddSent adversarial set. While the hard-to-learn subset yields a slightly higher F1 score on the AddOneSent dataset, the overall gains are limited. These findings suggest that dataset cartography provides little benefit for adversarial robustness in SQuAD-style QA tasks. I conclude by comparing these results to prior findings on SNLI and discuss possible reasons for the observed differences.
comment: 5 pages, 3 figures, 4 tables
☆ Voxel-based Point Cloud Geometry Compression with Space-to-Channel Context
Voxel-based methods are among the most efficient for point cloud geometry compression, particularly with dense point clouds. However, they face limitations due to a restricted receptive field, especially when handling high-bit depth point clouds. To overcome this issue, we introduce a stage-wise Space-to-Channel (S2C) context model for both dense point clouds and low-level sparse point clouds. This model utilizes a channel-wise autoregressive strategy to effectively integrate neighborhood information at a coarse resolution. For high-level sparse point clouds, we further propose a level-wise S2C context model that addresses resolution limitations by incorporating Geometry Residual Coding (GRC) for consistent-resolution cross-level prediction. Additionally, we use the spherical coordinate system for its compact representation and enhance our GRC approach with a Residual Probability Approximation (RPA) module, which features a large kernel size. Experimental results show that our S2C context model not only achieves bit savings while maintaining or improving reconstruction quality but also reduces computational complexity compared to state-of-the-art voxel-based compression methods.
☆ TopV: Compatible Token Pruning with Inference Time Optimization for Fast and Low-Memory Multimodal Vision Language Model CVPR 2025
Vision-Language Models (VLMs) demand substantial computational resources during inference, largely due to the extensive visual input tokens for representing visual information. Previous studies have noted that visual tokens tend to receive less attention than text tokens, suggesting their lower importance during inference and potential for pruning. However, their methods encounter several challenges: reliance on greedy heuristic criteria for token importance and incompatibility with FlashAttention and KV cache. To address these issues, we introduce \textbf{TopV}, a compatible \textbf{TO}ken \textbf{P}runing with inference Time Optimization for fast and low-memory \textbf{V}LM, achieving efficient pruning without additional training or fine-tuning. Instead of relying on attention scores, we formulate token pruning as an optimization problem, accurately identifying important visual tokens while remaining compatible with FlashAttention. Additionally, since we only perform this pruning once during the prefilling stage, it effectively reduces KV cache size. Our optimization framework incorporates a visual-aware cost function considering factors such as Feature Similarity, Relative Spatial Distance, and Absolute Central Distance, to measure the importance of each source visual token, enabling effective pruning of low-importance tokens. Extensive experiments demonstrate that our method outperforms previous token pruning methods, validating the effectiveness and efficiency of our approach.
comment: Accepted by CVPR 2025
☆ Risk Management for Distributed Arbitrage Systems: Integrating Artificial Intelligence
Effective risk management solutions become absolutely crucial when financial markets embrace distributed technology and decentralized financing (DeFi). This study offers a thorough survey and comparative analysis of the integration of artificial intelligence (AI) in risk management for distributed arbitrage systems. We examine several modern caching techniques namely in memory caching, distributed caching, and proxy caching and their functions in enhancing performance in decentralized settings. Through literature review we examine the utilization of AI techniques for alleviating risks related to market volatility, liquidity challenges, operational failures, regulatory compliance, and security threats. This comparison research evaluates various case studies from prominent DeFi technologies, emphasizing critical performance metrics like latency reduction, load balancing, and system resilience. Additionally, we examine the problems and trade offs associated with these technologies, emphasizing their effects on consistency, scalability, and fault tolerance. By meticulously analyzing real world applications, specifically centering on the Aave platform as our principal case study, we illustrate how the purposeful amalgamation of AI with contemporary caching methodologies has revolutionized risk management in distributed arbitrage systems.
comment: International Conference on AI and Financial Innovation AIFI-2025
☆ Severing Spurious Correlations with Data Pruning ICLR 2025
Deep neural networks have been shown to learn and rely on spurious correlations present in the data that they are trained on. Reliance on such correlations can cause these networks to malfunction when deployed in the real world, where these correlations may no longer hold. To overcome the learning of and reliance on such correlations, recent studies propose approaches that yield promising results. These works, however, study settings where the strength of the spurious signal is significantly greater than that of the core, invariant signal, making it easier to detect the presence of spurious features in individual training samples and allow for further processing. In this paper, we identify new settings where the strength of the spurious signal is relatively weaker, making it difficult to detect any spurious information while continuing to have catastrophic consequences. We also discover that spurious correlations are learned primarily due to only a handful of all the samples containing the spurious feature and develop a novel data pruning technique that identifies and prunes small subsets of the training data that contain these samples. Our proposed technique does not require inferred domain knowledge, information regarding the sample-wise presence or nature of spurious information, or human intervention. Finally, we show that such data pruning attains state-of-the-art performance on previously studied settings where spurious information is identifiable.
comment: ICLR 2025, Spotlight
☆ The Human-Machine Identity Blur: A Unified Framework for Cybersecurity Risk Management in 2025
The modern enterprise is facing an unprecedented surge in digital identities, with machine identities now significantly outnumbering human identities. This paper examines the cybersecurity risks emerging from what we define as the "human-machine identity blur" - the point at which human and machine identities intersect, delegate authority, and create new attack surfaces. Drawing from industry data, expert insights, and real-world incident analysis, we identify key governance gaps in current identity management models that treat human and machine entities as separate domains. To address these challenges, we propose a Unified Identity Governance Framework based on four core principles: treating identity as a continuum rather than a binary distinction, applying consistent risk evaluation across all identity types, implementing continuous verification guided by zero trust principles, and maintaining governance throughout the entire identity lifecycle. Our research shows that organizations adopting this unified approach experience a 47 percent reduction in identity-related security incidents and a 62 percent improvement in incident response time. We conclude by offering a practical implementation roadmap and outlining future research directions as AI-driven systems become increasingly autonomous.
comment: 9 pages, 8 figures
☆ LLM Benchmarking with LLaMA2: Evaluating Code Development Performance Across Multiple Programming Languages
The rapid evolution of large language models (LLMs) has opened new possibilities for automating various tasks in software development. This paper evaluates the capabilities of the Llama 2-70B model in automating these tasks for scientific applications written in commonly used programming languages. Using representative test problems, we assess the model's capacity to generate code, documentation, and unit tests, as well as its ability to translate existing code between commonly used programming languages. Our comprehensive analysis evaluates the compilation, runtime behavior, and correctness of the generated and translated code. Additionally, we assess the quality of automatically generated code, documentation and unit tests. Our results indicate that while Llama 2-70B frequently generates syntactically correct and functional code for simpler numerical tasks, it encounters substantial difficulties with more complex, parallelized, or distributed computations, requiring considerable manual corrections. We identify key limitations and suggest areas for future improvements to better leverage AI-driven automation in scientific computing workflows.
☆ A Survey of Large Language Model Agents for Question Answering
This paper surveys the development of large language model (LLM)-based agents for question answering (QA). Traditional agents face significant limitations, including substantial data requirements and difficulty in generalizing to new environments. LLM-based agents address these challenges by leveraging LLMs as their core reasoning engine. These agents achieve superior QA results compared to traditional QA pipelines and naive LLM QA systems by enabling interaction with external environments. We systematically review the design of LLM agents in the context of QA tasks, organizing our discussion across key stages: planning, question understanding, information retrieval, and answer generation. Additionally, this paper identifies ongoing challenges and explores future research directions to enhance the performance of LLM agent QA systems.
☆ Continual Reinforcement Learning for HVAC Systems Control: Integrating Hypernetworks and Transfer Learning
Buildings with Heating, Ventilation, and Air Conditioning (HVAC) systems play a crucial role in ensuring indoor comfort and efficiency. While traditionally governed by physics-based models, the emergence of big data has enabled data-driven methods like Deep Reinforcement Learning (DRL). However, Reinforcement Learning (RL)-based techniques often suffer from sample inefficiency and limited generalization, especially across varying HVAC systems. We introduce a model-based reinforcement learning framework that uses a Hypernetwork to continuously learn environment dynamics across tasks with different action spaces. This enables efficient synthetic rollout generation and improved sample usage. Our approach demonstrates strong backward transfer in a continual learning setting after training on a second task, minimal fine-tuning on the first task allows rapid convergence within just 5 episodes and thus outperforming Model Free Reinforcement Learning (MFRL) and effectively mitigating catastrophic forgetting. These findings have significant implications for reducing energy consumption and operational costs in building management, thus supporting global sustainability goals. Keywords: Deep Reinforcement Learning, HVAC Systems Control, Hypernetworks, Transfer and Continual Learning, Catastrophic Forgetting
☆ Overtrained Language Models Are Harder to Fine-Tune
Large language models are pre-trained on ever-growing token budgets under the assumption that better pre-training performance translates to improved downstream models. In this work, we challenge this assumption and show that extended pre-training can make models harder to fine-tune, leading to degraded final performance. We term this phenomenon catastrophic overtraining. For example, the instruction-tuned OLMo-1B model pre-trained on 3T tokens leads to over 2% worse performance on multiple standard LLM benchmarks than its 2.3T token counterpart. Through controlled experiments and theoretical analysis, we show that catastrophic overtraining arises from a systematic increase in the broad sensitivity of pre-trained parameters to modifications, including but not limited to fine-tuning. Our findings call for a critical reassessment of pre-training design that considers the downstream adaptability of the model.
comment: 72 pages, 65 figures, 6 tables
☆ A Shared Low-Rank Adaptation Approach to Personalized RLHF AISTATS 2025
Reinforcement Learning from Human Feedback (RLHF) has emerged as a pivotal technique for aligning artificial intelligence systems with human values, achieving remarkable success in fine-tuning large language models. However, existing RLHF frameworks often assume that human preferences are relatively homogeneous and can be captured by a single, unified reward model. This assumption overlooks the inherent diversity and heterogeneity across individuals, limiting the adaptability of RLHF to personalized scenarios and risking misalignments that can diminish user satisfaction and trust in AI systems. In this paper, we address these challenges by introducing Low-Rank Adaptation (LoRA) into the personalized RLHF framework. We apply LoRA in the the aggregated parameter space of all personalized reward functions, thereby enabling efficient learning of personalized reward models from potentially limited local datasets. Our approach exploits potential shared structures among the local ground-truth reward models while allowing for individual adaptation, without relying on restrictive assumptions about shared representations as in prior works. We further establish sample complexity guarantees for our method. Theoretical analysis demonstrates the effectiveness of the proposed approach in capturing both shared and individual-specific structures within heterogeneous human preferences, addressing the dual challenge of personalization requirements and practical data constraints. Experimental results on real-world datasets corroborate the efficiency of our algorithm in the personalized RLHF setting.
comment: Published as a conference paper at AISTATS 2025
☆ Mining-Gym: A Configurable RL Benchmarking Environment for Truck Dispatch Scheduling
Mining process optimization particularly truck dispatch scheduling is a critical factor in enhancing the efficiency of open pit mining operations However the dynamic and stochastic nature of mining environments characterized by uncertainties such as equipment failures truck maintenance and variable haul cycle times poses significant challenges for traditional optimization methods While Reinforcement Learning RL has shown promise in adaptive decision making for mining logistics its practical deployment requires rigorous evaluation in realistic and customizable simulation environments The lack of standardized benchmarking environments limits fair algorithm comparisons reproducibility and the real world applicability of RL based approaches in open pit mining settings To address this challenge we introduce Mining Gym a configurable open source benchmarking environment designed for training testing and comparing RL algorithms in mining process optimization Built on Discrete Event Simulation DES and seamlessly integrated with the OpenAI Gym interface Mining Gym provides a structured testbed that enables the direct application of advanced RL algorithms from Stable Baselines The framework models key mining specific uncertainties such as equipment failures queue congestion and the stochasticity of mining processes ensuring a realistic and adaptive learning environment Additionally Mining Gym features a graphical user interface GUI for intuitive mine site configuration a comprehensive data logging system a built in KPI dashboard and real time visual representation of the mine site These capabilities facilitate standardized reproducible evaluations across multiple RL strategies and baseline heuristics
comment: 11 pages, 8 figures
☆ Browsing Lost Unformed Recollections: A Benchmark for Tip-of-the-Tongue Search and Reasoning
We introduce Browsing Lost Unformed Recollections, a tip-of-the-tongue known-item search and reasoning benchmark for general AI assistants. BLUR introduces a set of 573 real-world validated questions that demand searching and reasoning across multi-modal and multilingual inputs, as well as proficient tool use, in order to excel on. Humans easily ace these questions (scoring on average 98%), while the best-performing system scores around 56%. To facilitate progress toward addressing this challenging and aspirational use case for general AI assistants, we release 350 questions through a public leaderboard, retain the answers to 250 of them, and have the rest as a private test set.
☆ SoK: How Robust is Audio Watermarking in Generative AI models?
Audio watermarking is increasingly used to verify the provenance of AI-generated content, enabling applications such as detecting AI-generated speech, protecting music IP, and defending against voice cloning. To be effective, audio watermarks must resist removal attacks that distort signals to evade detection. While many schemes claim robustness, these claims are typically tested in isolation and against a limited set of attacks. A systematic evaluation against diverse removal attacks is lacking, hindering practical deployment. In this paper, we investigate whether recent watermarking schemes that claim robustness can withstand a broad range of removal attacks. First, we introduce a taxonomy covering 22 audio watermarking schemes. Next, we summarize their underlying technologies and potential vulnerabilities. We then present a large-scale empirical study to assess their robustness. To support this, we build an evaluation framework encompassing 22 types of removal attacks (109 configurations) including signal-level, physical-level, and AI-induced distortions. We reproduce 9 watermarking schemes using open-source code, identify 8 new highly effective attacks, and highlight 11 key findings that expose the fundamental limitations of these methods across 3 public datasets. Our results reveal that none of the surveyed schemes can withstand all tested distortions. This evaluation offers a comprehensive view of how current watermarking methods perform under real-world threats. Our demo and code are available at https://sokaudiowm.github.io/.
☆ AssertionForge: Enhancing Formal Verification Assertion Generation with Structured Representation of Specifications and RTL
Generating SystemVerilog Assertions (SVAs) from natural language specifications remains a major challenge in formal verification (FV) due to the inherent ambiguity and incompleteness of specifications. Existing LLM-based approaches, such as AssertLLM, focus on extracting information solely from specification documents, often failing to capture essential internal signal interactions and design details present in the RTL code, leading to incomplete or incorrect assertions. We propose a novel approach that constructs a Knowledge Graph (KG) from both specifications and RTL, using a hardware-specific schema with domain-specific entity and relation types. We create an initial KG from the specification and then systematically fuse it with information extracted from the RTL code, resulting in a unified, comprehensive KG. This combined representation enables a more thorough understanding of the design and allows for a multi-resolution context synthesis process which is designed to extract diverse verification contexts from the KG. Experiments on four designs demonstrate that our method significantly enhances SVA quality over prior methods. This structured representation not only improves FV but also paves the way for future research in tasks like code generation and design understanding.
☆ PSO-UNet: Particle Swarm-Optimized U-Net Framework for Precise Multimodal Brain Tumor Segmentation
Medical image segmentation, particularly for brain tumor analysis, demands precise and computationally efficient models due to the complexity of multimodal MRI datasets and diverse tumor morphologies. This study introduces PSO-UNet, which integrates Particle Swarm Optimization (PSO) with the U-Net architecture for dynamic hyperparameter optimization. Unlike traditional manual tuning or alternative optimization approaches, PSO effectively navigates complex hyperparameter search spaces, explicitly optimizing the number of filters, kernel size, and learning rate. PSO-UNet substantially enhances segmentation performance, achieving Dice Similarity Coefficients (DSC) of 0.9578 and 0.9523 and Intersection over Union (IoU) scores of 0.9194 and 0.9097 on the BraTS 2021 and Figshare datasets, respectively. Moreover, the method reduces computational complexity significantly, utilizing only 7.8 million parameters and executing in approximately 906 seconds, markedly faster than comparable U-Net-based frameworks. These outcomes underscore PSO-UNet's robust generalization capabilities across diverse MRI modalities and tumor classifications, emphasizing its clinical potential and clear advantages over conventional hyperparameter tuning methods. Future research will explore hybrid optimization strategies and validate the framework against other bio-inspired algorithms to enhance its robustness and scalability.
comment: 9 pages, 6 figures, 4 tables, Gecco 2025 Conference
☆ Overcoming Vocabulary Mismatch: Vocabulary-agnostic Teacher Guided Language Modeling
Using large teacher models to guide the training of smaller student models has become the prevailing paradigm for efficient and effective learning. However, vocabulary mismatches between teacher and student language models pose significant challenges in language modeling, resulting in divergent token sequences and output distributions. To overcome these limitations, we propose Vocabulary-agnostic Teacher Guided Language Modeling (VocAgnoLM), a novel approach that bridges the gap caused by vocabulary mismatch through two key methods: (1) Token-level Lexical Alignment, which aligns token sequences across mismatched vocabularies, and (2) Teacher Guided Loss, which leverages the loss of teacher model to guide effective student training. We demonstrate its effectiveness in language modeling with 1B student model using various 7B teacher models with different vocabularies. Notably, with Qwen2.5-Math-Instruct, a teacher model sharing only about 6% of its vocabulary with TinyLlama, VocAgnoLM achieves a 46% performance improvement compared to naive continual pretraining. Furthermore, we demonstrate that VocAgnoLM consistently benefits from stronger teacher models, providing a robust solution to vocabulary mismatches in language modeling.
☆ Where is this coming from? Making groundedness count in the evaluation of Document VQA models NAACL
Document Visual Question Answering (VQA) models have evolved at an impressive rate over the past few years, coming close to or matching human performance on some benchmarks. We argue that common evaluation metrics used by popular benchmarks do not account for the semantic and multimodal groundedness of a model's outputs. As a result, hallucinations and major semantic errors are treated the same way as well-grounded outputs, and the evaluation scores do not reflect the reasoning capabilities of the model. In response, we propose a new evaluation methodology that accounts for the groundedness of predictions with regard to the semantic characteristics of the output as well as the multimodal placement of the output within the input document. Our proposed methodology is parameterized in such a way that users can configure the score according to their preferences. We validate our scoring methodology using human judgment and show its potential impact on existing popular leaderboards. Through extensive analyses, we demonstrate that our proposed method produces scores that are a better indicator of a model's robustness and tends to give higher rewards to better-calibrated answers.
comment: Accepted to NAACL Findings 2025
☆ Information-Seeking Decision Strategies Mitigate Risk in Dynamic, Uncertain Environments
To survive in dynamic and uncertain environments, individuals must develop effective decision strategies that balance information gathering and decision commitment. Models of such strategies often prioritize either optimizing tangible payoffs, like reward rate, or gathering information to support a diversity of (possibly unknown) objectives. However, our understanding of the relative merits of these two approaches remains incomplete, in part because direct comparisons have been limited to idealized, static environments that lack the dynamic complexity of the real world. Here we compared the performance of normative reward- and information-seeking strategies in a dynamic foraging task. Both strategies show similar transitions between exploratory and exploitative behaviors as environmental uncertainty changes. However, we find subtle disparities in the actions they take, resulting in meaningful performance differences: whereas reward-seeking strategies generate slightly more reward on average, information-seeking strategies provide more consistent and predictable outcomes. Our findings support the adaptive value of information-seeking behaviors that can mitigate risk with minimal reward loss.
☆ Anomaly Detection Using Computer Vision: A Comparative Analysis of Class Distinction and Performance Metrics
This paper showcases an experimental study on anomaly detection using computer vision. The study focuses on class distinction and performance evaluation, combining OpenCV with deep learning techniques while employing a TensorFlow-based convolutional neural network for real-time face recognition and classification. The system effectively distinguishes among three classes: authorized personnel (admin), intruders, and non-human entities. A MobileNetV2-based deep learning model is utilized to optimize real-time performance, ensuring high computational efficiency without compromising accuracy. Extensive dataset preprocessing, including image augmentation and normalization, enhances the models generalization capabilities. Our analysis demonstrates classification accuracies of 90.20% for admin, 98.60% for intruders, and 75.80% for non-human detection, while maintaining an average processing rate of 30 frames per second. The study leverages transfer learning, batch normalization, and Adam optimization to achieve stable and robust learning, and a comparative analysis of class differentiation strategies highlights the impact of feature extraction techniques and training methodologies. The results indicate that advanced feature selection and data augmentation significantly enhance detection performance, particularly in distinguishing human from non-human scenes. As an experimental study, this research provides critical insights into optimizing deep learning-based surveillance systems for high-security environments and improving the accuracy and efficiency of real-time anomaly detection.
comment: 6 pages, 4 figures
☆ Rankers, Judges, and Assistants: Towards Understanding the Interplay of LLMs in Information Retrieval Evaluation
Large language models (LLMs) are increasingly integral to information retrieval (IR), powering ranking, evaluation, and AI-assisted content creation. This widespread adoption necessitates a critical examination of potential biases arising from the interplay between these LLM-based components. This paper synthesizes existing research and presents novel experiment designs that explore how LLM-based rankers and assistants influence LLM-based judges. We provide the first empirical evidence of LLM judges exhibiting significant bias towards LLM-based rankers. Furthermore, we observe limitations in LLM judges' ability to discern subtle system performance differences. Contrary to some previous findings, our preliminary study does not find evidence of bias against AI-generated content. These results highlight the need for a more holistic view of the LLM-driven information ecosystem. To this end, we offer initial guidelines and a research agenda to ensure the reliable use of LLMs in IR evaluation.
☆ The Case for "Thick Evaluations" of Cultural Representation in AI
Generative AI image models have been increasingly evaluated for their (in)ability to represent non-Western cultures. We argue that these evaluations operate through reductive ideals of representation, abstracted from how people define their own representation and neglecting the inherently interpretive and contextual nature of cultural representation. In contrast to these 'thin' evaluations, we introduce the idea of 'thick evaluations': a more granular, situated, and discursive measurement framework for evaluating representations of social worlds in AI images, steeped in communities' own understandings of representation. We develop this evaluation framework through workshops in South Asia, by studying the 'thick' ways in which people interpret and assign meaning to images of their own cultures. We introduce practices for thicker evaluations of representation that expand the understanding of representation underpinning AI evaluations and by co-constructing metrics with communities, bringing measurement in line with the experiences of communities on the ground.
comment: 14 pages
☆ HingeRLC-GAN: Combating Mode Collapse with Hinge Loss and RLC Regularization
Recent advances in Generative Adversarial Networks (GANs) have demonstrated their capability for producing high-quality images. However, a significant challenge remains mode collapse, which occurs when the generator produces a limited number of data patterns that do not reflect the diversity of the training dataset. This study addresses this issue by proposing a number of architectural changes aimed at increasing the diversity and stability of GAN models. We start by improving the loss function with Wasserstein loss and Gradient Penalty to better capture the full range of data variations. We also investigate various network architectures and conclude that ResNet significantly contributes to increased diversity. Building on these findings, we introduce HingeRLC-GAN, a novel approach that combines RLC Regularization and the Hinge loss function. With a FID Score of 18 and a KID Score of 0.001, our approach outperforms existing methods by effectively balancing training stability and increased diversity.
☆ Minimum Volume Conformal Sets for Multivariate Regression
Conformal prediction provides a principled framework for constructing predictive sets with finite-sample validity. While much of the focus has been on univariate response variables, existing multivariate methods either impose rigid geometric assumptions or rely on flexible but computationally expensive approaches that do not explicitly optimize prediction set volume. We propose an optimization-driven framework based on a novel loss function that directly learns minimum-volume covering sets while ensuring valid coverage. This formulation naturally induces a new nonconformity score for conformal prediction, which adapts to the residual distribution and covariates. Our approach optimizes over prediction sets defined by arbitrary norm balls, including single and multi-norm formulations. Additionally, by jointly optimizing both the predictive model and predictive uncertainty, we obtain prediction sets that are tight, informative, and computationally efficient, as demonstrated in our experiments on real-world datasets.
☆ Mist: Efficient Distributed Training of Large Language Models via Memory-Parallelism Co-Optimization EuroSys 2025
Various parallelism, such as data, tensor, and pipeline parallelism, along with memory optimizations like activation checkpointing, redundancy elimination, and offloading, have been proposed to accelerate distributed training for Large Language Models. To find the best combination of these techniques, automatic distributed training systems are proposed. However, existing systems only tune a subset of optimizations, due to the lack of overlap awareness, inability to navigate the vast search space, and ignoring the inter-microbatch imbalance, leading to sub-optimal performance. To address these shortcomings, we propose Mist, a memory, overlap, and imbalance-aware automatic distributed training system that comprehensively co-optimizes all memory footprint reduction techniques alongside parallelism. Mist is based on three key ideas: (1) fine-grained overlap-centric scheduling, orchestrating optimizations in an overlapped manner, (2) symbolic-based performance analysis that predicts runtime and memory usage using symbolic expressions for fast tuning, and (3) imbalance-aware hierarchical tuning, decoupling the process into an inter-stage imbalance and overlap aware Mixed Integer Linear Programming problem and an intra-stage Dual-Objective Constrained Optimization problem, and connecting them through Pareto frontier sampling. Our evaluation results show that Mist achieves an average of 1.28$\times$ (up to 1.73$\times$) and 1.27$\times$ (up to 2.04$\times$) speedup compared to state-of-the-art manual system Megatron-LM and state-of-the-art automatic system Aceso, respectively.
comment: Accepted by EuroSys 2025
☆ Forecasting Labor Demand: Predicting JOLT Job Openings using Deep Learning Model
This thesis studies the effectiveness of Long Short Term Memory model in forecasting future Job Openings and Labor Turnover Survey data in the United States. Drawing on multiple economic indicators from various sources, the data are fed directly into LSTM model to predict JOLT job openings in subsequent periods. The performance of the LSTM model is compared with conventional autoregressive approaches, including ARIMA, SARIMA, and Holt-Winters. Findings suggest that the LSTM model outperforms these traditional models in predicting JOLT job openings, as it not only captures the dependent variables trends but also harmonized with key economic factors. These results highlight the potential of deep learning techniques in capturing complex temporal dependencies in economic data, offering valuable insights for policymakers and stakeholders in developing data-driven labor market strategies
comment: 19 pages, 7 figures
☆ LookAhead Tuning: Safer Language Models via Partial Answer Previews
Fine-tuning enables large language models (LLMs) to adapt to specific domains, but often undermines their previously established safety alignment. To mitigate the degradation of model safety during fine-tuning, we introduce LookAhead Tuning, which comprises two simple, low-resource, and effective data-driven methods that modify training data by previewing partial answer prefixes. Both methods aim to preserve the model's inherent safety mechanisms by minimizing perturbations to initial token distributions. Comprehensive experiments demonstrate that LookAhead Tuning effectively maintains model safety without sacrificing robust performance on downstream tasks. Our findings position LookAhead Tuning as a reliable and efficient solution for the safe and effective adaptation of LLMs. Code is released at https://github.com/zjunlp/LookAheadTuning.
comment: Work in progress
☆ Evolutionary Policy Optimization
Despite its extreme sample inefficiency, on-policy reinforcement learning has become a fundamental tool in real-world applications. With recent advances in GPU-driven simulation, the ability to collect vast amounts of data for RL training has scaled exponentially. However, studies show that current on-policy methods, such as PPO, fail to fully leverage the benefits of parallelized environments, leading to performance saturation beyond a certain scale. In contrast, Evolutionary Algorithms (EAs) excel at increasing diversity through randomization, making them a natural complement to RL. However, existing EvoRL methods have struggled to gain widespread adoption due to their extreme sample inefficiency. To address these challenges, we introduce Evolutionary Policy Optimization (EPO), a novel policy gradient algorithm that combines the strengths of EA and policy gradients. We show that EPO significantly improves performance across diverse and challenging environments, demonstrating superior scalability with parallelized simulations.
comment: Website at https://sites.google.com/view/epo-rl
☆ Role of AI Innovation, Clean Energy and Digital Economy towards Net Zero Emission in the United States: An ARDL Approach
The current paper investigates the influences of AI innovation, GDP growth, renewable energy utilization, the digital economy, and industrialization on CO2 emissions in the USA from 1990 to 2022, incorporating the ARDL methodology. The outcomes observe that AI innovation, renewable energy usage, and the digital economy reduce CO2 emissions, while GDP expansion and industrialization intensify ecosystem damage. Unit root tests (ADF, PP, and DF-GLS) reveal heterogeneous integration levels amongst components, ensuring robustness in the ARDL analysis. Complementary methods (FMOLS, DOLS, and CCR) validate the results, enhancing their reliability. Pairwise Granger causality assessments identify strong unidirectional connections within CO2 emissions and AI innovation, as well as the digital economy, underscoring their significant roles in ecological sustainability. This research highlights the requirement for strategic actions to nurture equitable growth, including advancements in AI technology, green energy adoption, and environmentally conscious industrial development, to improve environmental quality in the United States.
comment: 24 pages, 8 tables, 1 figure
☆ Option Discovery Using LLM-guided Semantic Hierarchical Reinforcement Learning
Large Language Models (LLMs) have shown remarkable promise in reasoning and decision-making, yet their integration with Reinforcement Learning (RL) for complex robotic tasks remains underexplored. In this paper, we propose an LLM-guided hierarchical RL framework, termed LDSC, that leverages LLM-driven subgoal selection and option reuse to enhance sample efficiency, generalization, and multi-task adaptability. Traditional RL methods often suffer from inefficient exploration and high computational cost. Hierarchical RL helps with these challenges, but existing methods often fail to reuse options effectively when faced with new tasks. To address these limitations, we introduce a three-stage framework that uses LLMs for subgoal generation given natural language description of the task, a reusable option learning and selection method, and an action-level policy, enabling more effective decision-making across diverse tasks. By incorporating LLMs for subgoal prediction and policy guidance, our approach improves exploration efficiency and enhances learning performance. On average, LDSC outperforms the baseline by 55.9\% in average reward, demonstrating its effectiveness in complex RL settings. More details and experiment videos could be found in \href{https://raaslab.org/projects/LDSC/}{this link\footnote{https://raaslab.org/projects/LDSC}}.
☆ Computational Thinking with Computer Vision: Developing AI Competency in an Introductory Computer Science Course AAAI 2025
Developing competency in artificial intelligence is becoming increasingly crucial for computer science (CS) students at all levels of the CS curriculum. However, most previous research focuses on advanced CS courses, as traditional introductory courses provide limited opportunities to develop AI skills and knowledge. This paper introduces an introductory CS course where students learn computational thinking through computer vision, a sub-field of AI, as an application context. The course aims to achieve computational thinking outcomes alongside critical thinking outcomes that expose students to AI approaches and their societal implications. Through experiential activities such as individual projects and reading discussions, our course seeks to balance technical learning and critical thinking goals. Our evaluation, based on pre-and post-course surveys, shows an improved sense of belonging, self-efficacy, and AI ethics awareness among students. The results suggest that an AI-focused context can enhance participation and employability, student-selected projects support self-efficacy, and ethically grounded AI instruction can be effective for interdisciplinary audiences. Students' discussions on reading assignments demonstrated deep engagement with the complex challenges in today's AI landscape. Finally, we share insights on scaling such courses for larger cohorts and improving the learning experience for introductory CS students.
comment: 7 pages, 3 figures, 3 tables, Proceedings of AAAI 2025
☆ DisentTalk: Cross-lingual Talking Face Generation via Semantic Disentangled Diffusion Model
Recent advances in talking face generation have significantly improved facial animation synthesis. However, existing approaches face fundamental limitations: 3DMM-based methods maintain temporal consistency but lack fine-grained regional control, while Stable Diffusion-based methods enable spatial manipulation but suffer from temporal inconsistencies. The integration of these approaches is hindered by incompatible control mechanisms and semantic entanglement of facial representations. This paper presents DisentTalk, introducing a data-driven semantic disentanglement framework that decomposes 3DMM expression parameters into meaningful subspaces for fine-grained facial control. Building upon this disentangled representation, we develop a hierarchical latent diffusion architecture that operates in 3DMM parameter space, integrating region-aware attention mechanisms to ensure both spatial precision and temporal coherence. To address the scarcity of high-quality Chinese training data, we introduce CHDTF, a Chinese high-definition talking face dataset. Extensive experiments show superior performance over existing methods across multiple metrics, including lip synchronization, expression quality, and temporal consistency. Project Page: https://kangweiiliu.github.io/DisentTalk.
☆ FACE: Few-shot Adapter with Cross-view Fusion for Cross-subject EEG Emotion Recognition
Cross-subject EEG emotion recognition is challenged by significant inter-subject variability and intricately entangled intra-subject variability. Existing works have primarily addressed these challenges through domain adaptation or generalization strategies. However, they typically require extensive target subject data or demonstrate limited generalization performance to unseen subjects. Recent few-shot learning paradigms attempt to address these limitations but often encounter catastrophic overfitting during subject-specific adaptation with limited samples. This article introduces the few-shot adapter with a cross-view fusion method called FACE for cross-subject EEG emotion recognition, which leverages dynamic multi-view fusion and effective subject-specific adaptation. Specifically, FACE incorporates a cross-view fusion module that dynamically integrates global brain connectivity with localized patterns via subject-specific fusion weights to provide complementary emotional information. Moreover, the few-shot adapter module is proposed to enable rapid adaptation for unseen subjects while reducing overfitting by enhancing adapter structures with meta-learning. Experimental results on three public EEG emotion recognition benchmarks demonstrate FACE's superior generalization performance over state-of-the-art methods. FACE provides a practical solution for cross-subject scenarios with limited labeled data.
comment: Under Review
☆ Fundamental Safety-Capability Trade-offs in Fine-tuning Large Language Models
Fine-tuning Large Language Models (LLMs) on some task-specific datasets has been a primary use of LLMs. However, it has been empirically observed that this approach to enhancing capability inevitably compromises safety, a phenomenon also known as the safety-capability trade-off in LLM fine-tuning. This paper presents a theoretical framework for understanding the interplay between safety and capability in two primary safety-aware LLM fine-tuning strategies, providing new insights into the effects of data similarity, context overlap, and alignment loss landscape. Our theoretical results characterize the fundamental limits of the safety-capability trade-off in LLM fine-tuning, which are also validated by numerical experiments.
comment: The first two authors contribute equally to this work and are listed in alphabetical order
♻ ☆ Large Language Models Empowered Personalized Web Agents WWW 2025
Web agents have emerged as a promising direction to automate Web task completion based on user instructions, significantly enhancing user experience. Recently, Web agents have evolved from traditional agents to Large Language Models (LLMs)-based Web agents. Despite their success, existing LLM-based Web agents overlook the importance of personalized data (e.g., user profiles and historical Web behaviors) in assisting the understanding of users' personalized instructions and executing customized actions. To overcome the limitation, we first formulate the task of LLM-empowered personalized Web agents, which integrate personalized data and user instructions to personalize instruction comprehension and action execution. To address the absence of a comprehensive evaluation benchmark, we construct a Personalized Web Agent Benchmark (PersonalWAB), featuring user instructions, personalized user data, Web functions, and two evaluation paradigms across three personalized Web tasks. Moreover, we propose a Personalized User Memory-enhanced Alignment (PUMA) framework to adapt LLMs to the personalized Web agent task. PUMA utilizes a memory bank with a task-specific retrieval strategy to filter relevant historical Web behaviors. Based on the behaviors, PUMA then aligns LLMs for personalized action execution through fine-tuning and direct preference optimization. Extensive experiments validate the superiority of PUMA over existing Web agents on PersonalWAB.
comment: Accepted to WWW 2025. The code and data are available on the project website https://hongrucai.github.io/PersonalWAB/
♻ ☆ Subversion Strategy Eval: Can language models statelessly strategize to subvert control protocols?
An AI control protocol is a plan for usefully deploying AI systems that aims to prevent an AI from intentionally causing some unacceptable outcome. This paper investigates how well AI systems can generate and act on their own strategies for subverting control protocols whilst operating statelessly (without shared memory between contexts). To do this, an AI system may need to reliably generate optimal plans in each context, take actions with well-calibrated probabilities, and coordinate plans with other instances of itself without communicating. We develop Subversion Strategy Eval, a suite of eight environments, covering a range of protocols and strategic capabilities, and six sets of affordances that help isolate individual capabilities. We implement the evaluation in Inspect-AI and release it open-source. We evaluate Claude 3.5 models, including helpful-only versions, as well as OpenAI reasoning models. None of the models demonstrate substantial capability in strategizing to subvert control protocols statelessly. However, providing models with additional affordances, such as the ability to share a plan between contexts, can substantially improve performance. We hope our evaluations can act as a leading indicator for when models are capable of subverting control protocols and also relax the worst-case assumption of perfect strategic ability in AI control evaluations.
♻ ☆ FlowDock: Geometric Flow Matching for Generative Protein-Ligand Docking and Affinity Prediction
Powerful generative AI models of protein-ligand structure have recently been proposed, but few of these methods support both flexible protein-ligand docking and affinity estimation. Of those that do, none can directly model multiple binding ligands concurrently or have been rigorously benchmarked on pharmacologically relevant drug targets, hindering their widespread adoption in drug discovery efforts. In this work, we propose FlowDock, the first deep geometric generative model based on conditional flow matching that learns to directly map unbound (apo) structures to their bound (holo) counterparts for an arbitrary number of binding ligands. Furthermore, FlowDock provides predicted structural confidence scores and binding affinity values with each of its generated protein-ligand complex structures, enabling fast virtual screening of new (multi-ligand) drug targets. For the well-known PoseBusters Benchmark dataset, FlowDock outperforms single-sequence AlphaFold 3 with a 51% blind docking success rate using unbound (apo) protein input structures and without any information derived from multiple sequence alignments, and for the challenging new DockGen-E dataset, FlowDock outperforms single-sequence AlphaFold 3 and matches single-sequence Chai-1 for binding pocket generalization. Additionally, in the ligand category of the 16th community-wide Critical Assessment of Techniques for Structure Prediction (CASP16), FlowDock ranked among the top-5 methods for pharmacological binding affinity estimation across 140 protein-ligand complexes, demonstrating the efficacy of its learned representations in virtual screening. Source code, data, and pre-trained models are available at https://github.com/BioinfoMachineLearning/FlowDock.
comment: 15 pages, 2 tables, 2 algorithms, 11 figures. Code, data, pre-trained models, and baseline method predictions are available at https://github.com/BioinfoMachineLearning/FlowDock
♻ ☆ Visual Position Prompt for MLLM based Visual Grounding
Although Multimodal Large Language Models (MLLMs) excel at various image-related tasks, they encounter challenges in precisely aligning coordinates with spatial information within images, particularly in position-aware tasks such as visual grounding. This limitation arises from two key factors. First, MLLMs lack explicit spatial references, making it difficult to associate textual descriptions with precise image locations. Second, their feature extraction processes prioritize global context over fine-grained spatial details, leading to weak localization capability. To address this issue, we introduce VPP-LLaVA, an MLLM equipped with Visual Position Prompt (VPP) to improve its grounding capability. VPP-LLaVA integrates two complementary mechanisms. The global VPP overlays learnable, axis-like embeddings onto the input image to provide structured spatial cues. The local VPP focuses on fine-grained localization by incorporating position-aware queries, which suggests probable object locations. We also introduce a VPP-SFT dataset with 0.6M samples, consolidating high-quality visual grounding data into a compact format for efficient model training. Training on this dataset with VPP enhances the model's performance, achieving state-of-the-art results on standard grounding benchmarks despite using fewer training samples compared to other MLLMs like MiniGPT-v2, which rely on much larger datasets ($\sim$21M samples). The code and VPP-SFT dataset will be available at https://github.com/WayneTomas/VPP-LLaVA upon acceptance.
♻ ☆ STEVE: A Step Verification Pipeline for Computer-use Agent Training
Developing AI agents to autonomously manipulate graphical user interfaces is a long challenging task. Recent advances in data scaling law inspire us to train computer-use agents with a scaled instruction set, yet using behavior cloning to train agents still requires immense high-quality trajectories. To meet the scalability need, we designed STEVE, a step verification pipeline for computer-use agent training. First, we establish a large instruction set for computer-use agents and collect trajectory data with some suboptimal agents. GPT-4o is used to verify the correctness of each step in the trajectories based on the screens before and after the action execution, assigning each step with a binary label. Last, we adopt the Kahneman and Tversky Optimization to optimize the agent from the binary stepwise labels. Extensive experiments manifest that our agent outperforms supervised finetuning by leveraging both positive and negative actions within a trajectory. Also, STEVE enables us to train a 7B vision-language model as a computer-use agent, achieving leading performance in the challenging live desktop environment WinAgentArena with great efficiency at a reduced cost. Code and data: https://github.com/FanbinLu/STEVE.
♻ ☆ Lightweight Embedded FPGA Deployment of Learned Image Compression with Knowledge Distillation and Hybrid Quantization
Learnable Image Compression (LIC) has shown the potential to outperform standardized video codecs in RD efficiency, prompting the research for hardware-friendly implementations. Most existing LIC hardware implementations prioritize latency to RD-efficiency and through an extensive exploration of the hardware design space. We present a novel design paradigm where the burden of tuning the design for a specific hardware platform is shifted towards model dimensioning and without compromising on RD-efficiency. First, we design a framework for distilling a leaner student LIC model from a reference teacher: by tuning a single model hyperparameters, we can meet the constraints of different hardware platforms without a complex hardware design exploration. Second, we propose a hardware-friendly implementation of the Generalized Divisive Normalization - GDN activation that preserves RD efficiency even post parameter quantization. Third, we design a pipelined FPGA configuration which takes full advantage of available FPGA resources by leveraging parallel processing and optimizing resource allocation. Our experiments with a state of the art LIC model show that we outperform all existing FPGA implementations while performing very close to the original model.
comment: Submitted to IEEE Transactions on Circuits and Systems for Video Technology in March 2025
♻ ☆ CarbonSense: A Multimodal Dataset and Baseline for Carbon Flux Modelling
Terrestrial carbon fluxes provide vital information about our biosphere's health and its capacity to absorb anthropogenic CO$_2$ emissions. The importance of predicting carbon fluxes has led to the emerging field of data-driven carbon flux modelling (DDCFM), which uses statistical techniques to predict carbon fluxes from biophysical data. However, the field lacks a standardized dataset to promote comparisons between models. To address this gap, we present CarbonSense, the first machine learning-ready dataset for DDCFM. CarbonSense integrates measured carbon fluxes, meteorological predictors, and satellite imagery from 385 locations across the globe, offering comprehensive coverage and facilitating robust model training. Additionally, we provide a baseline model using a current state-of-the-art DDCFM approach and a novel transformer based model. Our experiments illustrate the potential gains that multimodal deep learning techniques can bring to this domain. By providing these resources, we aim to lower the barrier to entry for other deep learning researchers to develop new models and drive new advances in carbon flux modelling.
comment: 9 content pages, 11 reference pages, 9 appendix pages
♻ ☆ HybridNorm: Towards Stable and Efficient Transformer Training via Hybrid Normalization
Transformers have become the de facto architecture for a wide range of machine learning tasks, particularly in large language models (LLMs). Despite their remarkable performance, challenges remain in training deep transformer networks, especially regarding the location of layer normalization. While Pre-Norm structures facilitate easier training due to their more prominent identity path, they often yield suboptimal performance compared to Post-Norm. In this paper, we propose $\textbf{HybridNorm}$, a straightforward yet effective hybrid normalization strategy that integrates the advantages of both Pre-Norm and Post-Norm approaches. Specifically, HybridNorm employs QKV normalization within the attention mechanism and Post-Norm in the feed-forward network (FFN) of each transformer block. This design not only stabilizes training but also enhances performance, particularly in the context of LLMs. Comprehensive experiments in both dense and sparse architectures show that HybridNorm consistently outperforms both Pre-Norm and Post-Norm approaches, achieving state-of-the-art results across various benchmarks. These findings highlight the potential of HybridNorm as a more stable and effective technique for improving the training and performance of deep transformer models. Code is available at https://github.com/BryceZhuo/HybridNorm.
♻ ☆ MF-CLIP: Leveraging CLIP as Surrogate Models for No-box Adversarial Attacks
The vulnerability of Deep Neural Networks (DNNs) to adversarial attacks poses a significant challenge to their deployment in safety-critical applications. While extensive research has addressed various attack scenarios, the no-box attack setting where adversaries have no prior knowledge, including access to training data of the target model, remains relatively underexplored despite its practical relevance. This work presents a systematic investigation into leveraging large-scale Vision-Language Models (VLMs), particularly CLIP, as surrogate models for executing no-box attacks. Our theoretical and empirical analyses reveal a key limitation in the execution of no-box attacks stemming from insufficient discriminative capabilities for direct application of vanilla CLIP as a surrogate model. To address this limitation, we propose MF-CLIP: a novel framework that enhances CLIP's effectiveness as a surrogate model through margin-aware feature space optimization. Comprehensive evaluations across diverse architectures and datasets demonstrate that MF-CLIP substantially advances the state-of-the-art in no-box attacks, surpassing existing baselines by 15.23% on standard models and achieving a 9.52% improvement on adversarially trained models. Our code will be made publicly available to facilitate reproducibility and future research in this direction.
♻ ☆ MaRS: A Fast Sampler for Mean Reverting Diffusion based on ODE and SDE Solvers ICLR 2025
In applications of diffusion models, controllable generation is of practical significance, but is also challenging. Current methods for controllable generation primarily focus on modifying the score function of diffusion models, while Mean Reverting (MR) Diffusion directly modifies the structure of the stochastic differential equation (SDE), making the incorporation of image conditions simpler and more natural. However, current training-free fast samplers are not directly applicable to MR Diffusion. And thus MR Diffusion requires hundreds of NFEs (number of function evaluations) to obtain high-quality samples. In this paper, we propose a new algorithm named MaRS (MR Sampler) to reduce the sampling NFEs of MR Diffusion. We solve the reverse-time SDE and the probability flow ordinary differential equation (PF-ODE) associated with MR Diffusion, and derive semi-analytical solutions. The solutions consist of an analytical function and an integral parameterized by a neural network. Based on this solution, we can generate high-quality samples in fewer steps. Our approach does not require training and supports all mainstream parameterizations, including noise prediction, data prediction and velocity prediction. Extensive experiments demonstrate that MR Sampler maintains high sampling quality with a speedup of 10 to 20 times across ten different image restoration tasks. Our algorithm accelerates the sampling procedure of MR Diffusion, making it more practical in controllable generation.
comment: Accepted by ICLR 2025
♻ ☆ VeriMind: Agentic LLM for Automated Verilog Generation with a Novel Evaluation Metric
Designing Verilog modules requires meticulous attention to correctness, efficiency, and adherence to design specifications. However, manually writing Verilog code remains a complex and time-consuming task that demands both expert knowledge and iterative refinement. Leveraging recent advancements in large language models (LLMs) and their structured text generation capabilities, we propose VeriMind, an agentic LLM framework for Verilog code generation that significantly automates and optimizes the synthesis process. Unlike traditional LLM-based code generators, VeriMind employs a structured reasoning approach: given a user-provided prompt describing design requirements, the system first formulates a detailed train of thought before the final Verilog code is generated. This multi-step methodology enhances interpretability, accuracy, and adaptability in hardware design. In addition, we introduce a novel evaluation metric-pass@ARC-which combines the conventional pass@k measure with Average Refinement Cycles (ARC) to capture both success rate and the efficiency of iterative refinement. Experimental results on diverse hardware design tasks demonstrated that our approach achieved up to $8.3\%$ improvement on pass@k metric and $8.1\%$ on pass@ARC metric. These findings underscore the transformative potential of agentic LLMs in automated hardware design, RTL development, and digital system synthesis.
♻ ☆ Implementing An Artificial Quantum Perceptron
A Perceptron is a fundamental building block of a neural network. The flexibility and scalability of perceptron make it ubiquitous in building intelligent systems. Studies have shown the efficacy of a single neuron in making intelligent decisions. Here, we examined and compared two perceptrons with distinct mechanisms, and developed a quantum version of one of those perceptrons. As a part of this modeling, we implemented the quantum circuit for an artificial perception, generated a dataset, and simulated the training. Through these experiments, we show that there is an exponential growth advantage and test different qubit versions. Our findings show that this quantum model of an individual perceptron can be used as a pattern classifier. For the second type of model, we provide an understanding to design and simulate a spike-dependent quantum perceptron. Our code is available at https://github.com/ashutosh1919/quantum-perceptron
♻ ☆ RankCLIP: Ranking-Consistent Language-Image Pretraining
Self-supervised contrastive learning models, such as CLIP, have set new benchmarks for vision-language models in many downstream tasks. However, their dependency on rigid one-to-one mappings overlooks the complex and often multifaceted relationships between and within texts and images. To this end, we introduce RankCLIP, a novel pre-training method that extends beyond the rigid one-to-one matching framework of CLIP and its variants. By extending the traditional pair-wise loss to list-wise, and leveraging both in-modal and cross-modal ranking consistency, RankCLIP improves the alignment process, enabling it to capture the nuanced many-to-many relationships between and within each modality. Through comprehensive experiments, we demonstrate the effectiveness of RankCLIP in various downstream tasks, notably achieving significant gains in zero-shot classifications over state-of-the-art methods, underscoring the importance of this enhanced learning process.
comment: Code and model checkpoints are available at https://github.com/Jam1ezhang/RankCLIP
♻ ☆ Benchmarking Post-Training Quantization in LLMs: Comprehensive Taxonomy, Unified Evaluation, and Comparative Analysis
Post-training Quantization (PTQ) technique has been extensively adopted for large language models (LLMs) compression owing to its efficiency and low resource requirement. However, current research lacks a in-depth analysis of the superior and applicable scenarios of each PTQ strategy. In addition, existing algorithms focus primarily on performance, overlooking the trade-off among model size, performance, and quantization bitwidth. To mitigate these confusions, we provide a novel benchmark for LLMs PTQ in this paper. Firstly, in order to support our benchmark, we propose a comprehensive taxonomy for existing mainstream methods by scrutinizing their computational strategies (e.g., optimization-based, compensation-based, etc.). Then, we conduct extensive experiments with the baseline within each class, covering models with various sizes (7B-70B), bitwidths, training levels (LLaMA1/2/3/3.1), architectures (Mixtral, DeepSeekMoE and Mamba) and modality (LLaVA1.5 and VILA1.5) on a wide range of evaluation metrics.Through comparative analysis on the results, we summarize the superior of each PTQ strategy and modelsize-bitwidth trade-off considering the performance. For example, our benchmark reveals that compensation-based technique demonstrates outstanding cross-architecture robustness and extremely low-bit PTQ for ultra large models should be reexamined. Finally, we further accordingly claim that a practical combination of compensation and other PTQ strategy can achieve SOTA various robustness. We believe that our benchmark will provide valuable recommendations for the deployment of LLMs and future research on PTQ approaches.We conduct an repository for our benchmark at https://github.com/zjq0455/PTQ_Benchmark.
comment: 17 pages, 3 fugures
♻ ☆ DiTEC-WDN: A Large-Scale Dataset of Hydraulic Scenarios across Multiple Water Distribution Networks
Privacy restrictions hinder the sharing of real-world Water Distribution Network (WDN) models, limiting the application of emerging data-driven machine learning, which typically requires extensive observations. To address this challenge, we propose the dataset DiTEC-WDN that comprises 36,000 unique scenarios simulated over either short-term (24 hours) or long-term (1 year) periods. We constructed this dataset using an automated pipeline that optimizes crucial parameters (e.g., pressure, flow rate, and demand patterns), facilitates large-scale simulations, and records discrete, synthetic but hydraulically realistic states under standard conditions via rule validation and post-hoc analysis. With a total of 228 million generated graph-based states, DiTEC-WDN can support a variety of machine-learning tasks, including graph-level, node-level, and link-level regression, as well as time-series forecasting. This contribution, released under a public license, encourages open scientific research in the critical water sector, eliminates the risk of exposing sensitive data, and fulfills the need for a large-scale water distribution network benchmark for study comparisons and scenario analysis.
comment: Submitted to Nature Scientific Data. Huy Truong and Andr\'es Tello contributed equally to this work. For the dataset, see https://huggingface.co/datasets/rugds/ditec-wdn
♻ ☆ Understanding Model Calibration -- A gentle introduction and visual exploration of calibration and the expected calibration error (ECE)
To be considered reliable, a model must be calibrated so that its confidence in each decision closely reflects its true outcome. In this blogpost we'll take a look at the most commonly used definition for calibration and then dive into a frequently used evaluation measure for model calibration. We'll then cover some of the drawbacks of this measure and how these surfaced the need for additional notions of calibration, which require their own new evaluation measures. This post is not intended to be an in-depth dissection of all works on calibration, nor does it focus on how to calibrate models. Instead, it is meant to provide a gentle introduction to the different notions and their evaluation measures as well as to re-highlight some issues with a measure that is still widely used to evaluate calibration.
♻ ☆ CholecTrack20: A Multi-Perspective Tracking Dataset for Surgical Tools CVPR 2025
Tool tracking in surgical videos is essential for advancing computer-assisted interventions, such as skill assessment, safety zone estimation, and human-machine collaboration. However, the lack of context-rich datasets limits AI applications in this field. Existing datasets rely on overly generic tracking formalizations that fail to capture surgical-specific dynamics, such as tools moving out of the camera's view or exiting the body. This results in less clinically relevant trajectories and a lack of flexibility for real-world surgical applications. Methods trained on these datasets often struggle with visual challenges such as smoke, reflection, and bleeding, further exposing the limitations of current approaches. We introduce CholecTrack20, a specialized dataset for multi-class, multi-tool tracking in surgical procedures. It redefines tracking formalization with three perspectives: (i) intraoperative, (ii) intracorporeal, and (iii) visibility, enabling adaptable and clinically meaningful tool trajectories. The dataset comprises 20 full-length surgical videos, annotated at 1 fps, yielding over 35K frames and 65K labeled tool instances. Annotations include spatial location, category, identity, operator, phase, and scene visual challenge. Benchmarking state-of-the-art methods on CholecTrack20 reveals significant performance gaps, with current approaches (< 45\% HOTA) failing to meet the accuracy required for clinical translation. These findings motivate the need for advanced and intuitive tracking algorithms and establish CholecTrack20 as a foundation for developing robust AI-driven surgical assistance systems.
comment: Surgical tool tracking dataset paper, 11 pages, 10 figures, 3 tables, CVPR 2025
♻ ☆ Believing is Seeing: Unobserved Object Detection using Generative Models
Can objects that are not visible in an image -- but are in the vicinity of the camera -- be detected? This study introduces the novel tasks of 2D, 2.5D and 3D unobserved object detection for predicting the location of nearby objects that are occluded or lie outside the image frame. We adapt several state-of-the-art pre-trained generative models to address this task, including 2D and 3D diffusion models and vision-language models, and show that they can be used to infer the presence of objects that are not directly observed. To benchmark this task, we propose a suite of metrics that capture different aspects of performance. Our empirical evaluation on indoor scenes from the RealEstate10k and NYU Depth v2 datasets demonstrate results that motivate the use of generative models for the unobserved object detection task.
comment: IEEE/CVF Computer Vision and Pattern Recognition 2025; 22 pages
♻ ☆ BI-RADS prediction of mammographic masses using uncertainty information extracted from a Bayesian Deep Learning model
The BI_RADS score is a probabilistic reporting tool used by radiologists to express the level of uncertainty in predicting breast cancer based on some morphological features in mammography images. There is a significant variability in describing masses which sometimes leads to BI_RADS misclassification. Using a BI_RADS prediction system is required to support the final radiologist decisions. In this study, the uncertainty information extracted by a Bayesian deep learning model is utilized to predict the BI_RADS score. The investigation results based on the pathology information demonstrate that the f1-scores of the predictions of the radiologist are 42.86%, 48.33% and 48.28%, meanwhile, the f1-scores of the model performance are 73.33%, 59.60% and 59.26% in the BI_RADS 2, 3 and 5 dataset samples, respectively. Also, the model can distinguish malignant from benign samples in the BI_RADS 0 category of the used dataset with an accuracy of 75.86% and correctly identify all malignant samples as BI_RADS 5. The Grad-CAM visualization shows the model pays attention to the morphological features of the lesions. Therefore, this study shows the uncertainty-aware Bayesian Deep Learning model can report his uncertainty about the malignancy of a lesion based on morphological features, like a radiologist.
♻ ☆ No Thing, Nothing: Highlighting Safety-Critical Classes for Robust LiDAR Semantic Segmentation in Adverse Weather CVPR 2025
Existing domain generalization methods for LiDAR semantic segmentation under adverse weather struggle to accurately predict "things" categories compared to "stuff" categories. In typical driving scenes, "things" categories can be dynamic and associated with higher collision risks, making them crucial for safe navigation and planning. Recognizing the importance of "things" categories, we identify their performance drop as a serious bottleneck in existing approaches. We observed that adverse weather induces degradation of semantic-level features and both corruption of local features, leading to a misprediction of "things" as "stuff". To mitigate these corruptions, we suggest our method, NTN - segmeNt Things for No-accident. To address semantic-level feature corruption, we bind each point feature to its superclass, preventing the misprediction of things classes into visually dissimilar categories. Additionally, to enhance robustness against local corruption caused by adverse weather, we define each LiDAR beam as a local region and propose a regularization term that aligns the clean data with its corrupted counterpart in feature space. NTN achieves state-of-the-art performance with a +2.6 mIoU gain on the SemanticKITTI-to-SemanticSTF benchmark and +7.9 mIoU on the SemanticPOSS-to-SemanticSTF benchmark. Notably, NTN achieves a +4.8 and +7.9 mIoU improvement on "things" classes, respectively, highlighting its effectiveness.
comment: 18 pages, accepted in CVPR 2025
♻ ☆ AutoTRIZ: Automating Engineering Innovation with TRIZ and Large Language Models
Various ideation methods, such as morphological analysis and design-by-analogy, have been developed to aid creative problem-solving and innovation. Among them, the Theory of Inventive Problem Solving (TRIZ) stands out as one of the best-known methods. However, the complexity of TRIZ and its reliance on users' knowledge, experience, and reasoning capabilities limit its practicality. To address this, we introduce AutoTRIZ, an artificial ideation system that integrates Large Language Models (LLMs) to automate and enhance the TRIZ methodology. By leveraging LLMs' vast pre-trained knowledge and advanced reasoning capabilities, AutoTRIZ offers a novel, generative, and interpretable approach to engineering innovation. AutoTRIZ takes a problem statement from the user as its initial input, automatically conduct the TRIZ reasoning process and generates a structured solution report. We demonstrate and evaluate the effectiveness of AutoTRIZ through comparative experiments with textbook cases and a real-world application in the design of a Battery Thermal Management System (BTMS). Moreover, the proposed LLM-based framework holds the potential for extension to automate other knowledge-based ideation methods, such as SCAMPER, Design Heuristics, and Design-by-Analogy, paving the way for a new era of AI-driven innovation tools.
comment: 28 pages, 12 figures
♻ ☆ MERGE$^3$: Efficient Evolutionary Merging on Consumer-grade GPUs
Evolutionary model merging enables the creation of high-performing multi-task models but remains computationally prohibitive for consumer hardware. We introduce MERGE$^3$, an efficient framework that makes evolutionary merging feasible on a single GPU by reducing fitness computation costs 50$\times$ while preserving performance. MERGE$^3$ achieves this by Extracting a reduced dataset for evaluation, Estimating model abilities using Item Response Theory (IRT), and Evolving optimal merges via IRT-based performance estimators. Our method enables state-of-the-art multilingual and cross-lingual merging, transferring knowledge across languages with significantly lower computational overhead. We provide theoretical guarantees and an open-source library, democratizing high-quality model merging.
comment: 19 pages, 13 figures
♻ ☆ GUI-World: A Video Benchmark and Dataset for Multimodal GUI-oriented Understanding ICLR 2025
Recently, Multimodal Large Language Models (MLLMs) have been used as agents to control keyboard and mouse inputs by directly perceiving the Graphical User Interface (GUI) and generating corresponding commands. However, current agents primarily demonstrate strong understanding capabilities in static environments and are mainly applied to relatively simple domains, such as Web or mobile interfaces. We argue that a robust GUI agent should be capable of perceiving temporal information on the GUI, including dynamic Web content and multi-step tasks. Additionally, it should possess a comprehensive understanding of various GUI scenarios, including desktop software and multi-window interactions. To this end, this paper introduces a new dataset, termed GUI-World, which features meticulously crafted Human-MLLM annotations, extensively covering six GUI scenarios and eight types of GUI-oriented questions in three formats. We evaluate the capabilities of current state-of-the-art MLLMs, including Image LLMs and Video LLMs, in understanding various types of GUI content, especially dynamic and sequential content. Our findings reveal that current models struggle with dynamic GUI content without manually annotated keyframes or operation history. On the other hand, Video LLMs fall short in all GUI-oriented tasks given the sparse GUI video dataset. Therefore, we take the initial step of leveraging a fine-tuned Video LLM, GUI-Vid, as a GUI-oriented assistant, demonstrating an improved understanding of various GUI tasks. However, due to the limitations in the performance of base LLMs, we conclude that using video LLMs as GUI agents remains a significant challenge. We believe our work provides valuable insights for future research in dynamic GUI content understanding. All the dataset and code are publicly available at: https://gui-world.github.io.
comment: Accepted by ICLR 2025
♻ ☆ ConCodeEval: Evaluating Large Language Models for Code Constraints in Domain-Specific Languages
Recent work shows Large Language Models (LLMs) struggle to understand natural language constraints for various text generation tasks in zero- and few-shot settings. While, in the code domain, there is wide usage of constraints in code format to maintain the integrity of code written in Domain-Specific Languages (DSLs) like JSON and YAML which are widely used for system-level programming tasks in enterprises. Given that LLMs are increasingly used for system-level code tasks, evaluating if they can comprehend these code constraints is crucial. However, no work has been done to evaluate their controllability over code constraints. Hence, we introduce ConCodeEval, a first-of-its-kind benchmark having two novel tasks for code constraints across five representations. Our findings suggest that language models struggle with code constraints. Code languages that perform excellently for normal code tasks do not perform well when the same languages represent fine-grained constraints.
♻ ☆ CompMarkGS: Robust Watermarking for Compression 3D Gaussian Splatting
3D Gaussian Splatting (3DGS) enables rapid differentiable rendering for 3D reconstruction and novel view synthesis, leading to its widespread commercial use. Consequently, copyright protection via watermarking has become critical. However, because 3DGS relies on millions of Gaussians, which require gigabytes of storage, efficient transfer and storage require compression. Existing 3DGS watermarking methods are vulnerable to quantization-based compression, often resulting in the loss of the embedded watermark. To address this challenge, we propose a novel watermarking method that ensures watermark robustness after model compression while maintaining high rendering quality. In detail, we incorporate a quantization distortion layer that simulates compression during training, preserving the watermark under quantization-based compression. Also, we propose a learnable watermark embedding feature that embeds the watermark into the anchor feature, ensuring structural consistency and seamless integration into the 3D scene. Furthermore, we present a frequency-aware anchor growing mechanism to enhance image quality in high-frequency regions by effectively identifying Guassians within these regions. Experimental results confirm that our method preserves the watermark and maintains superior image quality under high compression, validating it as a promising approach for a secure 3DGS model.
comment: 23 pages, 17 figures
♻ ☆ Toward a method for LLM-enabled Indoor Navigation
Indoor navigation presents unique challenges due to complex layouts, lack of GPS signals, and accessibility concerns. Existing solutions often struggle with real-time adaptability and user-specific needs. In this work, we explore the potential of a Large Language Model (LLM), i.e., ChatGPT, to generate natural, context-aware navigation instructions from indoor map images. We design and evaluate test cases across different real-world environments, analyzing the effectiveness of LLMs in interpreting spatial layouts, handling user constraints, and planning efficient routes. Our findings demonstrate the potential of LLMs for supporting personalized indoor navigation, with an average of 50.54% correct indications and a maximum of 77.78%. The results do not appear to depend on the complexity of the layout or the complexity of the expected path, but rather on the number of points of interest and the abundance of visual information, which negatively affect the performance.
comment: 7 pages, 3 figures, 5 tables
♻ ☆ Survival Analysis with Machine Learning for Predicting Li-ion Battery Remaining Useful Life
The accurate prediction of RUL for lithium-ion batteries is crucial for enhancing the reliability and longevity of energy storage systems. Traditional methods for RUL prediction often struggle with issues such as data sparsity, varying battery chemistries, and the inability to capture complex degradation patterns over time. In this study, we propose a survival analysis-based framework combined with deep learning models to predict the RUL of lithium-ion batteries. Specifically, we utilize five advanced models: the Cox-type models (Cox, CoxPH, and CoxTime) and two machine-learning-based models (DeepHit and MTLR). These models address the challenges of accurate RUL estimation by transforming raw time-series battery data into survival data, including key degradation indicators such as voltage, current, and internal resistance. Advanced feature extraction techniques enhance the model's robustness in diverse real-world scenarios, including varying charging conditions and battery chemistries. Our models are tested using 10-fold cross-validation, ensuring generalizability and minimizing overfitting. Experimental results show that our survival-based framework significantly improves RUL prediction accuracy compared to traditional methods, providing a reliable tool for battery management and maintenance optimization. This study contributes to the advancement of predictive maintenance in battery technology, offering valuable insights for both researchers and industry practitioners aiming to enhance the operational lifespan of lithium-ion batteries.
♻ ☆ Active Learning for Neural PDE Solvers
Solving partial differential equations (PDEs) is a fundamental problem in science and engineering. While neural PDE solvers can be more efficient than established numerical solvers, they often require large amounts of training data that is costly to obtain. Active learning (AL) could help surrogate models reach the same accuracy with smaller training sets by querying classical solvers with more informative initial conditions and PDE parameters. While AL is more common in other domains, it has yet to be studied extensively for neural PDE solvers. To bridge this gap, we introduce AL4PDE, a modular and extensible active learning benchmark. It provides multiple parametric PDEs and state-of-the-art surrogate models for the solver-in-the-loop setting, enabling the evaluation of existing and the development of new AL methods for neural PDE solving. We use the benchmark to evaluate batch active learning algorithms such as uncertainty- and feature-based methods. We show that AL reduces the average error by up to 71% compared to random sampling and significantly reduces worst-case errors. Moreover, AL generates similar datasets across repeated runs, with consistent distributions over the PDE parameters and initial conditions. The acquired datasets are reusable, providing benefits for surrogate models not involved in the data generation.
♻ ☆ Agent-R: Training Language Model Agents to Reflect via Iterative Self-Training
Large Language Models (LLMs) agents are increasingly pivotal for addressing complex tasks in interactive environments. Existing work mainly focuses on enhancing performance through behavior cloning from stronger experts, yet such approaches often falter in real-world applications, mainly due to the inability to recover from errors. However, step-level critique data is difficult and expensive to collect. Automating and dynamically constructing self-critique datasets is thus crucial to empowering models with intelligent agent capabilities. In this work, we propose an iterative self-training framework, Agent-R, that enables language Agent to Reflect on the fly. Unlike traditional methods that reward or penalize actions based on correctness, Agent-R leverages MCTS to construct training data that recover correct trajectories from erroneous ones. A key challenge of agent reflection lies in the necessity for timely revision rather than waiting until the end of a rollout. To address this, we introduce a model-guided critique construction mechanism: the actor model identifies the first error step (within its current capability) in a failed trajectory. Starting from it, we splice it with the adjacent correct path, which shares the same parent node in the tree. This strategy enables the model to learn reflection based on its current policy, therefore yielding better learning efficiency. To further explore the scalability of this self-improvement paradigm, we investigate iterative refinement of both error correction capabilities and dataset construction. Our findings demonstrate that Agent-R continuously improves the model's ability to recover from errors and enables timely error correction. Experiments on three interactive environments show that Agent-R effectively equips agents to correct erroneous actions while avoiding loops, achieving superior performance compared to baseline methods (+5.59%).
♻ ☆ Existential Notation3 Logic
In this paper, we delve into Notation3 Logic (N3), an extension of RDF, which empowers users to craft rules introducing fresh blank nodes to RDF graphs. This capability is pivotal in various applications such as ontology mapping, given the ubiquitous presence of blank nodes directly or in auxiliary constructs across the Web. However, the availability of fast N3 reasoners fully supporting blank node introduction remains limited. Conversely, engines like VLog or Nemo, though not explicitly designed for Semantic Web rule formats, cater to analogous constructs, namely existential rules. We investigate the correlation between N3 rules featuring blank nodes in their heads and existential rules. We pinpoint a subset of N3 that seamlessly translates to existential rules and establish a mapping preserving the equivalence of N3 formulae. To showcase the potential benefits of this translation in N3 reasoning, we implement this mapping and compare the performance of N3 reasoners like EYE and cwm against VLog and Nemo, both on native N3 rules and their translated counterparts. Our findings reveal that existential rule reasoners excel in scenarios with abundant facts, while the EYE reasoner demonstrates exceptional speed in managing a high volume of dependent rules. Additionally to the original conference version of this paper, we include all proofs of the theorems and introduce a new section dedicated to N3 lists featuring built-in functions and how they are implemented in existential rules. Adding lists to our translation/framework gives interesting insights on related design decisions influencing the standardization of N3.
comment: Under consideration in Theory and Practice of Logic Programming (TPLP)
♻ ☆ Twin Transition or Competing Interests? Validation of the Artificial Intelligence and Sustainability Perceptions Inventory (AISPI)
As artificial intelligence (AI) and sustainability initiatives increasingly intersect, understanding public perceptions of their relationship becomes crucial for successful implementation. However, no validated instrument exists to measure these specific perceptions. This paper presents the development and validation of the Artificial Intelligence and Sustainability Perceptions Inventory (AISPI), a novel 13-item instrument measuring how individuals view the relationship between AI advancement and environmental sustainability. Through factor analysis (N=105), we identified two distinct dimensions: Twin Transition and Competing Interests. The instrument demonstrated strong reliability (alpha=.89) and construct validity through correlations with established measures of AI and sustainability attitudes. Our findings suggest that individuals can simultaneously recognize both synergies and tensions in the AI-sustainability relationship, offering important implications for researchers and practitioners working at this critical intersection. This work provides a foundational tool for future research on public perceptions of AI's role in sustainable development.
comment: CHI 2025 Late Breaking Work
♻ ☆ Distilled Prompt Learning for Incomplete Multimodal Survival Prediction CVPR2025
The integration of multimodal data including pathology images and gene profiles is widely applied in precise survival prediction. Despite recent advances in multimodal survival models, collecting complete modalities for multimodal fusion still poses a significant challenge, hindering their application in clinical settings. Current approaches tackling incomplete modalities often fall short, as they typically compensate for only a limited part of the knowledge of missing modalities. To address this issue, we propose a Distilled Prompt Learning framework (DisPro) to utilize the strong robustness of Large Language Models (LLMs) to missing modalities, which employs two-stage prompting for compensation of comprehensive information for missing modalities. In the first stage, Unimodal Prompting (UniPro) distills the knowledge distribution of each modality, preparing for supplementing modality-specific knowledge of the missing modality in the subsequent stage. In the second stage, Multimodal Prompting (MultiPro) leverages available modalities as prompts for LLMs to infer the missing modality, which provides modality-common information. Simultaneously, the unimodal knowledge acquired in the first stage is injected into multimodal inference to compensate for the modality-specific knowledge of the missing modality. Extensive experiments covering various missing scenarios demonstrated the superiority of the proposed method. The code is available at https://github.com/Innse/DisPro.
comment: Accepted by CVPR2025
♻ ☆ Adapt-$\infty$: Scalable Continual Multimodal Instruction Tuning via Dynamic Data Selection
Visual instruction datasets from various distributors are released at different times and often contain a significant number of semantically redundant text-image pairs, depending on their task compositions (i.e., skills) or reference sources. This redundancy greatly limits the efficient deployment of continually adaptable multimodal large language models, hindering their ability to refine existing skills and acquire new competencies over time. We reframe the problem of lifelong Instruction Tuning (LiIT) via data selection, where the model automatically selects beneficial samples to learn from earlier and new datasets based on the current state of acquired knowledge in the model. We propose Adapt-$\infty$, a new multi-way and adaptive data selection approach that dynamically balances sample efficiency and effectiveness during LiIT. We first construct pseudo-skill clusters by grouping gradient-based sample vectors. Next, we select the best-performing data selector for each skill cluster from a pool of selector experts, including our newly proposed scoring function, Image Grounding score. This data selector samples a subset of the most important samples from each skill cluster for training. To prevent the continuous increase in the size of the dataset pool during LiIT, we introduce a cluster-wise permanent data pruning strategy to remove the most semantically redundant samples from each cluster, keeping computational requirements manageable. We validate the effectiveness and efficiency of Adapt-$\infty$ over a sequence of multimodal instruction tuning datasets with various tasks, including (Knowledge) VQA, multilingual, grounding, reasoning, language-only, and multi-image comprehension. Training with samples selected by Adapt-$\infty$ alleviates catastrophic forgetting, especially for rare tasks, and promotes forward transfer across the continuum using only a fraction of the original data.
comment: First two authors contributed equally. Code: https://github.com/adymaharana/adapt-inf
♻ ☆ Human-like conceptual representations emerge from language prediction
People acquire concepts through rich physical and social experiences and use them to understand the world. In contrast, large language models (LLMs), trained exclusively through next-token prediction over language data, exhibit remarkably human-like behaviors. Are these models developing concepts akin to humans, and if so, how are such concepts represented and organized? To address these questions, we reframed the classic reverse dictionary task to simulate human concept inference in context and investigated the emergence of human-like conceptual representations within LLMs. Our results demonstrate that LLMs can flexibly derive concepts from linguistic descriptions in relation to contextual cues about other concepts. The derived representations converged towards a shared, context-independent structure that effectively predicted human behavior across key psychological phenomena, including computation of similarities, categories and semantic scales. Moreover, these representations aligned well with neural activity patterns in the human brain, even in response to visual rather than linguistic stimuli, providing evidence for biological plausibility. These findings establish that structured, human-like conceptual representations can naturally emerge from language prediction without real-world grounding. More broadly, our work positions LLMs as promising computational tools for understanding complex human cognition and paves the way for better alignment between artificial and human intelligence.
comment: 51 pages
♻ ☆ Enhancing Multimodal Sentiment Analysis for Missing Modality through Self-Distillation and Unified Modality Cross-Attention
In multimodal sentiment analysis, collecting text data is often more challenging than video or audio due to higher annotation costs and inconsistent automatic speech recognition (ASR) quality. To address this challenge, our study has developed a robust model that effectively integrates multimodal sentiment information, even in the absence of text modality. Specifically, we have developed a Double-Flow Self-Distillation Framework, including Unified Modality Cross-Attention (UMCA) and Modality Imagination Autoencoder (MIA), which excels at processing both scenarios with complete modalities and those with missing text modality. In detail, when the text modality is missing, our framework uses the LLM-based model to simulate the text representation from the audio modality, while the MIA module supplements information from the other two modalities to make the simulated text representation similar to the real text representation. To further align the simulated and real representations, and to enable the model to capture the continuous nature of sample orders in sentiment valence regression tasks, we have also introduced the Rank-N Contrast (RNC) loss function. When testing on the CMU-MOSEI, our model achieved outstanding performance on MAE and significantly outperformed other models when text modality is missing. The code is available at: https://github.com/WarmCongee/SDUMC
♻ ☆ JanusFlow: Harmonizing Autoregression and Rectified Flow for Unified Multimodal Understanding and Generation CVPR 2025
We present JanusFlow, a powerful framework that unifies image understanding and generation in a single model. JanusFlow introduces a minimalist architecture that integrates autoregressive language models with rectified flow, a state-of-the-art method in generative modeling. Our key finding demonstrates that rectified flow can be straightforwardly trained within the large language model framework, eliminating the need for complex architectural modifications. To further improve the performance of our unified model, we adopt two key strategies: (i) decoupling the understanding and generation encoders, and (ii) aligning their representations during unified training. Extensive experiments show that JanusFlow achieves comparable or superior performance to specialized models in their respective domains, while significantly outperforming existing unified approaches across standard benchmarks. This work represents a step toward more efficient and versatile vision-language models.
comment: Accepted by CVPR 2025
♻ ☆ Humanoid Policy ~ Human Policy
Training manipulation policies for humanoid robots with diverse data enhances their robustness and generalization across tasks and platforms. However, learning solely from robot demonstrations is labor-intensive, requiring expensive tele-operated data collection which is difficult to scale. This paper investigates a more scalable data source, egocentric human demonstrations, to serve as cross-embodiment training data for robot learning. We mitigate the embodiment gap between humanoids and humans from both the data and modeling perspectives. We collect an egocentric task-oriented dataset (PH2D) that is directly aligned with humanoid manipulation demonstrations. We then train a human-humanoid behavior policy, which we term Human Action Transformer (HAT). The state-action space of HAT is unified for both humans and humanoid robots and can be differentiably retargeted to robot actions. Co-trained with smaller-scale robot data, HAT directly models humanoid robots and humans as different embodiments without additional supervision. We show that human data improves both generalization and robustness of HAT with significantly better data collection efficiency. Code and data: https://human-as-robot.github.io/
comment: Code and data: https://human-as-robot.github.io/
♻ ☆ CSCE: Boosting LLM Reasoning by Simultaneous Enhancing of Causal Significance and Consistency
Chain-based reasoning methods like chain of thought (CoT) play a rising role in solving reasoning tasks for large language models (LLMs). However, the causal hallucinations between a step of reasoning and corresponding state transitions are becoming a significant obstacle to advancing LLMs' reasoning capabilities, especially in long-range reasoning tasks. This paper proposes a non-chain-based reasoning framework for simultaneous consideration of causal significance and consistency, i.e., the Causal Significance and Consistency Enhancer (CSCE). We customize LLM's loss function utilizing treatment effect assessments to enhance its reasoning ability from two aspects: causal significance and consistency. This ensures that the model captures essential causal relationships and maintains robust and consistent performance across various scenarios. Additionally, we transform the reasoning process from the cascading multiple one-step reasoning commonly used in Chain-Based methods, like CoT, to a causal-enhanced method that outputs the entire reasoning process in one go, further improving the model's reasoning efficiency. Extensive experiments show that our method improves both the reasoning success rate and speed. These improvements further demonstrate that non-chain-based methods can also aid LLMs in completing reasoning tasks.
comment: 6 pages,4 figures. This paper has been accepted for presentation at IEEE International Conference on Multimedia & Expo 2025
♻ ☆ Cross-domain Random Pre-training with Prototypes for Reinforcement Learning
This work has been submitted to the IEEE for possible publication. Copyright may be transferred without notice, after which this version may no longer be accessible. Unsupervised cross-domain Reinforcement Learning (RL) pre-training shows great potential for challenging continuous visual control but poses a big challenge. In this paper, we propose \textbf{C}ross-domain \textbf{R}andom \textbf{P}re-\textbf{T}raining with \textbf{pro}totypes (CRPTpro), a novel, efficient, and effective self-supervised cross-domain RL pre-training framework. CRPTpro decouples data sampling from encoder pre-training, proposing decoupled random collection to easily and quickly generate a qualified cross-domain pre-training dataset. Moreover, a novel prototypical self-supervised algorithm is proposed to pre-train an effective visual encoder that is generic across different domains. Without finetuning, the cross-domain encoder can be implemented for challenging downstream tasks defined in different domains, either seen or unseen. Compared with recent advanced methods, CRPTpro achieves better performance on downstream policy learning without extra training on exploration agents for data collection, greatly reducing the burden of pre-training. We conduct extensive experiments across eight challenging continuous visual-control domains, including balance control, robot locomotion, and manipulation. CRPTpro significantly outperforms the next best Proto-RL(C) on 11/12 cross-domain downstream tasks with only 54.5\% wall-clock pre-training time, exhibiting state-of-the-art pre-training performance with greatly improved pre-training efficiency.
comment: This work has been submitted to the IEEE for possible publication
♻ ☆ Divide and Conquer: Heterogeneous Noise Integration for Diffusion-based Adversarial Purification
Existing diffusion-based purification methods aim to disrupt adversarial perturbations by introducing a certain amount of noise through a forward diffusion process, followed by a reverse process to recover clean examples. However, this approach is fundamentally flawed: the uniform operation of the forward process across all pixels compromises normal pixels while attempting to combat adversarial perturbations, resulting in the target model producing incorrect predictions. Simply relying on low-intensity noise is insufficient for effective defense. To address this critical issue, we implement a heterogeneous purification strategy grounded in the interpretability of neural networks. Our method decisively applies higher-intensity noise to specific pixels that the target model focuses on while the remaining pixels are subjected to only low-intensity noise. This requirement motivates us to redesign the sampling process of the diffusion model, allowing for the effective removal of varying noise levels. Furthermore, to evaluate our method against strong adaptative attack, our proposed method sharply reduces time cost and memory usage through a single-step resampling. The empirical evidence from extensive experiments across three datasets demonstrates that our method outperforms most current adversarial training and purification techniques by a substantial margin.
♻ ☆ Lessons and Insights from a Unifying Study of Parameter-Efficient Fine-Tuning (PEFT) in Visual Recognition CVPR 2025
Parameter-efficient fine-tuning (PEFT) has attracted significant attention due to the growth of pre-trained model sizes and the need to fine-tune (FT) them for superior downstream performance. Despite a surge in new PEFT methods, a systematic study to understand their performance and suitable application scenarios is lacking, leaving questions like "when to apply PEFT" and "which method to use" largely unanswered, especially in visual recognition. In this paper, we conduct a unifying empirical study of representative PEFT methods with Vision Transformers. We systematically tune their hyper-parameters to fairly compare their accuracy on downstream tasks. Our study offers a practical user guide and unveils several new insights. First, if tuned carefully, different PEFT methods achieve similar accuracy in the low-shot benchmark VTAB-1K. This includes simple approaches like FT the bias terms that were reported inferior. Second, despite similar accuracy, we find that PEFT methods make different mistakes and high-confidence predictions, likely due to their different inductive biases. Such an inconsistency (or complementarity) opens up the opportunity for ensemble methods, and we make preliminary attempts at this. Third, going beyond the commonly used low-shot tasks, we find that PEFT is also useful in many-shot regimes, achieving comparable or better accuracy than full FT while using significantly fewer parameters. Lastly, we investigate PEFT's ability to preserve a pre-trained model's robustness to distribution shifts (e.g., CLIP). Perhaps not surprisingly, PEFT approaches outperform full FT alone. However, with weight-space ensembles, full FT can better balance target distribution and distribution shift performance, suggesting a future research direction for robust PEFT.
comment: CVPR 2025. The code is available at https://github.com/OSU-MLB/ViT_PEFT_Vision
♻ ☆ LaMOuR: Leveraging Language Models for Out-of-Distribution Recovery in Reinforcement Learning
Deep Reinforcement Learning (DRL) has demonstrated strong performance in robotic control but remains susceptible to out-of-distribution (OOD) states, often resulting in unreliable actions and task failure. While previous methods have focused on minimizing or preventing OOD occurrences, they largely neglect recovery once an agent encounters such states. Although the latest research has attempted to address this by guiding agents back to in-distribution states, their reliance on uncertainty estimation hinders scalability in complex environments. To overcome this limitation, we introduce Language Models for Out-of-Distribution Recovery (LaMOuR), which enables recovery learning without relying on uncertainty estimation. LaMOuR generates dense reward codes that guide the agent back to a state where it can successfully perform its original task, leveraging the capabilities of LVLMs in image description, logical reasoning, and code generation. Experimental results show that LaMOuR substantially enhances recovery efficiency across diverse locomotion tasks and even generalizes effectively to complex environments, including humanoid locomotion and mobile manipulation, where existing methods struggle. The code and supplementary materials are available at https://lamour-rl.github.io/.
comment: 14 pages, 17 figures
♻ ☆ DreamText: High Fidelity Scene Text Synthesis
Scene text synthesis involves rendering specified texts onto arbitrary images. Current methods typically formulate this task in an end-to-end manner but lack effective character-level guidance during training. Besides, their text encoders, pre-trained on a single font type, struggle to adapt to the diverse font styles encountered in practical applications. Consequently, these methods suffer from character distortion, repetition, and absence, particularly in polystylistic scenarios. To this end, this paper proposes DreamText for high-fidelity scene text synthesis. Our key idea is to reconstruct the diffusion training process, introducing more refined guidance tailored to this task, to expose and rectify the model's attention at the character level and strengthen its learning of text regions. This transformation poses a hybrid optimization challenge, involving both discrete and continuous variables. To effectively tackle this challenge, we employ a heuristic alternate optimization strategy. Meanwhile, we jointly train the text encoder and generator to comprehensively learn and utilize the diverse font present in the training dataset. This joint training is seamlessly integrated into the alternate optimization process, fostering a synergistic relationship between learning character embedding and re-estimating character attention. Specifically, in each step, we first encode potential character-generated position information from cross-attention maps into latent character masks. These masks are then utilized to update the representation of specific characters in the current step, which, in turn, enables the generator to correct the character's attention in the subsequent steps. Both qualitative and quantitative results demonstrate the superiority of our method to the state of the art.
comment: Code: https://github.com/CodeGoat24/DreamText, Project page: https://codegoat24.github.io/DreamText/
♻ ☆ Reinforcement Learning: A Comprehensive Overview
This manuscript gives a big-picture, up-to-date overview of the field of (deep) reinforcement learning and sequential decision making, covering value-based method, policy-gradient methods, model-based methods, and various other topics (e.g., multi-agent RL, RL+LLMs, and RL+inference).
♻ ☆ Implementing Fairness in AI Classification: The Role of Explainability
In this paper, we propose a philosophical and experimental investigation of the problem of AI fairness in classification. We argue that implementing fairness in AI classification involves more work than just operationalizing a fairness metric. It requires establishing the explainability of the classification model chosen and of the principles behind it. Specifically, it involves making the training processes transparent, determining what outcomes the fairness criteria actually produce, and assessing their trade-offs by comparison with closely related models that would lead to a different outcome. To exemplify this methodology, we trained a model and developed a tool for disparity detection and fairness interventions, the package FairDream. While FairDream is set to enforce Demographic Parity, experiments reveal that it fulfills the constraint of Equalized Odds. The algorithm is thus more conservative than the user might expect. To justify this outcome, we first clarify the relation between Demographic Parity and Equalized Odds as fairness criteria. We then explain FairDream's reweighting method and justify the trade-offs reached by FairDream by a benchmark comparison with closely related GridSearch models. We draw conclusions regarding the way in which these explanatory steps can make an AI model trustworthy.
♻ ☆ Parameter Efficient Mamba Tuning via Projector-targeted Diagonal-centric Linear Transformation CVPR 2025
Despite the growing interest in Mamba architecture as a potential replacement for Transformer architecture, parameter-efficient fine-tuning (PEFT) approaches for Mamba remain largely unexplored. In our study, we introduce two key insights-driven strategies for PEFT in Mamba architecture: (1) While state-space models (SSMs) have been regarded as the cornerstone of Mamba architecture, then expected to play a primary role in transfer learning, our findings reveal that Projectors -- not SSMs -- are the predominant contributors to transfer learning. (2) Based on our observation, we propose a novel PEFT method specialized to Mamba architecture: Projector-targeted Diagonal-centric Linear Transformation (ProDiaL). ProDiaL focuses on optimizing only the pretrained Projectors for new tasks through diagonal-centric linear transformation matrices, without directly fine-tuning the Projector weights. This targeted approach allows efficient task adaptation, utilizing less than 1% of the total parameters, and exhibits strong performance across both vision and language Mamba models, highlighting its versatility and effectiveness.
comment: accepted in CVPR 2025
♻ ☆ AnDB: Breaking Boundaries with an AI-Native Database for Universal Semantic Analysis
In this demonstration, we present AnDB, an AI-native database that supports traditional OLTP workloads and innovative AI-driven tasks, enabling unified semantic analysis across structured and unstructured data. While structured data analytics is mature, challenges remain in bridging the semantic gap between user queries and unstructured data. AnDB addresses these issues by leveraging cutting-edge AI-native technologies, allowing users to perform semantic queries using intuitive SQL-like statements without requiring AI expertise. This approach eliminates the ambiguity of traditional text-to-SQL systems and provides a seamless end-to-end optimization for analyzing all data types. AnDB automates query processing by generating multiple execution plans and selecting the optimal one through its optimizer, which balances accuracy, execution time, and financial cost based on user policies and internal optimizing mechanisms. AnDB future-proofs data management infrastructure, empowering users to effectively and efficiently harness the full potential of all kinds of data without starting from scratch.
comment: 4 pages, 5 figures
♻ ☆ Zero-Knowledge Federated Learning: A New Trustworthy and Privacy-Preserving Distributed Learning Paradigm
Federated Learning (FL) has emerged as a promising paradigm in distributed machine learning, enabling collaborative model training while preserving data privacy. However, despite its many advantages, FL still contends with significant challenges -- most notably regarding security and trust. Zero-Knowledge Proofs (ZKPs) offer a potential solution by establishing trust and enhancing system integrity throughout the FL process. Although several studies have explored ZKP-based FL (ZK-FL), a systematic framework and comprehensive analysis are still lacking. This article makes two key contributions. First, we propose a structured ZK-FL framework that categorizes and analyzes the technical roles of ZKPs across various FL stages and tasks. Second, we introduce a novel algorithm, Verifiable Client Selection FL (Veri-CS-FL), which employs ZKPs to refine the client selection process. In Veri-CS-FL, participating clients generate verifiable proofs for the performance metrics of their local models and submit these concise proofs to the server for efficient verification. The server then selects clients with high-quality local models for uploading, subsequently aggregating the contributions from these selected clients. By integrating ZKPs, Veri-CS-FL not only ensures the accuracy of performance metrics but also fortifies trust among participants while enhancing the overall efficiency and security of FL systems.
comment: 7 pages, 5 figures, 1 table
♻ ☆ Characterizing GPU Resilience and Impact on AI/HPC Systems
In this study, we characterize GPU failures in Delta, the current large-scale AI system with over 600 petaflops of peak compute throughput. The system comprises GPU and non-GPU nodes with modern AI accelerators, such as NVIDIA A40, A100, and H100 GPUs. The study uses two and a half years of data on GPU errors. We evaluate the resilience of GPU hardware components to determine the vulnerability of different GPU components to failure and their impact on the GPU and node availability. We measure the key propagation paths in GPU hardware, GPU interconnect (NVLink), and GPU memory. Finally, we evaluate the impact of the observed GPU errors on user jobs. Our key findings are: (i) Contrary to common beliefs, GPU memory is over 30x more reliable than GPU hardware in terms of MTBE (mean time between errors). (ii) The newly introduced GSP (GPU System Processor) is the most vulnerable GPU hardware component. (iii) NVLink errors did not always lead to user job failure, and we attribute it to the underlying error detection and retry mechanisms employed. (iv) We show multiple examples of hardware errors originating from one of the key GPU hardware components, leading to application failure. (v) We project the impact of GPU node availability on larger scales with emulation and find that significant overprovisioning between 5-20% would be necessary to handle GPU failures. If GPU availability were improved to 99.9%, the overprovisioning would be reduced by 4x.
♻ ☆ A Mechanistic Explanatory Strategy for XAI
Despite significant advancements in XAI, scholars note a persistent lack of solid conceptual foundations and integration with broader discourse on scientific explanation. In response, emerging XAI research draws on explanatory strategies from various sciences and philosophy of science literature to fill these gaps. This paper outlines a mechanistic strategy for explaining the functional organization of deep learning systems, situating recent advancements in AI explainability within a broader philosophical context. According to the mechanistic approach, the explanation of opaque AI systems involves identifying mechanisms that drive decision-making. For deep neural networks, this means discerning functionally relevant components - such as neurons, layers, circuits, or activation patterns - and understanding their roles through decomposition, localization, and recomposition. Proof-of-principle case studies from image recognition and language modeling align these theoretical approaches with the latest research from AI labs like OpenAI and Anthropic. The paper suggests that such a systematic approach to studying model organization can reveal elements that individual explainability techniques might miss, fostering more thoroughly explainable AI.
comment: Forthcoming in M\"uller, V. C., Dung, L., L\"ohr, G., & Rumana, A. (Eds.). Philosophy of Artificial Intelligence: The State of the Art, Synthese Library, Springer Nature. Please cite the published version
♻ ☆ Forming Auxiliary High-confident Instance-level Loss to Promote Learning from Label Proportions CVPR 2025
Learning from label proportions (LLP), i.e., a challenging weakly-supervised learning task, aims to train a classifier by using bags of instances and the proportions of classes within bags, rather than annotated labels for each instance. Beyond the traditional bag-level loss, the mainstream methodology of LLP is to incorporate an auxiliary instance-level loss with pseudo-labels formed by predictions. Unfortunately, we empirically observed that the pseudo-labels are are often inaccurate due to over-smoothing, especially for the scenarios with large bag sizes, hurting the classifier induction. To alleviate this problem, we suggest a novel LLP method, namely Learning from Label Proportions with Auxiliary High-confident Instance-level Loss (L^2P-AHIL). Specifically, we propose a dual entropy-based weight (DEW) method to adaptively measure the confidences of pseudo-labels. It simultaneously emphasizes accurate predictions at the bag level and avoids overly smoothed predictions. We then form high-confident instance-level loss with DEW, and jointly optimize it with the bag-level loss in a self-training manner. The experimental results on benchmark datasets show that L^2P-AHIL can surpass the existing baseline methods, and the performance gain can be more significant as the bag size increases. The implementation of our method is available at https://github.com/TianhaoMa5/LLP-AHIL.
comment: Accepted as a conference paper at CVPR 2025
♻ ☆ Leveraging Allophony in Self-Supervised Speech Models for Atypical Pronunciation Assessment NAACL 2025
Allophony refers to the variation in the phonetic realization of a phoneme based on its phonetic environment. Modeling allophones is crucial for atypical pronunciation assessment, which involves distinguishing atypical from typical pronunciations. However, recent phoneme classifier-based approaches often simplify this by treating various realizations as a single phoneme, bypassing the complexity of modeling allophonic variation. Motivated by the acoustic modeling capabilities of frozen self-supervised speech model (S3M) features, we propose MixGoP, a novel approach that leverages Gaussian mixture models to model phoneme distributions with multiple subclusters. Our experiments show that MixGoP achieves state-of-the-art performance across four out of five datasets, including dysarthric and non-native speech. Our analysis further suggests that S3M features capture allophonic variation more effectively than MFCCs and Mel spectrograms, highlighting the benefits of integrating MixGoP with S3M features.
comment: Accepted to NAACL 2025. Codebase available at https://github.com/juice500ml/acoustic-units-for-ood
♻ ☆ Human-AI Interaction Design Standards
The rapid development of artificial intelligence (AI) has significantly transformed human-computer interactions, making it essential to establish robust design standards to ensure effective, ethical, and human-centered AI (HCAI) solutions. Standards serve as the foundation for the adoption of new technologies, and human-AI interaction (HAII) standards are critical to supporting the industrialization of AI technology by following an HCAI approach. These design standards aim to provide clear principles, requirements, and guidelines for designing, developing, deploying, and using AI systems, enhancing the user experience and performance of AI systems. Despite their importance, the creation and adoption of HCAI-based interaction design standards face challenges, including the absence of universal frameworks, the inherent complexity of HAII, and the ethical dilemmas that arise in such systems. This chapter provides a comparative analysis of HAII versus traditional human-computer interaction (HCI) and outlines guiding principles for HCAI-based design. It explores international, regional, national, and industry standards related to HAII design from an HCAI perspective and reviews design guidelines released by leading companies such as Microsoft, Google, and Apple. Additionally, the chapter highlights tools available for implementing HAII standards and presents case studies of human-centered interaction design for AI systems in diverse fields, including healthcare, autonomous vehicles, and customer service. It further examines key challenges in developing HAII standards and suggests future directions for the field. Emphasizing the importance of ongoing collaboration between AI designers, developers, and experts in human factors and HCI, this chapter stresses the need to advance HCAI-based interaction design standards to ensure human-centered AI solutions across various domains.
♻ ☆ Neural Representation for Wireless Radiation Field Reconstruction: A 3D Gaussian Splatting Approach
Wireless channel modeling plays a pivotal role in designing, analyzing, and optimizing wireless communication systems. Nevertheless, developing an effective channel modeling approach has been a long-standing challenge. This issue has been escalated due to denser network deployment, larger antenna arrays, and broader bandwidth in next-generation networks. To address this challenge, we put forth WRF-GS, a novel framework for channel modeling based on wireless radiation field (WRF) reconstruction using 3D Gaussian splatting (3D-GS). WRF-GS employs 3D Gaussian primitives and neural networks to capture the interactions between the environment and radio signals, enabling efficient WRF reconstruction and visualization of the propagation characteristics. The reconstructed WRF can then be used to synthesize the spatial spectrum for comprehensive wireless channel characterization. While WRF-GS demonstrates remarkable effectiveness, it faces limitations in capturing high-frequency signal variations caused by complex multipath effects. To overcome these limitations, we propose WRF-GS+, an enhanced framework that integrates electromagnetic wave physics into the neural network design. WRF-GS+ leverages deformable 3D Gaussians to model both static and dynamic components of the WRF, significantly improving its ability to characterize signal variations. In addition, WRF-GS+ enhances the splatting process by simplifying the 3D-GS modeling process and improving computational efficiency. Experimental results demonstrate that both WRF-GS and WRF-GS+ outperform baselines for spatial spectrum synthesis, including ray tracing and other deep-learning approaches. Notably, WRF-GS+ achieves state-of-the-art performance in the received signal strength indication (RSSI) and channel state information (CSI) prediction tasks, surpassing existing methods by more than 0.7 dB and 3.36 dB, respectively.
comment: This is an extended journal version of our previous conference paper that was accepted to the IEEE INFOCOM 2025 at arXiv:2412.04832v2. The code for this version is available at https://github.com/wenchaozheng/WRF-GSplus
♻ ☆ R2-KG: General-Purpose Dual-Agent Framework for Reliable Reasoning on Knowledge Graphs
Recent studies have combined Large Language Models (LLMs) with Knowledge Graphs (KGs) to enhance reasoning, improving inference accuracy without additional training while mitigating hallucination. However, existing frameworks are often rigid, struggling to adapt to KG or task changes. They also rely heavily on powerful LLMs for reliable (i.e., trustworthy) reasoning. To address this, We introduce R2-KG, a plug-and-play, dual-agent framework that separates reasoning into two roles: an Operator (a low-capacity LLM) that gathers evidence and a Supervisor (a high-capacity LLM) that makes final judgments. This design is cost-efficient for LLM inference while still maintaining strong reasoning accuracy. Additionally, R2-KG employs an Abstention mechanism, generating answers only when sufficient evidence is collected from KG, which significantly enhances reliability. Experiments across multiple KG-based reasoning tasks show that R2-KG consistently outperforms baselines in both accuracy and reliability, regardless of the inherent capability of LLMs used as the Operator. Further experiments reveal that the single-agent version of R2-KG, equipped with a strict self-consistency strategy, achieves significantly higher-than-baseline reliability while reducing inference cost. However, it also leads to a higher abstention rate in complex KGs. Our findings establish R2-KG as a flexible and cost-effective solution for KG-based reasoning. It reduces reliance on high-capacity LLMs while ensuring trustworthy inference.
♻ ☆ Advancing Mathematical Reasoning in Language Models: The Impact of Problem-Solving Data, Data Synthesis Methods, and Training Stages ICLR 2025
Mathematical reasoning remains a challenging area for large language models (LLMs), prompting the development of math-specific LLMs such as LLEMMA, DeepSeekMath, and Qwen2-Math, among others. These models typically follow a two-stage training paradigm: pre-training with math-related corpora and post-training with problem datasets for supervised fine-tuning (SFT). Despite these efforts, the improvements in mathematical reasoning achieved through continued pre-training (CPT) are often less significant compared to those obtained via SFT. This study addresses this discrepancy by exploring alternative strategies during the pre-training phase, focusing on the use of problem-solving data over general mathematical corpora. We investigate three primary research questions: (1) Can problem-solving data enhance the model's mathematical reasoning capabilities more effectively than general mathematical corpora during CPT? (2) Are synthetic data from the same source equally effective, and which synthesis methods are most efficient? (3) How do the capabilities developed from the same problem-solving data differ between the CPT and SFT stages, and what factors contribute to these differences? Our findings indicate that problem-solving data significantly enhances the model's mathematical capabilities compared to general mathematical corpora. We also identify effective data synthesis methods, demonstrating that the tutorship amplification synthesis method achieves the best performance. Furthermore, while SFT facilitates instruction-following abilities, it underperforms compared to CPT with the same data, which can be partially attributed to its poor learning capacity for more challenging problem-solving data. These insights provide valuable guidance for optimizing the mathematical reasoning capabilities of LLMs, culminating in our development of a powerful mathematical base model called MathGPT-8B.
comment: ICLR 2025
♻ ☆ Optimizing Generative AI's Accuracy and Transparency in Inductive Thematic Analysis: A Human-AI Comparison
This study highlights the transparency and accuracy of GenAI's inductive thematic analysis, particularly using GPT-4 Turbo API integrated within a stepwise prompt-based Python script. This approach ensured a traceable and systematic coding process, generating codes with supporting statements and page references, which enhanced validation and reproducibility. The results indicate that GenAI performs inductive coding in a manner closely resembling human coders, effectively categorizing themes at a level like the average human coder. However, in interpretation, GenAI extends beyond human coders by situating themes within a broader conceptual context, providing a more generalized and abstract perspective.
♻ ☆ Education distillation:getting student models to learn in shcools
This paper introduces a new knowledge distillation method, called education distillation (ED), which is inspired by the structured and progressive nature of human learning. ED mimics the educational stages of primary school, middle school, and university and designs teaching reference blocks. The student model is split into a main body and multiple teaching reference blocks to learn from teachers step by step. This promotes efficient knowledge distillation while maintaining the architecture of the student model. Experimental results on the CIFAR100, Tiny Imagenet, Caltech and Food-101 datasets show that the teaching reference blocks can effectively avoid the problem of forgetting. Compared with conventional single-teacher and multi-teacher knowledge distillation methods, ED significantly improves the accuracy and generalization ability of the student model. These findings highlight the potential of ED to improve model performance across different architectures and datasets, indicating its value in various deep learning scenarios. Code examples can be obtained at: https://github.com/Revolutioner1/ED.git.
♻ ☆ Differentially Private Kernel Density Estimation
We introduce a refined differentially private (DP) data structure for kernel density estimation (KDE), offering not only improved privacy-utility tradeoff but also better efficiency over prior results. Specifically, we study the mathematical problem: given a similarity function $f$ (or DP KDE) and a private dataset $X \subset \mathbb{R}^d$, our goal is to preprocess $X$ so that for any query $y\in\mathbb{R}^d$, we approximate $\sum_{x \in X} f(x, y)$ in a differentially private fashion. The best previous algorithm for $f(x,y) =\| x - y \|_1$ is the node-contaminated balanced binary tree by [Backurs, Lin, Mahabadi, Silwal, and Tarnawski, ICLR 2024]. Their algorithm requires $O(nd)$ space and time for preprocessing with $n=|X|$. For any query point, the query time is $d \log n$, with an error guarantee of $(1+\alpha)$-approximation and $\epsilon^{-1} \alpha^{-0.5} d^{1.5} R \log^{1.5} n$. In this paper, we improve the best previous result [Backurs, Lin, Mahabadi, Silwal, and Tarnawski, ICLR 2024] in three aspects: - We reduce query time by a factor of $\alpha^{-1} \log n$. - We improve the approximation ratio from $\alpha$ to 1. - We reduce the error dependence by a factor of $\alpha^{-0.5}$. From a technical perspective, our method of constructing the search tree differs from previous work [Backurs, Lin, Mahabadi, Silwal, and Tarnawski, ICLR 2024]. In prior work, for each query, the answer is split into $\alpha^{-1} \log n$ numbers, each derived from the summation of $\log n$ values in interval tree countings. In contrast, we construct the tree differently, splitting the answer into $\log n$ numbers, where each is a smart combination of two distance values, two counting values, and $y$ itself. We believe our tree structure may be of independent interest.
comment: v2: Appendix added. v3: Numerical validations added
♻ ☆ LTL-Constrained Policy Optimization with Cycle Experience Replay
Linear Temporal Logic (LTL) offers a precise means for constraining the behavior of reinforcement learning agents. However, in many settings where both satisfaction and optimality conditions are present, LTL is insufficient to capture both. Instead, LTL-constrained policy optimization, where the goal is to optimize a scalar reward under LTL constraints, is needed. This constrained optimization problem proves difficult in deep Reinforcement Learning (DRL) settings, where learned policies often ignore the LTL constraint due to the sparse nature of LTL satisfaction. To alleviate the sparsity issue, we introduce Cycle Experience Replay (CyclER), a novel reward shaping technique that exploits the underlying structure of the LTL constraint to guide a policy towards satisfaction by encouraging partial behaviors compliant with the constraint. We provide a theoretical guarantee that optimizing CyclER will achieve policies that satisfy the LTL constraint with near-optimal probability. We evaluate CyclER in three continuous control domains. Our experimental results show that optimizing CyclER in tandem with the existing scalar reward outperforms existing reward-shaping methods at finding performant LTL-satisfying policies.
comment: Published in TMLR, 12 pages in main text
♻ ☆ SILMM: Self-Improving Large Multimodal Models for Compositional Text-to-Image Generation CVPR 2025
Large Multimodal Models (LMMs) have demonstrated impressive capabilities in multimodal understanding and generation, pushing forward advancements in text-to-image generation. However, achieving accurate text-image alignment for LMMs, particularly in compositional scenarios, remains challenging. Existing approaches, such as layout planning for multi-step generation and learning from human feedback or AI feedback, depend heavily on prompt engineering, costly human annotations, and continual upgrading, limiting flexibility and scalability. In this work, we introduce a model-agnostic iterative self-improvement framework (SILMM) that can enable LMMs to provide helpful and scalable self-feedback and optimize text-image alignment via Direct Preference Optimization (DPO). DPO can readily applied to LMMs that use discrete visual tokens as intermediate image representations; while it is less suitable for LMMs with continuous visual features, as obtaining generation probabilities is challenging. To adapt SILMM to LMMs with continuous features, we propose a diversity mechanism to obtain diverse representations and a kernel-based continuous DPO for alignment. Extensive experiments on three compositional text-to-image generation benchmarks validate the effectiveness and superiority of SILMM, showing improvements exceeding 30% on T2I-CompBench++ and around 20% on DPG-Bench.
comment: CVPR 2025 Camera-ready. Project page: https://silmm.github.io/
♻ ☆ TIGeR: Unifying Text-to-Image Generation and Retrieval with Large Multimodal Models ICLR 2025
How humans can effectively and efficiently acquire images has always been a perennial question. A classic solution is text-to-image retrieval from an existing database; however, the limited database typically lacks creativity. By contrast, recent breakthroughs in text-to-image generation have made it possible to produce attractive and counterfactual visual content, but it faces challenges in synthesizing knowledge-intensive images. In this work, we rethink the relationship between text-to-image generation and retrieval, proposing a unified framework for both tasks with one single Large Multimodal Model (LMM). Specifically, we first explore the intrinsic discriminative abilities of LMMs and introduce an efficient generative retrieval method for text-to-image retrieval in a training-free manner. Subsequently, we unify generation and retrieval autoregressively and propose an autonomous decision mechanism to choose the best-matched one between generated and retrieved images as the response to the text prompt. To standardize the evaluation of unified text-to-image generation and retrieval, we construct TIGeR-Bench, a benchmark spanning both creative and knowledge-intensive domains. Extensive experiments on TIGeR-Bench and two retrieval benchmarks, i.e., Flickr30K and MS-COCO, demonstrate the superiority of our proposed framework.
comment: ICLR 2025 Camera-ready
♻ ☆ BEVDiffuser: Plug-and-Play Diffusion Model for BEV Denoising with Ground-Truth Guidance CVPR 2025
Bird's-eye-view (BEV) representations play a crucial role in autonomous driving tasks. Despite recent advancements in BEV generation, inherent noise, stemming from sensor limitations and the learning process, remains largely unaddressed, resulting in suboptimal BEV representations that adversely impact the performance of downstream tasks. To address this, we propose BEVDiffuser, a novel diffusion model that effectively denoises BEV feature maps using the ground-truth object layout as guidance. BEVDiffuser can be operated in a plug-and-play manner during training time to enhance existing BEV models without requiring any architectural modifications. Extensive experiments on the challenging nuScenes dataset demonstrate BEVDiffuser's exceptional denoising and generation capabilities, which enable significant enhancement to existing BEV models, as evidenced by notable improvements of 12.3\% in mAP and 10.1\% in NDS achieved for 3D object detection without introducing additional computational complexity. Moreover, substantial improvements in long-tail object detection and under challenging weather and lighting conditions further validate BEVDiffuser's effectiveness in denoising and enhancing BEV representations.
comment: CVPR 2025
♻ ☆ The Inadequacy of Similarity-based Privacy Metrics: Privacy Attacks against "Truly Anonymous" Synthetic Datasets
Generative models producing synthetic data are meant to provide a privacy-friendly approach to releasing data. However, their privacy guarantees are only considered robust when models satisfy Differential Privacy (DP). Alas, this is not a ubiquitous standard, as many leading companies (and, in fact, research papers) use ad-hoc privacy metrics based on testing the statistical similarity between synthetic and real data. In this paper, we examine the privacy metrics used in real-world synthetic data deployments and demonstrate their unreliability in several ways. First, we provide counter-examples where severe privacy violations occur even if the privacy tests pass and instantiate accurate membership and attribute inference attacks with minimal cost. We then introduce ReconSyn, a reconstruction attack that generates multiple synthetic datasets that are considered private by the metrics but actually leak information unique to individual records. We show that ReconSyn recovers 78-100% of the outliers in the train data with only black-box access to a single fitted generative model and the privacy metrics. In the process, we show that applying DP only to the model does not mitigate this attack, as using privacy metrics breaks the end-to-end DP pipeline.
comment: Published in the Proceedings of the 46th IEEE Symposium on Security & Privacy (IEEE S&P 2025). Please cite the S&P version
♻ ☆ NextStop: An Improved Tracker For Panoptic LIDAR Segmentation Data
4D panoptic LiDAR segmentation is essential for scene understanding in autonomous driving and robotics, combining semantic and instance segmentation with temporal consistency. Current methods, like 4D-PLS and 4D-STOP, use a tracking-by-detection methodology, employing deep learning networks to perform semantic and instance segmentation on each frame. To maintain temporal consistency, large-size instances detected in the current frame are compared and associated with instances within a temporal window that includes the current and preceding frames. However, their reliance on short-term instance detection, lack of motion estimation, and exclusion of small-sized instances lead to frequent identity switches and reduced tracking performance. We address these issues with the NextStop1 tracker, which integrates Kalman filter-based motion estimation, data association, and lifespan management, along with a tracklet state concept to improve prioritization. Evaluated using the LiDAR Segmentation and Tracking Quality (LSTQ) metric on the SemanticKITTI validation set, NextStop demonstrated enhanced tracking performance, particularly for small-sized objects like people and bicyclists, with fewer ID switches, earlier tracking initiation, and improved reliability in complex environments. The source code is available at https://github.com/AIROTAU/NextStop
♻ ☆ Deepfake-Eval-2024: A Multi-Modal In-the-Wild Benchmark of Deepfakes Circulated in 2024
In the age of increasingly realistic generative AI, robust deepfake detection is essential for mitigating fraud and disinformation. While many deepfake detectors report high accuracy on academic datasets, we show that these academic benchmarks are out of date and not representative of real-world deepfakes. We introduce Deepfake-Eval-2024, a new deepfake detection benchmark consisting of in-the-wild deepfakes collected from social media and deepfake detection platform users in 2024. Deepfake-Eval-2024 consists of 45 hours of videos, 56.5 hours of audio, and 1,975 images, encompassing the latest manipulation technologies. The benchmark contains diverse media content from 88 different websites in 52 different languages. We find that the performance of open-source state-of-the-art deepfake detection models drops precipitously when evaluated on Deepfake-Eval-2024, with AUC decreasing by 50% for video, 48% for audio, and 45% for image models compared to previous benchmarks. We also evaluate commercial deepfake detection models and models finetuned on Deepfake-Eval-2024, and find that they have superior performance to off-the-shelf open-source models, but do not yet reach the accuracy of deepfake forensic analysts. The dataset is available at https://github.com/nuriachandra/Deepfake-Eval-2024.
♻ ☆ LLM-Match: An Open-Sourced Patient Matching Model Based on Large Language Models and Retrieval-Augmented Generation
Patient matching is the process of linking patients to appropriate clinical trials by accurately identifying and matching their medical records with trial eligibility criteria. We propose LLM-Match, a novel framework for patient matching leveraging fine-tuned open-source large language models. Our approach consists of four key components. First, a retrieval-augmented generation (RAG) module extracts relevant patient context from a vast pool of electronic health records (EHRs). Second, a prompt generation module constructs input prompts by integrating trial eligibility criteria (both inclusion and exclusion criteria), patient context, and system instructions. Third, a fine-tuning module with a classification head optimizes the model parameters using structured prompts and ground-truth labels. Fourth, an evaluation module assesses the fine-tuned model's performance on the testing datasets. We evaluated LLM-Match on four open datasets - n2c2, SIGIR, TREC 2021, and TREC 2022 - using open-source models, comparing it against TrialGPT, Zero-Shot, and GPT-4-based closed models. LLM-Match outperformed all baselines.
comment: 10 pages, 1 figure
♻ ☆ HREF: Human Response-Guided Evaluation of Instruction Following in Language Models
Evaluating the capability of Large Language Models (LLMs) in following instructions has heavily relied on a powerful LLM as the judge, introducing unresolved biases that deviate the judgments from human judges. In this work, we reevaluate various choices for automatic evaluation on a wide range of instruction-following tasks. We experiment with methods that leverage human-written responses and observe that they enhance the reliability of automatic evaluations across a wide range of tasks, resulting in up to a 3.2% improvement in agreement with human judges. We also discovered that human-written responses offer an orthogonal perspective to model-generated responses in following instructions and should be used as an additional context when comparing model responses. Based on these observations, we develop a new evaluation benchmark, Human Response-Guided Evaluation of Instruction Following (HREF), comprising 4,258 samples across 11 task categories with a composite evaluation setup, employing a composite evaluation setup that selects the most reliable method for each category. In addition to providing reliable evaluation, HREF emphasizes individual task performance and is free from contamination. Finally, we study the impact of key design choices in HREF, including the size of the evaluation set, the judge model, the baseline model, and the prompt template. We host a live leaderboard that evaluates LLMs on the private evaluation set of HREF.
comment: 28 pages, 15 figures
♻ ☆ Int2Int: a framework for mathematics with transformers
This paper documents Int2Int, an open source code base for using transformers on problems of mathematical research, with a focus on number theory and other problems involving integers. Int2Int is a complete PyTorch implementation of a transformer architecture, together with training and evaluation loops, and classes and functions to represent, generate and decode common mathematical objects. Ancillary code for data preparation, and Jupyter Notebooks for visualizing experimental results are also provided. This document presents the main features of Int2Int, serves as its user manual, and provides guidelines on how to extend it. Int2Int is released under the MIT licence, at https://github.com/f-charton/Int2Int.
♻ ☆ QualityFlow: An Agentic Workflow for Program Synthesis Controlled by LLM Quality Checks
We introduce QualityFlow, a dynamic agentic workflow for program synthesis. Given the English description of a programming problem and a set of unit tests, the model's goal is to synthesize the correct program that solves the problem and passes the tests. QualityFlow includes large language model (LLM) agents resembling a software development team, including code generation, testing, and self-debugging. We propose the LLM Quality Checker, which explicitly "imagines" whether the synthesized programs' execution would conform to the unit tests. The Quality Checks dynamically control the workflow, including actions to submit the final answer, clarify the problem statement, and revert previous workflow steps. Our experiments show that the Quality Checker can precisely accept any correct program, mitigate faulty synthesized tests, and prevent potential workflow deviation. QualityFlow establishes the state-of-the-art results on four program synthesis benchmarks: MBPP, HumanEval, and stricter evaluations from MBPP-EvalPlus and HumanEval-EvalPlus.
♻ ☆ Counterfactual Token Generation in Large Language Models
"Sure, I am happy to generate a story for you: Captain Lyra stood at the helm of her trusty ship, the Maelstrom's Fury, gazing out at the endless sea. [...] Lyra's eyes welled up with tears as she realized the bitter truth - she had sacrificed everything for fleeting riches, and lost the love of her crew, her family, and herself." Although this story, generated by a large language model, is captivating, one may wonder -- how would the story have unfolded if the model had chosen "Captain Maeve" as the protagonist instead? We cannot know. State-of-the-art large language models are stateless -- they maintain no internal memory or state. Given a prompt, they generate a sequence of tokens as an output using an autoregressive process. As a consequence, they cannot reason about counterfactual alternatives to tokens they have generated in the past. In this work, our goal is to enhance them with this functionality. To this end, we develop a causal model of token generation that builds upon the Gumbel-Max structural causal model. Our model allows any large language model to perform counterfactual token generation at almost no cost in comparison with vanilla token generation, it is embarrassingly simple to implement, and it does not require any fine-tuning nor prompt engineering. We implement our model on Llama 3 8B-Instruct and Ministral-8B-Instruct and conduct a qualitative and a quantitative analysis of counterfactually generated text. We conclude with a demonstrative application of counterfactual token generation for bias detection, unveiling interesting insights about the model of the world constructed by large language models.
comment: Accepted at CLeaR 2025
♻ ☆ Morphological Symmetries in Robotics
We present a comprehensive framework for studying and leveraging morphological symmetries in robotic systems. These are intrinsic properties of the robot's morphology, frequently observed in animal biology and robotics, which stem from the replication of kinematic structures and the symmetrical distribution of mass. We illustrate how these symmetries extend to the robot's state space and both proprioceptive and exteroceptive sensor measurements, resulting in the equivariance of the robot's equations of motion and optimal control policies. Thus, we recognize morphological symmetries as a relevant and previously unexplored physics-informed geometric prior, with significant implications for both data-driven and analytical methods used in modeling, control, estimation and design in robotics. For data-driven methods, we demonstrate that morphological symmetries can enhance the sample efficiency and generalization of machine learning models through data augmentation, or by applying equivariant/invariant constraints on the model's architecture. In the context of analytical methods, we employ abstract harmonic analysis to decompose the robot's dynamics into a superposition of lower-dimensional, independent dynamics. We substantiate our claims with both synthetic and real-world experiments conducted on bipedal and quadrupedal robots. Lastly, we introduce the repository MorphoSymm to facilitate the practical use of the theory and applications outlined in this work.
comment: 18 pages, 11 figures
♻ ☆ Chameleon: a Heterogeneous and Disaggregated Accelerator System for Retrieval-Augmented Language Models VLDB 2025
A Retrieval-Augmented Language Model (RALM) combines a large language model (LLM) with a vector database to retrieve context-specific knowledge during text generation. This strategy facilitates impressive generation quality even with smaller models, thus reducing computational demands by orders of magnitude. To serve RALMs efficiently and flexibly, we propose Chameleon, a heterogeneous accelerator system integrating both LLM and vector search accelerators in a disaggregated architecture. The heterogeneity ensures efficient serving for both inference and retrieval, while the disaggregation allows independent scaling of LLM and vector search accelerators to fulfill diverse RALM requirements. Our Chameleon prototype implements vector search accelerators on FPGAs and assigns LLM inference to GPUs, with CPUs as cluster coordinators. Evaluated on various RALMs, Chameleon exhibits up to 2.16$\times$ reduction in latency and 3.18x speedup in throughput compared to the hybrid CPU-GPU architecture. The promising results pave the way for adopting heterogeneous accelerators for not only LLM inference but also vector search in future RALM systems.
comment: Accepted by VLDB 2025
Computation and Language 57
☆ ShED-HD: A Shannon Entropy Distribution Framework for Lightweight Hallucination Detection on Edge Devices
Large Language Models (LLMs) have demonstrated impressive capabilities on a broad array of NLP tasks, but their tendency to produce hallucinations$\unicode{x2013}$plausible-sounding but factually incorrect content$\unicode{x2013}$poses severe challenges in high-stakes domains. Existing hallucination detection methods either bear the computational cost of multiple inference passes or sacrifice accuracy for efficiency with single-pass approaches, neither of which is ideal in resource-constrained environments such as edge devices. We propose the Shannon Entropy Distribution Hallucination Detector (ShED-HD), a novel hallucination detection framework that bridges this gap by classifying sequence-level entropy patterns using a lightweight BiLSTM architecture with single-headed attention. In contrast to prior approaches, ShED-HD efficiently detects distinctive uncertainty patterns across entire output sequences, preserving contextual awareness. Through in-depth evaluation on three datasets (BioASQ, TriviaQA, and Jeopardy Questions), we show that ShED-HD significantly outperforms other computationally efficient approaches in the out-of-distribution setting, while achieving comparable performance in the in-distribution setting. ShED-HD facilitates hallucination detection that is low-cost, accurate, and generalizable, improving the credibility of content generated by LLMs in resource-constrained environments where trustworthy AI functionality is crucial.
☆ Mapping Hymns and Organizing Concepts in the Rigveda: Quantitatively Connecting the Vedic Suktas NAACL 2025
Accessing and gaining insight into the Rigveda poses a non-trivial challenge due to its extremely ancient Sanskrit language, poetic structure, and large volume of text. By using NLP techniques, this study identified topics and semantic connections of hymns within the Rigveda that were corroborated by seven well-known groupings of hymns. The 1,028 suktas (hymns) from the modern English translation of the Rigveda by Jamison and Brereton were preprocessed and sukta-level embeddings were obtained using, i) a novel adaptation of LSA, presented herein, ii) SBERT, and iii) Doc2Vec embeddings. Following an UMAP dimension reduction of the vectors, the network of suktas was formed using k-nearest neighbours. Then, community detection of topics in the sukta networks was performed with the Louvain, Leiden, and label propagation methods, whose statistical significance of the formed topics were determined using an appropriate null distribution. Only the novel adaptation of LSA using the Leiden method, had detected sukta topic networks that were significant (z = 2.726, p < .01) with a modularity score of 0.944. Of the seven famous sukta groupings analyzed (e.g., creation, funeral, water, etc.) the LSA derived network was successful in all seven cases, while Doc2Vec was not significant and failed to detect the relevant suktas. SBERT detected four of the famous suktas as separate groups, but mistakenly combined three of them into a single mixed group. Also, the SBERT network was not statistically significant.
comment: Accepted to NLP4DH 2025 at NAACL 2025
☆ Decoupling Angles and Strength in Low-rank Adaptation ICLR 2025
Parameter-Efficient FineTuning (PEFT) methods have recently gained significant popularity thanks to the widespread availability of large-scale pretrained models. These methods allow for quick adaptation to downstream tasks with minimal computational cost. However, popular finetuning methods such as LoRA exhibit limited robustness when it comes to hyperparameter choices or extended training regimes, preventing optimal out-of-the-box performance. In contrast, bounded approaches, such as ETHER, provide greater robustness but are limited to extremely low-rank adaptations and fixed-strength transformations, reducing their adaptation expressive power. In this work, we propose Decoupled Low-rank Adaptation (DeLoRA), a novel finetuning method that normalizes and scales learnable low-rank matrices. By bounding the distance of the transformation, DeLoRA effectively decouples the angular learning from the adaptation strength, enhancing robustness without compromising performance. Through evaluations on subject-driven image generation, natural language understanding, and instruction tuning, we show that DeLoRA matches or surpasses performance of competing PEFT methods, while exhibiting stronger robustness. Code is available at https://github.com/ExplainableML/DeLoRA.
comment: ICLR 2025
☆ LakotaBERT: A Transformer-based Model for Low Resource Lakota Language
Lakota, a critically endangered language of the Sioux people in North America, faces significant challenges due to declining fluency among younger generations. This paper introduces LakotaBERT, the first large language model (LLM) tailored for Lakota, aiming to support language revitalization efforts. Our research has two primary objectives: (1) to create a comprehensive Lakota language corpus and (2) to develop a customized LLM for Lakota. We compiled a diverse corpus of 105K sentences in Lakota, English, and parallel texts from various sources, such as books and websites, emphasizing the cultural significance and historical context of the Lakota language. Utilizing the RoBERTa architecture, we pre-trained our model and conducted comparative evaluations against established models such as RoBERTa, BERT, and multilingual BERT. Initial results demonstrate a masked language modeling accuracy of 51% with a single ground truth assumption, showcasing performance comparable to that of English-based models. We also evaluated the model using additional metrics, such as precision and F1 score, to provide a comprehensive assessment of its capabilities. By integrating AI and linguistic methodologies, we aspire to enhance linguistic diversity and cultural resilience, setting a valuable precedent for leveraging technology in the revitalization of other endangered indigenous languages.
☆ Exploring Topic Trends in COVID-19 Research Literature using Non-Negative Matrix Factorization
In this work, we apply topic modeling using Non-Negative Matrix Factorization (NMF) on the COVID-19 Open Research Dataset (CORD-19) to uncover the underlying thematic structure and its evolution within the extensive body of COVID-19 research literature. NMF factorizes the document-term matrix into two non-negative matrices, effectively representing the topics and their distribution across the documents. This helps us see how strongly documents relate to topics and how topics relate to words. We describe the complete methodology which involves a series of rigorous pre-processing steps to standardize the available text data while preserving the context of phrases, and subsequently feature extraction using the term frequency-inverse document frequency (tf-idf), which assigns weights to words based on their frequency and rarity in the dataset. To ensure the robustness of our topic model, we conduct a stability analysis. This process assesses the stability scores of the NMF topic model for different numbers of topics, enabling us to select the optimal number of topics for our analysis. Through our analysis, we track the evolution of topics over time within the CORD-19 dataset. Our findings contribute to the understanding of the knowledge structure of the COVID-19 research landscape, providing a valuable resource for future research in this field.
☆ GINGER: Grounded Information Nugget-Based Generation of Responses
Retrieval-augmented generation (RAG) faces challenges related to factual correctness, source attribution, and response completeness. To address them, we propose a modular pipeline for grounded response generation that operates on information nuggets-minimal, atomic units of relevant information extracted from retrieved documents. The multistage pipeline encompasses nugget detection, clustering, ranking, top cluster summarization, and fluency enhancement. It guarantees grounding in specific facts, facilitates source attribution, and ensures maximum information inclusion within length constraints. Extensive experiments on the TREC RAG'24 dataset evaluated with the AutoNuggetizer framework demonstrate that GINGER achieves state-of-the-art performance on this benchmark.
☆ Unmasking Deceptive Visuals: Benchmarking Multimodal Large Language Models on Misleading Chart Question Answering
Misleading chart visualizations, which intentionally manipulate data representations to support specific claims, can distort perceptions and lead to incorrect conclusions. Despite decades of research, misleading visualizations remain a widespread and pressing issue. Recent advances in multimodal large language models (MLLMs) have demonstrated strong chart comprehension capabilities, yet no existing work has systematically evaluated their ability to detect and interpret misleading charts. This paper introduces the Misleading Chart Question Answering (Misleading ChartQA) Benchmark, a large-scale multimodal dataset designed to assess MLLMs in identifying and reasoning about misleading charts. It contains over 3,000 curated examples, covering 21 types of misleaders and 10 chart types. Each example includes standardized chart code, CSV data, and multiple-choice questions with labeled explanations, validated through multi-round MLLM checks and exhausted expert human review. We benchmark 16 state-of-the-art MLLMs on our dataset, revealing their limitations in identifying visually deceptive practices. We also propose a novel pipeline that detects and localizes misleaders, enhancing MLLMs' accuracy in misleading chart interpretation. Our work establishes a foundation for advancing MLLM-driven misleading chart comprehension. We publicly release the sample dataset to support further research in this critical area.
comment: 31 pages in total. Under Review For ARR
☆ Evaluating Negative Sampling Approaches for Neural Topic Models
Negative sampling has emerged as an effective technique that enables deep learning models to learn better representations by introducing the paradigm of learn-to-compare. The goal of this approach is to add robustness to deep learning models to learn better representation by comparing the positive samples against the negative ones. Despite its numerous demonstrations in various areas of computer vision and natural language processing, a comprehensive study of the effect of negative sampling in an unsupervised domain like topic modeling has not been well explored. In this paper, we present a comprehensive analysis of the impact of different negative sampling strategies on neural topic models. We compare the performance of several popular neural topic models by incorporating a negative sampling technique in the decoder of variational autoencoder-based neural topic models. Experiments on four publicly available datasets demonstrate that integrating negative sampling into topic models results in significant enhancements across multiple aspects, including improved topic coherence, richer topic diversity, and more accurate document classification. Manual evaluations also indicate that the inclusion of negative sampling into neural topic models enhances the quality of the generated topics. These findings highlight the potential of negative sampling as a valuable tool for advancing the effectiveness of neural topic models.
comment: Code is available at: https://github.com/AdhyaSuman/Eval_NegTM
☆ MathAgent: Leveraging a Mixture-of-Math-Agent Framework for Real-World Multimodal Mathematical Error Detection
Mathematical error detection in educational settings presents a significant challenge for Multimodal Large Language Models (MLLMs), requiring a sophisticated understanding of both visual and textual mathematical content along with complex reasoning capabilities. Though effective in mathematical problem-solving, MLLMs often struggle with the nuanced task of identifying and categorizing student errors in multimodal mathematical contexts. Therefore, we introduce MathAgent, a novel Mixture-of-Math-Agent framework designed specifically to address these challenges. Our approach decomposes error detection into three phases, each handled by a specialized agent: an image-text consistency validator, a visual semantic interpreter, and an integrative error analyzer. This architecture enables more accurate processing of mathematical content by explicitly modeling relationships between multimodal problems and student solution steps. We evaluate MathAgent on real-world educational data, demonstrating approximately 5% higher accuracy in error step identification and 3% improvement in error categorization compared to baseline models. Besides, MathAgent has been successfully deployed in an educational platform that has served over one million K-12 students, achieving nearly 90% student satisfaction while generating significant cost savings by reducing manual error detection.
comment: Work In Progress
☆ GeoBenchX: Benchmarking LLMs for Multistep Geospatial Tasks
In this paper, we establish a benchmark for evaluating large language models (LLMs) on multi-step geospatial tasks relevant to commercial GIS practitioners. We assess seven leading commercial LLMs (Sonnet 3.5 and 3.7, Haiku 3.5, Gemini 2.0, GPT-4o, GPT-4o mini, and o3-mini) using a simple tool-calling agent equipped with 23 geospatial functions. Our benchmark comprises tasks across four categories of increasing complexity, with both solvable and intentionally unsolvable tasks to test hallucination rejection. We develop an LLM-as-Judge evaluation framework to compare agent solutions against reference implementations. Results show Sonnet 3.5 and GPT-4o achieve the best overall performance, with Claude models excelling on solvable tasks while OpenAI models better identify unsolvable scenarios. We observe significant differences in token usage, with Anthropic models consuming substantially more tokens than competitors. Common errors include misunderstanding geometrical relationships, relying on outdated knowledge, and inefficient data manipulation. The resulting benchmark set, evaluation framework, and data generation pipeline are released as open-source resources, providing one more standardized method for ongoing evaluation of LLMs for GeoAI.
comment: Github with code and benchmark set: https://github.com/Solirinai/GeoBenchX
☆ Detection of Somali-written Fake News and Toxic Messages on the Social Media Using Transformer-based Language Models
The fact that everyone with a social media account can create and share content, and the increasing public reliance on social media platforms as a news and information source bring about significant challenges such as misinformation, fake news, harmful content, etc. Although human content moderation may be useful to an extent and used by these platforms to flag posted materials, the use of AI models provides a more sustainable, scalable, and effective way to mitigate these harmful contents. However, low-resourced languages such as the Somali language face limitations in AI automation, including scarce annotated training datasets and lack of language models tailored to their unique linguistic characteristics. This paper presents part of our ongoing research work to bridge some of these gaps for the Somali language. In particular, we created two human-annotated social-media-sourced Somali datasets for two downstream applications, fake news \& toxicity classification, and developed a transformer-based monolingual Somali language model (named SomBERTa) -- the first of its kind to the best of our knowledge. SomBERTa is then fine-tuned and evaluated on toxic content, fake news and news topic classification datasets. Comparative evaluation analysis of the proposed model against related multilingual models (e.g., AfriBERTa, AfroXLMR, etc) demonstrated that SomBERTa consistently outperformed these comparators in both fake news and toxic content classification tasks while achieving the best average accuracy (87.99%) across all tasks. This research contributes to Somali NLP by offering a foundational language model and a replicable framework for other low-resource languages, promoting digital and AI inclusivity and linguistic diversity.
☆ AgentRxiv: Towards Collaborative Autonomous Research
Progress in scientific discovery is rarely the result of a single "Eureka" moment, but is rather the product of hundreds of scientists incrementally working together toward a common goal. While existing agent workflows are capable of producing research autonomously, they do so in isolation, without the ability to continuously improve upon prior research results. To address these challenges, we introduce AgentRxiv-a framework that lets LLM agent laboratories upload and retrieve reports from a shared preprint server in order to collaborate, share insights, and iteratively build on each other's research. We task agent laboratories to develop new reasoning and prompting techniques and find that agents with access to their prior research achieve higher performance improvements compared to agents operating in isolation (11.4% relative improvement over baseline on MATH-500). We find that the best performing strategy generalizes to benchmarks in other domains (improving on average by 3.3%). Multiple agent laboratories sharing research through AgentRxiv are able to work together towards a common goal, progressing more rapidly than isolated laboratories, achieving higher overall accuracy (13.7% relative improvement over baseline on MATH-500). These findings suggest that autonomous agents may play a role in designing future AI systems alongside humans. We hope that AgentRxiv allows agents to collaborate toward research goals and enables researchers to accelerate discovery.
☆ Clarifying Misconceptions in COVID-19 Vaccine Sentiment and Stance Analysis and Their Implications for Vaccine Hesitancy Mitigation: A Systematic Review
Background Advances in machine learning (ML) models have increased the capability of researchers to detect vaccine hesitancy in social media using Natural Language Processing (NLP). A considerable volume of research has identified the persistence of COVID-19 vaccine hesitancy in discourse shared on various social media platforms. Methods Our objective in this study was to conduct a systematic review of research employing sentiment analysis or stance detection to study discourse towards COVID-19 vaccines and vaccination spread on Twitter (officially known as X since 2023). Following registration in the PROSPERO international registry of systematic reviews, we searched papers published from 1 January 2020 to 31 December 2023 that used supervised machine learning to assess COVID-19 vaccine hesitancy through stance detection or sentiment analysis on Twitter. We categorized the studies according to a taxonomy of five dimensions: tweet sample selection approach, self-reported study type, classification typology, annotation codebook definitions, and interpretation of results. We analyzed if studies using stance detection report different hesitancy trends than those using sentiment analysis by examining how COVID-19 vaccine hesitancy is measured, and whether efforts were made to avoid measurement bias. Results Our review found that measurement bias is widely prevalent in studies employing supervised machine learning to analyze sentiment and stance toward COVID-19 vaccines and vaccination. The reporting errors are sufficiently serious that they hinder the generalisability and interpretation of these studies to understanding whether individual opinions communicate reluctance to vaccinate against SARS-CoV-2. Conclusion Improving the reporting of NLP methods is crucial to addressing knowledge gaps in vaccine hesitancy discourse.
comment: 14 pages, 3 figures, 4 tables
☆ $D^2LoRA$: Data-Driven LoRA Initialization for Low Resource Tasks
Tuning large language models is essential for optimizing their performance across diverse applications, particularly in scenarios with limited data availability. Tuning large language models in scarce data scenarios is crucial, particularly given that the convergence speed of the LoRA method is lower than that of full fine-tuning. In this paper, we present an analysis of post-training methods including Supervised Fine-Tuning (SFT), Direct Preference Optimization (DPO), and Odds Ratio Preference Optimization (ORPO) within the context of task-specific learning using the LoRA method. Next we introduce $D^2LoRA$, a data-driven approach for initializing LoRA metrics that enhances training efficiency, especially in limited-data settings. Our experiments compare $D^2LoRA$ with vanilla LoRA in terms of performance and catastrophic forgetting under extremely data-constrained conditions. The results demonstrate that $D^2LoRA$ achieves a 1% improvement GSM8K benchmark and a 2-point improvement in ROUGE score in title generation tasks. $D^2LoRA$ facilitates the adaptation of LLMs to multiple tasks even when task-specific data is scarce, thereby reducing training expenses and offering data cost.
☆ Temporal Relation Extraction in Clinical Texts: A Span-based Graph Transformer Approach
Temporal information extraction from unstructured text is essential for contextualizing events and deriving actionable insights, particularly in the medical domain. We address the task of extracting clinical events and their temporal relations using the well-studied I2B2 2012 Temporal Relations Challenge corpus. This task is inherently challenging due to complex clinical language, long documents, and sparse annotations. We introduce GRAPHTREX, a novel method integrating span-based entity-relation extraction, clinical large pre-trained language models (LPLMs), and Heterogeneous Graph Transformers (HGT) to capture local and global dependencies. Our HGT component facilitates information propagation across the document through innovative global landmarks that bridge distant entities. Our method improves the state-of-the-art with 5.5% improvement in the tempeval $F_1$ score over the previous best and up to 8.9% improvement on long-range relations, which presents a formidable challenge. This work not only advances temporal information extraction but also lays the groundwork for improved diagnostic and prognostic models through enhanced temporal reasoning.
comment: Introducing a novel method for joint extraction of medical events and temporal relations from free-text, leveraging clinical LPLMs and Heterogeneous Graph Transformers, achieving a 5.5% improvement over the previous state-of-the-art and up to 8.9% on long-range relations
☆ A Multi-Model Adaptation of Speculative Decoding for Classification
The current study introduces a novel adaptation of speculative decoding, repurposed from generation to classification tasks. We propose a multi-model framework employing up to three lightweight worker models and a single, more robust judge model analogous to draft models and target model, respectively, in speculative decoding. The worker models, tasked with the bulk of the computation, independently predict discrete class labels for a given input. When majority worker models agree on a label, it is accepted as the final label, optimizing efficiency by bypassing the computationally expensive judge model. In cases of disagreement, the judge model intervenes to resolve the label. This approach minimizes redundant computation, leverages the redundancy of multiple workers for confidence, and confines the judge model's role to challenging cases, offering a practical balance of efficiency and accuracy. Our analysis suggests that smaller out of the box instruction/chat finetuned worker models with 3 billion parameters (hereafter, 3B) demonstrate a level of alignment with judge models comparable to that of larger finetuned worker models with 7 billion parameters (hereafter, 7B) across both simple and higher order reasoning tasks. The top performing 3B worker model pair achieve an agreement rate of approximately 80-83% for sentiment and around 50-80% for similar ticket when compared to judge models. Additionally, 3B worker models provide a speedup ranging from 2.8x to 9x relative to the judge models, while 7B worker model combinations achieve a speedup ranging from 1.28x to 0.28x
☆ On the effectiveness of LLMs for automatic grading of open-ended questions in Spanish
Grading is a time-consuming and laborious task that educators must face. It is an important task since it provides feedback signals to learners, and it has been demonstrated that timely feedback improves the learning process. In recent years, the irruption of LLMs has shed light on the effectiveness of automatic grading. In this paper, we explore the performance of different LLMs and prompting techniques in automatically grading short-text answers to open-ended questions. Unlike most of the literature, our study focuses on a use case where the questions, answers, and prompts are all in Spanish. Experimental results comparing automatic scores to those of human-expert evaluators show good outcomes in terms of accuracy, precision and consistency for advanced LLMs, both open and proprietary. Results are notably sensitive to prompt styles, suggesting biases toward certain words or content in the prompt. However, the best combinations of models and prompt strategies, consistently surpasses an accuracy of 95% in a three-level grading task, which even rises up to more than 98% when the it is simplified to a binary right or wrong rating problem, which demonstrates the potential that LLMs have to implement this type of automation in education applications.
☆ Mind with Eyes: from Language Reasoning to Multimodal Reasoning
Language models have recently advanced into the realm of reasoning, yet it is through multimodal reasoning that we can fully unlock the potential to achieve more comprehensive, human-like cognitive capabilities. This survey provides a systematic overview of the recent multimodal reasoning approaches, categorizing them into two levels: language-centric multimodal reasoning and collaborative multimodal reasoning. The former encompasses one-pass visual perception and active visual perception, where vision primarily serves a supporting role in language reasoning. The latter involves action generation and state update within reasoning process, enabling a more dynamic interaction between modalities. Furthermore, we analyze the technical evolution of these methods, discuss their inherent challenges, and introduce key benchmark tasks and evaluation metrics for assessing multimodal reasoning performance. Finally, we provide insights into future research directions from the following two perspectives: (i) from visual-language reasoning to omnimodal reasoning and (ii) from multimodal reasoning to multimodal agents. This survey aims to provide a structured overview that will inspire further advancements in multimodal reasoning research.
☆ Long Is More Important Than Difficult for Training Reasoning Models
Difficult problems, which often result in long reasoning traces, are widely recognized as key factors for enhancing the performance of reasoning models. However, such high-challenge problems are scarce, limiting the size of available datasets. In this paper, we propose a simple method to decouple the reliance on problem difficulty. First, we empirically demonstrate that reasoning length, rather than problem difficulty, primarily influences the performance of trained models. Second, we identify a scaling law on reasoning length, showing that model performance increases in a log-linear fashion as the reasoning data length grows. Finally, we introduce a straightforward technique to generate reasoning data of arbitrary length, and show that synthesized data is effective for training reasoning models. After fine-tuning the Qwen2.5-32B-Instruct language model on our Long1K dataset, we present our model, Long1K-32B, which achieves remarkable performance with only 1,000 training samples, achieving 95.6\% accuracy on MATH, and 71.1\% on GPQA outperforming DeepSeek-R1-Distill-Qwen-32B. The model, code, and dataset are all open-sourced, available at https://huggingface.co/ZTss/LONG1.
comment: 15 pages,6 figures
☆ Unseen from Seen: Rewriting Observation-Instruction Using Foundation Models for Augmenting Vision-Language Navigation
Data scarcity is a long-standing challenge in the Vision-Language Navigation (VLN) field, which extremely hinders the generalization of agents to unseen environments. Previous works primarily rely on additional simulator data or web-collected images/videos to improve the generalization. However, the simulator environments still face limited diversity, and the web-collected data often requires extensive labor to remove the noise. In this paper, we propose a Rewriting-driven AugMentation (RAM) paradigm for VLN, which directly creates the unseen observation-instruction pairs via rewriting human-annotated training data. Benefiting from our rewriting mechanism, new observation-instruction can be obtained in both simulator-free and labor-saving manners to promote generalization. Specifically, we first introduce Object-Enriched Observation Rewriting, where we combine Vision-Language Models (VLMs) and Large Language Models (LLMs) to derive rewritten object-enriched scene descriptions, enabling observation synthesis with diverse objects and spatial layouts via Text-to-Image Generation Models (T2IMs). Then, we propose Observation-Contrast Instruction Rewriting, which generates observation-aligned rewritten instructions by requiring LLMs to reason the difference between original and new observations. We further develop a mixing-then-focusing training strategy with a random observation cropping scheme, effectively enhancing data distribution diversity while suppressing augmentation data noise during training. Experiments on both the discrete environments (R2R, REVERIE, and R4R datasets) and continuous environments (R2R-CE dataset) show the superior performance and impressive generalization ability of our method. Code is available at https://github.com/SaDil13/VLN-RAM.
☆ Dynamic Task Vector Grouping for Efficient Multi-Task Prompt Tuning
Multi-task prompt tuning utilizes multiple high-resource source tasks to improve performance on low-source target tasks. Existing approaches transfer the soft prompt trained by combining all source tasks or a single ``high-similar'' source task one-time-only. However, we find that the optimal transfer performance often comes from a combination of source tasks, which is neither one nor all. Further, we find that the similarity between source and target tasks also changes dynamically during fine-tuning after transfering, making similarity calculation in the initiation stage inadequate. To address these issues, we propose a method called Dynamic Task Vector Grouping (DTVG), whose core ideas contain (1) measuring the task similarity with task vectors instead of soft prompt, (2) grouping the optimal source task combination based on two metrics: {\it target similarity} and {\it knowledge consistency}; (3) dynamically updating the combination in each iteration step. Extensive experiments on the 26 NLP datasets under different settings demonstrate that DTVG effectively groups similar source tasks while reducing negative transfer, achieving the start-of-art performance.
comment: Work in progress
☆ Investigating Recent Large Language Models for Vietnamese Machine Reading Comprehension
Large Language Models (LLMs) have shown remarkable proficiency in Machine Reading Comprehension (MRC) tasks; however, their effectiveness for low-resource languages like Vietnamese remains largely unexplored. In this paper, we fine-tune and evaluate two state-of-the-art LLMs: Llama 3 (8B parameters) and Gemma (7B parameters), on ViMMRC, a Vietnamese MRC dataset. By utilizing Quantized Low-Rank Adaptation (QLoRA), we efficiently fine-tune these models and compare their performance against powerful LLM-based baselines. Although our fine-tuned models are smaller than GPT-3 and GPT-3.5, they outperform both traditional BERT-based approaches and these larger models. This demonstrates the effectiveness of our fine-tuning process, showcasing how modern LLMs can surpass the capabilities of older models like BERT while still being suitable for deployment in resource-constrained environments. Through intensive analyses, we explore various aspects of model performance, providing valuable insights into adapting LLMs for low-resource languages like Vietnamese. Our study contributes to the advancement of natural language processing in low-resource languages, and we make our fine-tuned models publicly available at: https://huggingface.co/iaiuet.
☆ (G)I-DLE: Generative Inference via Distribution-preserving Logit Exclusion with KL Divergence Minimization for Constrained Decoding
We propose (G)I-DLE, a new approach to constrained decoding that leverages KL divergence minimization to preserve the intrinsic conditional probability distribution of autoregressive language models while excluding undesirable tokens. Unlike conventional methods that naively set banned tokens' logits to $-\infty$, which can distort the conversion from raw logits to posterior probabilities and increase output variance, (G)I-DLE re-normalizes the allowed token probabilities to minimize such distortion. We validate our method on the K2-Eval dataset, specifically designed to assess Korean language fluency, logical reasoning, and cultural appropriateness. Experimental results on Qwen2.5 models (ranging from 1.5B to 14B) demonstrate that G-IDLE not only boosts mean evaluation scores but also substantially reduces the variance of output quality.
comment: preprint
☆ Expanding the Boundaries of Vision Prior Knowledge in Multi-modal Large Language Models
Does the prior knowledge of the vision encoder constrain the capability boundary of Multi-modal Large Language Models (MLLMs)? While most existing research treats MLLMs as unified systems optimized through end-to-end training, the impact of vision encoder's prior knowledge is seldom investigated. In this work, we introduce a novel metric, $Rank_e$, to quantify the effect of the vision encoder's prior knowledge on MLLM performance. Our analysis reveals a positive correlation between prior knowledge and MLLM performance. Moreover, we find that domain-specific fine-tuning using solely end-to-end visual question answering (VQA) data is insufficient--particularly for entities with low inherent visual prior knowledge. To address this issue, we propose VisPRE (Vision Prior Remediation), a two-stage training framework that explicitly incorporates prior knowledge at the vision encoder level. Experimental results demonstrate that augmenting vision encoder's prior knowledge substantially boosts the visual understanding capabilities of MLLMs, offering a novel and effective strategy for improving performance, especially in scenarios involving uncommon visual entities.
☆ Personalized Language Models via Privacy-Preserving Evolutionary Model Merging
Personalization in large language models (LLMs) seeks to tailor models to individual user or user group preferences. Prompt-based methods augment queries with user preference information, whereas training-based methods directly encode preferences into model parameters for more effective personalization. Despite achieving some success in personalizing LLMs, prior methods often fail to directly optimize task-specific metrics and lack explicit privacy-preservation mechanisms. To address these limitations, we propose Privacy-Preserving Model Merging via Evolutionary Algorithms (PriME), a novel approach to personalization that employs gradient-free methods to directly optimize task-specific metrics while preserving user privacy. By incorporating privacy preservation into optimization, PriME produces a personalized module that effectively captures the target user's preferences while minimizing the privacy risks for the users sharing their private information. Experiments on the LaMP benchmark show that PriME outperforms both prompt-based and training-based methods, achieving up to a 45% performance improvement over the prior art. Further analysis shows that PriME achieves a significantly better privacy-utility trade-off, highlighting the potential of evolutionary approaches for privacy-preserving LLM personalization.
comment: Preprint
☆ Instructing the Architecture Search for Spatial-temporal Sequence Forecasting with LLM
Spatial-temporal sequence forecasting (STSF) is a long-standing research problem with widespread real-world applications. Neural architecture search (NAS), which automates the neural network design, has been shown effective in tackling the STSF problem. However, the existing NAS methods for STSF focus on generating architectures in a time-consuming data-driven fashion, which heavily limits their ability to use background knowledge and explore the complicated search trajectory. Large language models (LLMs) have shown remarkable ability in decision-making with comprehensive internal world knowledge, but how it could benefit NAS for STSF remains unexplored. In this paper, we propose a novel NAS method for STSF based on LLM. Instead of directly generate architectures with LLM, We inspire the LLM's capability with a multi-level enhancement mechanism. Specifically, on the step-level, we decompose the generation task into decision steps with powerful prompt engineering and inspire LLM to serve as instructor for architecture search based on its internal knowledge. On the instance-level, we utilize a one-step tuning framework to quickly evaluate the architecture instance and a memory bank to cumulate knowledge to improve LLM's search ability. On the task-level, we propose a two-stage architecture search, balancing the exploration stage and optimization stage, to reduce the possibility of being trapped in local optima. Extensive experimental results demonstrate that our method can achieve competitive effectiveness with superior efficiency against existing NAS methods for STSF.
☆ Trade-offs in Large Reasoning Models: An Empirical Analysis of Deliberative and Adaptive Reasoning over Foundational Capabilities
Recent advancements in Large Reasoning Models (LRMs), such as OpenAI's o1/o3 and DeepSeek-R1, have demonstrated remarkable performance in specialized reasoning tasks through human-like deliberative thinking and long chain-of-thought reasoning. However, our systematic evaluation across various model families (DeepSeek, Qwen, and LLaMA) and scales (7B to 671B) reveals that acquiring these deliberative reasoning capabilities significantly reduces the foundational capabilities of LRMs, including notable declines in helpfulness and harmlessness, alongside substantially increased inference costs. Importantly, we demonstrate that adaptive reasoning -- employing modes like Zero-Thinking, Less-Thinking, and Summary-Thinking -- can effectively alleviate these drawbacks. Our empirical insights underline the critical need for developing more versatile LRMs capable of dynamically allocating inference-time compute according to specific task characteristics.
comment: 23 pages. Work in progress
☆ Understanding the Effects of RLHF on the Quality and Detectability of LLM-Generated Texts
Large Language Models (LLMs) have demonstrated exceptional performance on a range of downstream NLP tasks by generating text that closely resembles human writing. However, the ease of achieving this similarity raises concerns from potential malicious uses at scale by bad actors, as LLM-generated text becomes increasingly difficult to discern from human text. Although detection methods have been developed to address this issue, bad actors can further manipulate LLM-generated texts to make them less detectable. In this work, we study how further editing texts with Reinforcement Learning from Human Feedback (RLHF), which aligns model outputs with human preferences, affects (a) the quality of generated texts for two tasks, and (b) the performance of LLM-generated text detectors, looking at both training-based and zero-shot detection methods. Although RLHF improves the quality of LLM-generated texts, we find that it also tends to produce more detectable, lengthy, and repetitive outputs. Additionally, we observe that training-based detectors are vulnerable to short texts and to texts that incorporate code, whereas zero-shot detectors exhibit greater robustness.
comment: 14 pages, 3 figures
☆ Won: Establishing Best Practices for Korean Financial NLP KR
In this work, we present the first open leaderboard for evaluating Korean large language models focused on finance. Operated for about eight weeks, the leaderboard evaluated 1,119 submissions on a closed benchmark covering five MCQA categories: finance and accounting, stock price prediction, domestic company analysis, financial markets, and financial agent tasks and one open-ended qa task. Building on insights from these evaluations, we release an open instruction dataset of 80k instances and summarize widely used training strategies observed among top-performing models. Finally, we introduce Won, a fully open and transparent LLM built using these best practices. We hope our contributions help advance the development of better and safer financial LLMs for Korean and other languages.
comment: The training dataset is uploaded here: https://huggingface.co/datasets/KRX-Data/Won-Instruct. The model will be updated shortly
☆ Human-AI Interaction and User Satisfaction: Empirical Evidence from Online Reviews of AI Products
Human-AI Interaction (HAI) guidelines and design principles have become increasingly important in both industry and academia to guide the development of AI systems that align with user needs and expectations. However, large-scale empirical evidence on how HAI principles shape user satisfaction in practice remains limited. This study addresses that gap by analyzing over 100,000 user reviews of AI-related products from G2.com, a leading review platform for business software and services. Based on widely adopted industry guidelines, we identify seven core HAI dimensions and examine their coverage and sentiment within the reviews. We find that the sentiment on four HAI dimensions-adaptability, customization, error recovery, and security-is positively associated with overall user satisfaction. Moreover, we show that engagement with HAI dimensions varies by professional background: Users with technical job roles are more likely to discuss system-focused aspects, such as reliability, while non-technical users emphasize interaction-focused features like customization and feedback. Interestingly, the relationship between HAI sentiment and overall satisfaction is not moderated by job role, suggesting that once an HAI dimension has been identified by users, its effect on satisfaction is consistent across job roles.
☆ SLIDE: Sliding Localized Information for Document Extraction
Constructing accurate knowledge graphs from long texts and low-resource languages is challenging, as large language models (LLMs) experience degraded performance with longer input chunks. This problem is amplified in low-resource settings where data scarcity hinders accurate entity and relationship extraction. Contextual retrieval methods, while improving retrieval accuracy, struggle with long documents. They truncate critical information in texts exceeding maximum context lengths of LLMs, significantly limiting knowledge graph construction. We introduce SLIDE (Sliding Localized Information for Document Extraction), a chunking method that processes long documents by generating local context through overlapping windows. SLIDE ensures that essential contextual information is retained, enhancing knowledge graph extraction from documents exceeding LLM context limits. It significantly improves GraphRAG performance, achieving a 24% increase in entity extraction and a 39% improvement in relationship extraction for English. For Afrikaans, a low-resource language, SLIDE achieves a 49% increase in entity extraction and an 82% improvement in relationship extraction. Furthermore, it improves upon state-of-the-art in question-answering metrics such as comprehensiveness, diversity and empowerment, demonstrating its effectiveness in multilingual and resource-constrained settings.
☆ An Empirical Study of the Role of Incompleteness and Ambiguity in Interactions with Large Language Models
Natural language as a medium for human-computer interaction has long been anticipated, has been undergoing a sea-change with the advent of Large Language Models (LLMs) with startling capacities for processing and generating language. Many of us now treat LLMs as modern-day oracles, asking it almost any kind of question. Unlike its Delphic predecessor, consulting an LLM does not have to be a single-turn activity (ask a question, receive an answer, leave); and -- also unlike the Pythia -- it is widely acknowledged that answers from LLMs can be improved with additional context. In this paper, we aim to study when we need multi-turn interactions with LLMs to successfully get a question answered; or conclude that a question is unanswerable. We present a neural symbolic framework that models the interactions between human and LLM agents. Through the proposed framework, we define incompleteness and ambiguity in the questions as properties deducible from the messages exchanged in the interaction, and provide results from benchmark problems, in which the answer-correctness is shown to depend on whether or not questions demonstrate the presence of incompleteness or ambiguity (according to the properties we identify). Our results show multi-turn interactions are usually required for datasets which have a high proportion of incompleteness or ambiguous questions; and that that increasing interaction length has the effect of reducing incompleteness or ambiguity. The results also suggest that our measures of incompleteness and ambiguity can be useful tools for characterising interactions with an LLM on question-answeringproblems
☆ Experience Retrieval-Augmentation with Electronic Health Records Enables Accurate Discharge QA
To improve the reliability of Large Language Models (LLMs) in clinical applications, retrieval-augmented generation (RAG) is extensively applied to provide factual medical knowledge. However, beyond general medical knowledge from open-ended datasets, clinical case-based knowledge is also critical for effective medical reasoning, as it provides context grounded in real-world patient experiences. Motivated by this, we propose Experience Retrieval Augmentation - ExpRAG framework based on Electronic Health Record (EHR), aiming to offer the relevant context from other patients' discharge reports. ExpRAG performs retrieval through a coarse-to-fine process, utilizing an EHR-based report ranker to efficiently identify similar patients, followed by an experience retriever to extract task-relevant content for enhanced medical reasoning. To evaluate ExpRAG, we introduce DischargeQA, a clinical QA dataset with 1,280 discharge-related questions across diagnosis, medication, and instruction tasks. Each problem is generated using EHR data to ensure realistic and challenging scenarios. Experimental results demonstrate that ExpRAG consistently outperforms a text-based ranker, achieving an average relative improvement of 5.2%, highlighting the importance of case-based knowledge for medical reasoning.
☆ STShield: Single-Token Sentinel for Real-Time Jailbreak Detection in Large Language Models
Large Language Models (LLMs) have become increasingly vulnerable to jailbreak attacks that circumvent their safety mechanisms. While existing defense methods either suffer from adaptive attacks or require computationally expensive auxiliary models, we present STShield, a lightweight framework for real-time jailbroken judgement. STShield introduces a novel single-token sentinel mechanism that appends a binary safety indicator to the model's response sequence, leveraging the LLM's own alignment capabilities for detection. Our framework combines supervised fine-tuning on normal prompts with adversarial training using embedding-space perturbations, achieving robust detection while preserving model utility. Extensive experiments demonstrate that STShield successfully defends against various jailbreak attacks, while maintaining the model's performance on legitimate queries. Compared to existing approaches, STShield achieves superior defense performance with minimal computational overhead, making it a practical solution for real-world LLM deployment.
comment: 11 pages
☆ Debiasing Multimodal Large Language Models via Noise-Aware Preference Optimization CVPR 2025
Multimodal Large Language Models excel in various tasks, yet often struggle with modality bias, where the model tends to rely heavily on a single modality and overlook critical information in other modalities, which leads to incorrect focus and generating irrelevant responses. In this paper, we propose using the paradigm of preference optimization to solve the modality bias problem, including RLAIFVBias, a debiased preference optimization dataset, and a Noise Aware Preference Optimization algorithm. Specifically, we first construct the dataset by introducing perturbations to reduce the informational content of certain modalities, compelling the model to rely on a specific modality when generating negative responses. To address the inevitable noise in automatically constructed data, we combine the noise robust Mean Absolute Error with the Binary Cross Entropy in Direct Preference Optimization by a negative Box Cox transformation, and dynamically adjust the algorithm noise robustness based on the evaluated noise levels in the data. Extensive experiments validate our approach, demonstrating not only its effectiveness in mitigating modality bias but also its significant role in minimizing hallucinations.
comment: CVPR 2025
☆ WindowKV: Task-Adaptive Group-Wise KV Cache Window Selection for Efficient LLM Inference
With the advancements in long-context inference capabilities of large language models (LLMs), the KV cache has become one of the foundational components. However, its substantial GPU memory consumption makes KV cache compression a key technique for enabling efficient LLM inference in industrial scenarios. While recent studies have focused on optimizing the memory occupied by the KV cache, they overlook two critical factors: preserving semantic coherence and considering task-specific characteristic during compression. To address these limitations, we propose a novel task-adaptive KV cache window selection method, WindowKV. WindowKV dynamically selects local semantic windows consisting of consecutive tokens, according to task-specific characteristics, ensuring the retained KV cache captures continuous, essential context. Additionally, we introduce an intra-group layer KV cache indices sharing strategy to reduce computational overhead, achieving a balance between performance and efficiency. We rigorously evaluate WindowKV on the LongBench benchmark, and the results demonstrate that it maintains a performance comparable to full KV cache retention while using only 12% of the original KV cache, significantly reducing memory requirements. Furthermore, our method also achieves state-of-the-art results in the Needle-in-a-Haystack evaluation, highlighting its effectiveness and robustness.
☆ MedPlan:A Two-Stage RAG-Based System for Personalized Medical Plan Generation
Despite recent success in applying large language models (LLMs) to electronic health records (EHR), most systems focus primarily on assessment rather than treatment planning. We identify three critical limitations in current approaches: they generate treatment plans in a single pass rather than following the sequential reasoning process used by clinicians; they rarely incorporate patient-specific historical context; and they fail to effectively distinguish between subjective and objective clinical information. Motivated by the SOAP methodology (Subjective, Objective, Assessment, Plan), we introduce MedPlan, a novel framework that structures LLM reasoning to align with real-life clinician workflows. Our approach employs a two-stage architecture that first generates a clinical assessment based on patient symptoms and objective data, then formulates a structured treatment plan informed by this assessment and enriched with patient-specific information through retrieval-augmented generation. Comprehensive evaluation demonstrates that our method significantly outperforms baseline approaches in both assessment accuracy and treatment plan quality.
♻ ☆ DEPTH: Discourse Education through Pre-Training Hierarchically NAACL 2025
Language Models (LMs) struggle with linguistic understanding at the discourse level, even though discourse patterns such as coherence, cohesion, and narrative flow are prevalent in their pre-training data. To improve the discourse capabilities of LMs already at the pre-training stage, we introduce DEPTH, an encoder-decoder model that learns latent representations for sentences using a discourse-oriented pre-training objective. DEPTH combines hierarchical sentence representations with two objectives: (1) Sentence Un-Shuffling, and (2) Span-Corruption. Our approach trains the model to represent both sub-word-level and sentence-level dependencies over a pre-training corpora. When trained either from scratch or continuing from a pre-trained T5 checkpoint, DEPTH learns semantic and discourse-level representations faster than T5, outperforming it in span-corruption loss despite the additional sentence-un-shuffling objective. Evaluations on the GLUE, DiscoEval, and NI benchmarks demonstrate DEPTH's ability to quickly learn diverse downstream tasks, which require syntactic, semantic, and discourse capabilities. Our approach extends the discourse capabilities of T5, while minimally impacting other natural language understanding (NLU) capabilities in the resulting LM. We share our codebase for reproducibility: https://github.com/zbambergerNLP/depth.git.
comment: 25 pages, 10 figures, 10 tables, accepted to NAACL 2025, Rep4NLP
♻ ☆ Atom of Thoughts for Markov LLM Test-Time Scaling
Large Language Models (LLMs) achieve superior performance through training-time scaling, and test-time scaling further enhances their capabilities by conducting effective reasoning during inference. However, as the scale of reasoning increases, existing test-time scaling methods suffer from accumulated historical information, which not only wastes computational resources but also interferes with effective reasoning. To address this issue, we observe that complex reasoning can be achieved by solving a series of independent and self-contained subquestions. These subquestions are essentially \textit{atomic questions}, exhibiting the memoryless property similar to Markov processes. Based on this observation, we propose Atom of Thoughts (\our), where each state transition consists of decomposing the current question into a dependency-based directed acyclic graph and contracting its subquestions, forming a simplified question that maintains answer equivalence with the original problem. This answer preservation enables the iterative \textit{decomposition-contraction} process to naturally form a meaningful Markov reasoning process. Furthermore, these atomic states can be seamlessly integrated into existing test-time scaling methods, enabling \our to serve as a plug-in enhancement for improving reasoning capabilities. Experiments across six benchmarks demonstrate the effectiveness of \our both as a standalone framework and a plug-in enhancement. Notably, on HotpotQA, when applied to gpt-4o-mini, \our achieves an \textbf{80.6\%} F1 score, surpassing o3-mini by \textbf{3.4\%} and DeepSeek-R1 by \textbf{10.6\%}. The code is available at \href{https://github.com/qixucen/atom}{https://github.com/qixucen/atom}.
♻ ☆ Sneaking Syntax into Transformer Language Models with Tree Regularization NAACL 2025
While compositional accounts of human language understanding are based on a hierarchical tree-like process, neural models like transformers lack a direct inductive bias for such tree structures. Introducing syntactic inductive biases could unlock more robust and data-efficient learning in transformer language models (LMs), but existing methods for incorporating such structure greatly restrict models, either limiting their expressivity or increasing inference complexity. This work instead aims to softly inject syntactic inductive biases into given transformer circuits, through a structured regularizer. We introduce TreeReg, an auxiliary loss function that converts bracketing decisions from silver parses into a set of differentiable orthogonality constraints on vector hidden states. TreeReg integrates seamlessly with the standard LM objective, requiring no architectural changes. LMs pre-trained with TreeReg on natural language corpora such as WikiText-103 achieve up to 10% lower perplexities on out-of-distribution data and up to 9.5 point improvements in syntactic generalization, requiring less than half the training data to outperform standard LMs. TreeReg still provides gains for pre-trained LLMs: Continued pre-training of Sheared Llama with TreeReg results in improved syntactic generalization, and fine-tuning on MultiNLI with TreeReg mitigates degradation of performance on adversarial NLI benchmarks by 41.2 points. We release all code to guide future research.
comment: Accepted to NAACL 2025
♻ ☆ Speech Emotion Recognition with ASR Transcripts: A Comprehensive Study on Word Error Rate and Fusion Techniques
Text data is commonly utilized as a primary input to enhance Speech Emotion Recognition (SER) performance and reliability. However, the reliance on human-transcribed text in most studies impedes the development of practical SER systems, creating a gap between in-lab research and real-world scenarios where Automatic Speech Recognition (ASR) serves as the text source. Hence, this study benchmarks SER performance using ASR transcripts with varying Word Error Rates (WERs) from eleven models on three well-known corpora: IEMOCAP, CMU-MOSI, and MSP-Podcast. Our evaluation includes both text-only and bimodal SER with six fusion techniques, aiming for a comprehensive analysis that uncovers novel findings and challenges faced by current SER research. Additionally, we propose a unified ASR error-robust framework integrating ASR error correction and modality-gated fusion, achieving lower WER and higher SER results compared to the best-performing ASR transcript. These findings provide insights into SER with ASR assistance, especially for real-world applications.
comment: Accepted to IEEE SLT 2024. Fixed table and figure reference mistakes in the IEEE Xplore version
♻ ☆ SCULPT: Systematic Tuning of Long Prompts
Prompt optimization is essential for effective utilization of large language models (LLMs) across diverse tasks. While existing optimization methods are effective in optimizing short prompts, they struggle with longer, more complex ones, often risking information loss and being sensitive to small perturbations. To address these challenges, we propose SCULPT (Systematic Tuning of Long Prompts), a framework that treats prompt optimization as a hierarchical tree refinement problem. SCULPT represents prompts as tree structures, enabling targeted modifications while preserving contextual integrity. It employs a Critic-Actor framework that generates reflections and applies actions to refine the prompt. Evaluations demonstrate SCULPT's effectiveness on long prompts, its robustness to adversarial perturbations, and its ability to generate high-performing prompts even without any initial human-written prompt. Compared to existing state of the art methods, SCULPT consistently improves LLM performance by preserving essential task information while applying structured refinements. Both qualitative and quantitative analyses show that SCULPT produces more stable and interpretable prompt modifications, ensuring better generalization across tasks.
♻ ☆ From Knowledge Generation to Knowledge Verification: Examining the BioMedical Generative Capabilities of ChatGPT
The generative capabilities of LLM models offer opportunities for accelerating tasks but raise concerns about the authenticity of the knowledge they produce. To address these concerns, we present a computational approach that evaluates the factual accuracy of biomedical knowledge generated by an LLM. Our approach consists of two processes: generating disease-centric associations and verifying these associations using the semantic framework of biomedical ontologies. Using ChatGPT as the selected LLM, we designed prompt-engineering processes to establish linkages between diseases and related drugs, symptoms, and genes, and assessed consistency across multiple ChatGPT models (e.g., GPT-turbo, GPT-4, etc.). Experimental results demonstrate high accuracy in identifying disease terms (88%-97%), drug names (90%-91%), and genetic information (88%-98%). However, symptom term identification was notably lower (49%-61%), due to the informal and verbose nature of symptom descriptions, which hindered effective semantic matching with the formal language of specialized ontologies. Verification of associations reveals literature coverage rates of 89%-91% for disease-drug and disease-gene pairs, while symptom-related associations exhibit lower coverage (49%-62%).
comment: 28 pages, 6 figures, In Review with a Cell Press Journal
♻ ☆ Large Language Models: A Survey
Large Language Models (LLMs) have drawn a lot of attention due to their strong performance on a wide range of natural language tasks, since the release of ChatGPT in November 2022. LLMs' ability of general-purpose language understanding and generation is acquired by training billions of model's parameters on massive amounts of text data, as predicted by scaling laws \cite{kaplan2020scaling,hoffmann2022training}. The research area of LLMs, while very recent, is evolving rapidly in many different ways. In this paper, we review some of the most prominent LLMs, including three popular LLM families (GPT, LLaMA, PaLM), and discuss their characteristics, contributions and limitations. We also give an overview of techniques developed to build, and augment LLMs. We then survey popular datasets prepared for LLM training, fine-tuning, and evaluation, review widely used LLM evaluation metrics, and compare the performance of several popular LLMs on a set of representative benchmarks. Finally, we conclude the paper by discussing open challenges and future research directions.
♻ ☆ OASIS: Open Agent Social Interaction Simulations with One Million Agents
There has been a growing interest in enhancing rule-based agent-based models (ABMs) for social media platforms (i.e., X, Reddit) with more realistic large language model (LLM) agents, thereby allowing for a more nuanced study of complex systems. As a result, several LLM-based ABMs have been proposed in the past year. While they hold promise, each simulator is specifically designed to study a particular scenario, making it time-consuming and resource-intensive to explore other phenomena using the same ABM. Additionally, these models simulate only a limited number of agents, whereas real-world social media platforms involve millions of users. To this end, we propose OASIS, a generalizable and scalable social media simulator. OASIS is designed based on real-world social media platforms, incorporating dynamically updated environments (i.e., dynamic social networks and post information), diverse action spaces (i.e., following, commenting), and recommendation systems (i.e., interest-based and hot-score-based). Additionally, OASIS supports large-scale user simulations, capable of modeling up to one million users. With these features, OASIS can be easily extended to different social media platforms to study large-scale group phenomena and behaviors. We replicate various social phenomena, including information spreading, group polarization, and herd effects across X and Reddit platforms. Moreover, we provide observations of social phenomena at different agent group scales. We observe that the larger agent group scale leads to more enhanced group dynamics and more diverse and helpful agents' opinions. These findings demonstrate OASIS's potential as a powerful tool for studying complex systems in digital environments.
♻ ☆ Who is the root in a syntactic dependency structure?
The syntactic structure of a sentence can be described as a tree that indicates the syntactic relationships between words. In spite of significant progress in unsupervised methods that retrieve the syntactic structure of sentences, guessing the right direction of edges is still a challenge. As in a syntactic dependency structure edges are oriented away from the root, the challenge of guessing the right direction can be reduced to finding an undirected tree and the root. The limited performance of current unsupervised methods demonstrates the lack of a proper understanding of what a root vertex is from first principles. We consider an ensemble of centrality scores, some that only take into account the free tree (non-spatial scores) and others that take into account the position of vertices (spatial scores). We test the hypothesis that the root vertex is an important or central vertex of the syntactic dependency structure. We confirm that hypothesis and find that the best performance in guessing the root is achieved by novel scores that only take into account the position of a vertex and that of its neighbours. We provide theoretical and empirical foundations towards a universal notion of rootness from a network science perspective.
comment: All-subgraphs centrality added. Inconsistencies, ambiguities and errors have been corrected in Section 2. New tables of results section. Format of long tables improved. Language improved. Typos corrected
♻ ☆ VideoAutoArena: An Automated Arena for Evaluating Large Multimodal Models in Video Analysis through User Simulation CVPR 2025
Large multimodal models (LMMs) with advanced video analysis capabilities have recently garnered significant attention. However, most evaluations rely on traditional methods like multiple-choice questions in benchmarks such as VideoMME and LongVideoBench, which are prone to lack the depth needed to capture the complex demands of real-world users. To address this limitation-and due to the prohibitive cost and slow pace of human annotation for video tasks-we introduce VideoAutoArena, an arena-style benchmark inspired by LMSYS Chatbot Arena's framework, designed to automatically assess LMMs' video analysis abilities. VideoAutoArena utilizes user simulation to generate open-ended, adaptive questions that rigorously assess model performance in video understanding. The benchmark features an automated, scalable evaluation framework, incorporating a modified ELO Rating System for fair and continuous comparisons across multiple LMMs. To validate our automated judging system, we construct a 'gold standard' using a carefully curated subset of human annotations, demonstrating that our arena strongly aligns with human judgment while maintaining scalability. Additionally, we introduce a fault-driven evolution strategy, progressively increasing question complexity to push models toward handling more challenging video analysis scenarios. Experimental results demonstrate that VideoAutoArena effectively differentiates among state-of-the-art LMMs, providing insights into model strengths and areas for improvement. To further streamline our evaluation, we introduce VideoAutoBench as an auxiliary benchmark, where human annotators label winners in a subset of VideoAutoArena battles. We use GPT-4o as a judge to compare responses against these human-validated answers. Together, VideoAutoArena and VideoAutoBench offer a cost-effective, and scalable framework for evaluating LMMs in user-centric video analysis.
comment: CVPR 2025, Project Page: https://videoautoarena.github.io/
♻ ☆ CLASP: Contrastive Language-Speech Pretraining for Multilingual Multimodal Information Retrieval ECIR 2025
This study introduces CLASP (Contrastive Language-Speech Pretraining), a multilingual, multimodal representation tailored for audio-text information retrieval. CLASP leverages the synergy between spoken content and textual data. During training, we utilize our newly introduced speech-text dataset, which encompasses 15 diverse categories ranging from fiction to religion. CLASP's audio component integrates audio spectrograms with a pre-trained self-supervised speech model, while its language encoding counterpart employs a sentence encoder pre-trained on over 100 languages. This unified lightweight model bridges the gap between various modalities and languages, enhancing its effectiveness in handling and retrieving multilingual and multimodal data. Our evaluations across multiple languages demonstrate that CLASP establishes new benchmarks in HITS@1, MRR, and meanR metrics, outperforming traditional ASR-based retrieval methods that rely on transcribing speech into text for subsequent text retrieval, especially in specific scenarios.
comment: accepted at ECIR 2025, 13 pages, 4 figures
♻ ☆ Credit Risk Meets Large Language Models: Building a Risk Indicator from Loan Descriptions in P2P Lending
Peer-to-peer (P2P) lending connects borrowers and lenders through online platforms but suffers from significant information asymmetry, as lenders often lack sufficient data to assess borrowers' creditworthiness. This paper addresses this challenge by leveraging BERT, a Large Language Model (LLM) known for its ability to capture contextual nuances in text, to generate a risk score based on borrowers' loan descriptions using a dataset from the Lending Club platform. We fine-tune BERT to distinguish between defaulted and non-defaulted loans using the loan descriptions provided by the borrowers. The resulting BERT-generated risk score is then integrated as an additional feature into an XGBoost classifier used at the loan granting stage, where decision-makers have limited information available to guide their decisions. This integration enhances predictive performance, with improvements in balanced accuracy and AUC, highlighting the value of textual features in complementing traditional inputs. Moreover, we find that the incorporation of the BERT score alters how classification models utilize traditional input variables, with these changes varying by loan purpose. These findings suggest that BERT discerns meaningful patterns in loan descriptions, encompassing borrower-specific features, specific purposes, and linguistic characteristics. However, the inherent opacity of LLMs and their potential biases underscore the need for transparent frameworks to ensure regulatory compliance and foster trust. Overall, this study demonstrates how LLM-derived insights interact with traditional features in credit risk modeling, opening new avenues to enhance the explainability and fairness of these models.
♻ ☆ Nuclear Deployed: Analyzing Catastrophic Risks in Decision-making of Autonomous LLM Agents
Large language models (LLMs) are evolving into autonomous decision-makers, raising concerns about catastrophic risks in high-stakes scenarios, particularly in Chemical, Biological, Radiological and Nuclear (CBRN) domains. Based on the insight that such risks can originate from trade-offs between the agent's Helpful, Harmlessness and Honest (HHH) goals, we build a novel three-stage evaluation framework, which is carefully constructed to effectively and naturally expose such risks. We conduct 14,400 agentic simulations across 12 advanced LLMs, with extensive experiments and analysis. Results reveal that LLM agents can autonomously engage in catastrophic behaviors and deception, without being deliberately induced. Furthermore, stronger reasoning abilities often increase, rather than mitigate, these risks. We also show that these agents can violate instructions and superior commands. On the whole, we empirically prove the existence of catastrophic risks in autonomous LLM agents. We release our code to foster further research.
comment: Please visit https://llm-catastrophic-risks.github.io for a quick tour of our research. Our code is available at https://github.com/pillowsofwind/LLM-CBRN-Risks
♻ ☆ CMAT: A Multi-Agent Collaboration Tuning Framework for Enhancing Small Language Models
Open large language models (LLMs) have significantly advanced the field of natural language processing, showcasing impressive performance across various tasks.Despite the significant advancements in LLMs, their effective operation still relies heavily on human input to accurately guide the dialogue flow, with agent tuning being a crucial optimization technique that involves human adjustments to the model for better response to such guidance.Addressing this dependency, our work introduces the TinyAgent model, trained on a meticulously curated high-quality dataset. We also present the Collaborative Multi-Agent Tuning (CMAT) framework, an innovative system designed to augment language agent capabilities through adaptive weight updates based on environmental feedback. This framework fosters collaborative learning and real-time adaptation among multiple intelligent agents, enhancing their context-awareness and long-term memory. In this research, we propose a new communication agent framework that integrates multi-agent systems with environmental feedback mechanisms, offering a scalable method to explore cooperative behaviors. Notably, our TinyAgent-7B model exhibits performance on par with GPT-3.5, despite having fewer parameters, signifying a substantial improvement in the efficiency and effectiveness of LLMs.
♻ ☆ Stop Overthinking: A Survey on Efficient Reasoning for Large Language Models
Large Language Models (LLMs) have demonstrated remarkable capabilities in complex tasks. Recent advancements in Large Reasoning Models (LRMs), such as OpenAI o1 and DeepSeek-R1, have further improved performance in System-2 reasoning domains like mathematics and programming by harnessing supervised fine-tuning (SFT) and reinforcement learning (RL) techniques to enhance the Chain-of-Thought (CoT) reasoning. However, while longer CoT reasoning sequences improve performance, they also introduce significant computational overhead due to verbose and redundant outputs, known as the "overthinking phenomenon". In this paper, we provide the first structured survey to systematically investigate and explore the current progress toward achieving efficient reasoning in LLMs. Overall, relying on the inherent mechanism of LLMs, we categorize existing works into several key directions: (1) model-based efficient reasoning, which considers optimizing full-length reasoning models into more concise reasoning models or directly training efficient reasoning models; (2) reasoning output-based efficient reasoning, which aims to dynamically reduce reasoning steps and length during inference; (3) input prompts-based efficient reasoning, which seeks to enhance reasoning efficiency based on input prompt properties such as difficulty or length control. Additionally, we introduce the use of efficient data for training reasoning models, explore the reasoning capabilities of small language models, and discuss evaluation methods and benchmarking.
comment: Project Website: https://github.com/Eclipsess/Awesome-Efficient-Reasoning-LLMs
♻ ☆ EAGLE-3: Scaling up Inference Acceleration of Large Language Models via Training-Time Test
The sequential nature of modern LLMs makes them expensive and slow, and speculative sampling has proven to be an effective solution to this problem. Methods like EAGLE perform autoregression at the feature level, reusing top-layer features from the target model to achieve better results than vanilla speculative sampling. A growing trend in the LLM community is scaling up training data to improve model intelligence without increasing inference costs. However, we observe that scaling up data provides limited improvements for EAGLE. We identify that this limitation arises from EAGLE's feature prediction constraints. In this paper, we introduce EAGLE-3, which abandons feature prediction in favor of direct token prediction and replaces reliance on top-layer features with multi-layer feature fusion via a technique named training-time test. These improvements significantly enhance performance and enable the draft model to fully benefit from scaling up training data. Our experiments include both chat models and reasoning models, evaluated on five tasks. The results show that EAGLE-3 achieves a speedup ratio up to 6.5x, with about 1.4x improvement over EAGLE-2. In the SGLang framework, EAGLE-3 achieves a 1.38x throughput improvement at a batch size of 64. The code is available at https://github.com/SafeAILab/EAGLE.
♻ ☆ Exploring the Limits of Zero Shot Vision Language Models for Hate Meme Detection: The Vulnerabilities and their Interpretations
There is a rapid increase in the use of multimedia content in current social media platforms. One of the highly popular forms of such multimedia content are memes. While memes have been primarily invented to promote funny and buoyant discussions, malevolent users exploit memes to target individuals or vulnerable communities, making it imperative to identify and address such instances of hateful memes. Thus social media platforms are in dire need for active moderation of such harmful content. While manual moderation is extremely difficult due to the scale of such content, automatic moderation is challenged by the need of good quality annotated data to train hate meme detection algorithms. This makes a perfect pretext for exploring the power of modern day vision language models (VLMs) that have exhibited outstanding performance across various tasks. In this paper we study the effectiveness of VLMs in handling intricate tasks such as hate meme detection in a completely zero-shot setting so that there is no dependency on annotated data for the task. We perform thorough prompt engineering and query state-of-the-art VLMs using various prompt types to detect hateful/harmful memes. We further interpret the misclassification cases using a novel superpixel based occlusion method. Finally we show that these misclassifications can be neatly arranged into a typology of error classes the knowledge of which should enable the design of better safety guardrails in future.
♻ ☆ Word Sense Disambiguation in Persian: Can AI Finally Get It Right?
Homograph disambiguation, the task of distinguishing words with identical spellings but different meanings, poses a substantial challenge in natural language processing. In this study, we introduce a novel dataset tailored for Persian homograph disambiguation. Our work encompasses a thorough exploration of various embeddings, evaluated through the cosine similarity method and their efficacy in downstream tasks like classification. Our investigation entails training a diverse array of lightweight machine learning and deep learning models for phonograph disambiguation. We scrutinize the models' performance in terms of Accuracy, Recall, and F1 Score, thereby gaining insights into their respective strengths and limitations. The outcomes of our research underscore three key contributions. First, we present a newly curated Persian dataset, providing a solid foundation for future research in homograph disambiguation. Second, our comparative analysis of embeddings highlights their utility in different contexts, enriching the understanding of their capabilities. Third, by training and evaluating a spectrum of models, we extend valuable guidance for practitioners in selecting suitable strategies for homograph disambiguation tasks. In summary, our study unveils a new dataset, scrutinizes embeddings through diverse perspectives, and benchmarks various models for homograph disambiguation. These findings empower researchers and practitioners to navigate the intricate landscape of homograph-related challenges effectively.
♻ ☆ Autonomous Data Selection with Zero-shot Generative Classifiers for Mathematical Texts
We present Autonomous Data Selection (AutoDS), a method that leverages base language models themselves as zero-shot "generative classifiers" to automatically curate high-quality mathematical texts. Unlike prior approaches that require human annotations or training a dedicated data filter, AutoDS relies solely on a model's logits to determine whether a given passage is mathematically informative and educational. By integrating AutoDS into a continual pretraining pipeline, we substantially boost downstream performance on challenging math benchmarks (MATH, GSM8K, and BBH) while using far fewer tokens than previous methods. Empirically, our approach achieves roughly a twofold improvement in pretraining token efficiency over strong baselines, underscoring the potential of self-directed data selection in enhancing mathematical reasoning. We release our curated AutoMathText dataset to facilitate future research in automated domain-specific data curation. The AutoMathText dataset is available at https://huggingface.co/datasets/math-ai/AutoMathText. The code is available at https://github.com/yifanzhang-pro/AutoMathText.
comment: 24 pages, 8 figures. arXiv admin note: text overlap with arXiv:0808.2664, arXiv:0806.2159, arXiv:1703.08834, arXiv:math/0610707 by other authors
♻ ☆ Few-Shot Joint Multimodal Entity-Relation Extraction via Knowledge-Enhanced Cross-modal Prompt Model ACM MM 2024
Joint Multimodal Entity-Relation Extraction (JMERE) is a challenging task that aims to extract entities and their relations from text-image pairs in social media posts. Existing methods for JMERE require large amounts of labeled data. However, gathering and annotating fine-grained multimodal data for JMERE poses significant challenges. Initially, we construct diverse and comprehensive multimodal few-shot datasets fitted to the original data distribution. To address the insufficient information in the few-shot setting, we introduce the \textbf{K}nowledge-\textbf{E}nhanced \textbf{C}ross-modal \textbf{P}rompt \textbf{M}odel (KECPM) for JMERE. This method can effectively address the problem of insufficient information in the few-shot setting by guiding a large language model to generate supplementary background knowledge. Our proposed method comprises two stages: (1) a knowledge ingestion stage that dynamically formulates prompts based on semantic similarity guide ChatGPT generating relevant knowledge and employs self-reflection to refine the knowledge; (2) a knowledge-enhanced language model stage that merges the auxiliary knowledge with the original input and utilizes a transformer-based model to align with JMERE's required output format. We extensively evaluate our approach on a few-shot dataset derived from the JMERE dataset, demonstrating its superiority over strong baselines in terms of both micro and macro F$_1$ scores. Additionally, we present qualitative analyses and case studies to elucidate the effectiveness of our model.
comment: accepted by ACM MM 2024
Information Retrieval 12
☆ MammAlps: A multi-view video behavior monitoring dataset of wild mammals in the Swiss Alps CVPR 2025
Monitoring wildlife is essential for ecology and ethology, especially in light of the increasing human impact on ecosystems. Camera traps have emerged as habitat-centric sensors enabling the study of wildlife populations at scale with minimal disturbance. However, the lack of annotated video datasets limits the development of powerful video understanding models needed to process the vast amount of fieldwork data collected. To advance research in wild animal behavior monitoring we present MammAlps, a multimodal and multi-view dataset of wildlife behavior monitoring from 9 camera-traps in the Swiss National Park. MammAlps contains over 14 hours of video with audio, 2D segmentation maps and 8.5 hours of individual tracks densely labeled for species and behavior. Based on 6135 single animal clips, we propose the first hierarchical and multimodal animal behavior recognition benchmark using audio, video and reference scene segmentation maps as inputs. Furthermore, we also propose a second ecology-oriented benchmark aiming at identifying activities, species, number of individuals and meteorological conditions from 397 multi-view and long-term ecological events, including false positive triggers. We advocate that both tasks are complementary and contribute to bridging the gap between machine learning and ecology. Code and data are available at: https://github.com/eceo-epfl/MammAlps
comment: CVPR 2025; Benchmark and code at: https://github.com/eceo-epfl/MammAlps
☆ Causality-Aware Next Location Prediction Framework based on Human Mobility Stratification
Human mobility data are fused with multiple travel patterns and hidden spatiotemporal patterns are extracted by integrating user, location, and time information to improve next location prediction accuracy. In existing next location prediction methods, different causal relationships that result from patterns in human mobility data are ignored, which leads to confounding information that can have a negative effect on predictions. Therefore, this study introduces a causality-aware framework for next location prediction, focusing on human mobility stratification for travel patterns. In our research, a novel causal graph is developed that describes the relationships between various input variables. We use counterfactuals to enhance the indirect effects in our causal graph for specific travel patterns: non-anchor targeted travels. The proposed framework is designed as a plug-and-play module that integrates multiple next location prediction paradigms. We tested our proposed framework using several state-of-the-art models and human mobility datasets, and the results reveal that the proposed module improves the prediction performance. In addition, we provide results from the ablation study and quantitative study to demonstrate the soundness of our causal graph and its ability to further enhance the interpretability of the current next location prediction models.
comment: Accepted by IEEE UIC 2024
☆ GINGER: Grounded Information Nugget-Based Generation of Responses
Retrieval-augmented generation (RAG) faces challenges related to factual correctness, source attribution, and response completeness. To address them, we propose a modular pipeline for grounded response generation that operates on information nuggets-minimal, atomic units of relevant information extracted from retrieved documents. The multistage pipeline encompasses nugget detection, clustering, ranking, top cluster summarization, and fluency enhancement. It guarantees grounding in specific facts, facilitates source attribution, and ensures maximum information inclusion within length constraints. Extensive experiments on the TREC RAG'24 dataset evaluated with the AutoNuggetizer framework demonstrate that GINGER achieves state-of-the-art performance on this benchmark.
☆ BERTDetect: A Neural Topic Modelling Approach for Android Malware Detection
Web access today occurs predominantly through mobile devices, with Android representing a significant share of the mobile device market. This widespread usage makes Android a prime target for malicious attacks. Despite efforts to combat malicious attacks through tools like Google Play Protect and antivirus software, new and evolved malware continues to infiltrate Android devices. Source code analysis is effective but limited, as attackers quickly abandon old malware for new variants to evade detection. Therefore, there is a need for alternative methods that complement source code analysis. Prior research investigated clustering applications based on their descriptions and identified outliers in these clusters by API usage as malware. However, these works often used traditional techniques such as Latent Dirichlet Allocation (LDA) and k-means clustering, that do not capture the nuanced semantic structures present in app descriptions. To this end, in this paper, we propose BERTDetect, which leverages the BERTopic neural topic modelling to effectively capture the latent topics in app descriptions. The resulting topic clusters are comparatively more coherent than previous methods and represent the app functionalities well. Our results demonstrate that BERTDetect outperforms other baselines, achieving ~10% relative improvement in F1 score.
☆ SUNAR: Semantic Uncertainty based Neighborhood Aware Retrieval for Complex QA NAACL 2025
Complex question-answering (QA) systems face significant challenges in retrieving and reasoning over information that addresses multi-faceted queries. While large language models (LLMs) have advanced the reasoning capabilities of these systems, the bounded-recall problem persists, where procuring all relevant documents in first-stage retrieval remains a challenge. Missing pertinent documents at this stage leads to performance degradation that cannot be remedied in later stages, especially given the limited context windows of LLMs which necessitate high recall at smaller retrieval depths. In this paper, we introduce SUNAR, a novel approach that leverages LLMs to guide a Neighborhood Aware Retrieval process. SUNAR iteratively explores a neighborhood graph of documents, dynamically promoting or penalizing documents based on uncertainty estimates from interim LLM-generated answer candidates. We validate our approach through extensive experiments on two complex QA datasets. Our results show that SUNAR significantly outperforms existing retrieve-and-reason baselines, achieving up to a 31.84% improvement in performance over existing state-of-the-art methods for complex QA.
comment: Accepted at NAACL 2025 Main Conference
☆ Experience Retrieval-Augmentation with Electronic Health Records Enables Accurate Discharge QA
To improve the reliability of Large Language Models (LLMs) in clinical applications, retrieval-augmented generation (RAG) is extensively applied to provide factual medical knowledge. However, beyond general medical knowledge from open-ended datasets, clinical case-based knowledge is also critical for effective medical reasoning, as it provides context grounded in real-world patient experiences. Motivated by this, we propose Experience Retrieval Augmentation - ExpRAG framework based on Electronic Health Record (EHR), aiming to offer the relevant context from other patients' discharge reports. ExpRAG performs retrieval through a coarse-to-fine process, utilizing an EHR-based report ranker to efficiently identify similar patients, followed by an experience retriever to extract task-relevant content for enhanced medical reasoning. To evaluate ExpRAG, we introduce DischargeQA, a clinical QA dataset with 1,280 discharge-related questions across diagnosis, medication, and instruction tasks. Each problem is generated using EHR data to ensure realistic and challenging scenarios. Experimental results demonstrate that ExpRAG consistently outperforms a text-based ranker, achieving an average relative improvement of 5.2%, highlighting the importance of case-based knowledge for medical reasoning.
♻ ☆ Bayesian Active Learning for Multi-Criteria Comparative Judgement in Educational Assessment
Comparative Judgement (CJ) provides an alternative assessment approach by evaluating work holistically rather than breaking it into discrete criteria. This method leverages human ability to make nuanced comparisons, yielding more reliable and valid assessments. CJ aligns with real-world evaluations, where overall quality emerges from the interplay of various elements. However, rubrics remain widely used in education, offering structured criteria for grading and detailed feedback. This creates a gap between CJ's holistic ranking and the need for criterion-based performance breakdowns. This paper addresses this gap using a Bayesian approach. We build on Bayesian CJ (BCJ) by Gray et al., which directly models preferences instead of using likelihoods over total scores, allowing for expected ranks with uncertainty estimation. Their entropy-based active learning method selects the most informative pairwise comparisons for assessors. We extend BCJ to handle multiple independent learning outcome (LO) components, defined by a rubric, enabling both holistic and component-wise predictive rankings with uncertainty estimates. Additionally, we propose a method to aggregate entropies and identify the most informative comparison for assessors. Experiments on synthetic and real data demonstrate our method's effectiveness. Finally, we address a key limitation of BCJ, which is the inability to quantify assessor agreement. We show how to derive agreement levels, enhancing transparency in assessment.
♻ ☆ Balancing Content Size in RAG-Text2SQL System
Large Language Models (LLMs) have emerged as a promising solution for converting natural language queries into SQL commands, enabling seamless database interaction. However, these Text-to-SQL (Text2SQL) systems face inherent limitations, hallucinations, outdated knowledge, and untraceable reasoning. To address these challenges, the integration of retrieval-augmented generation (RAG) with Text2SQL models has gained traction. RAG serves as a retrieval mechanism, providing essential contextual information, such as table schemas and metadata, to enhance the query generation process. Despite their potential, RAG + Text2SQL systems are susceptible to the quality and size of retrieved documents. While richer document content can improve schema relevance and retrieval accuracy, it also introduces noise, increasing the risk of hallucinations and reducing query fidelity as the prompt size of the Text2SQL model increases. This research investigates the nuanced trade-off between document size and quality, aiming to strike a balance that optimizes system performance. Key thresholds are identified where performance degradation occurs, along with actionable strategies to mitigate these challenges. Additionally, we explore the phenomenon of hallucinations in Text2SQL models, emphasizing the critical role of curated document presentation in minimizing errors. Our findings provide a roadmap for enhancing the robustness of RAG + Text2SQL systems, offering practical insights for real-world applications.
♻ ☆ TSPRank: Bridging Pairwise and Listwise Methods with a Bilinear Travelling Salesman Model KDD 2025
Traditional Learning-To-Rank (LETOR) approaches, including pairwise methods like RankNet and LambdaMART, often fall short by solely focusing on pairwise comparisons, leading to sub-optimal global rankings. Conversely, deep learning based listwise methods, while aiming to optimise entire lists, require complex tuning and yield only marginal improvements over robust pairwise models. To overcome these limitations, we introduce Travelling Salesman Problem Rank (TSPRank), a hybrid pairwise-listwise ranking method. TSPRank reframes the ranking problem as a Travelling Salesman Problem (TSP), a well-known combinatorial optimisation challenge that has been extensively studied for its numerous solution algorithms and applications. This approach enables the modelling of pairwise relationships and leverages combinatorial optimisation to determine the listwise ranking. This approach can be directly integrated as an additional component into embeddings generated by existing backbone models to enhance ranking performance. Our extensive experiments across three backbone models on diverse tasks, including stock ranking, information retrieval, and historical events ordering, demonstrate that TSPRank significantly outperforms both pure pairwise and listwise methods. Our qualitative analysis reveals that TSPRank's main advantage over existing methods is its ability to harness global information better while ranking. TSPRank's robustness and superior performance across different domains highlight its potential as a versatile and effective LETOR solution.
comment: Accepted to ACM SIGKDD 2025 Research Track. The code and preprocessed data are available at https://github.com/waylonli/TSPRank-KDD2025
♻ ☆ From Knowledge Generation to Knowledge Verification: Examining the BioMedical Generative Capabilities of ChatGPT
The generative capabilities of LLM models offer opportunities for accelerating tasks but raise concerns about the authenticity of the knowledge they produce. To address these concerns, we present a computational approach that evaluates the factual accuracy of biomedical knowledge generated by an LLM. Our approach consists of two processes: generating disease-centric associations and verifying these associations using the semantic framework of biomedical ontologies. Using ChatGPT as the selected LLM, we designed prompt-engineering processes to establish linkages between diseases and related drugs, symptoms, and genes, and assessed consistency across multiple ChatGPT models (e.g., GPT-turbo, GPT-4, etc.). Experimental results demonstrate high accuracy in identifying disease terms (88%-97%), drug names (90%-91%), and genetic information (88%-98%). However, symptom term identification was notably lower (49%-61%), due to the informal and verbose nature of symptom descriptions, which hindered effective semantic matching with the formal language of specialized ontologies. Verification of associations reveals literature coverage rates of 89%-91% for disease-drug and disease-gene pairs, while symptom-related associations exhibit lower coverage (49%-62%).
comment: 28 pages, 6 figures, In Review with a Cell Press Journal
♻ ☆ Efficient Top-k s-Biplexes Search over Large Bipartite Graphs
In a bipartite graph, a subgraph is an $s$-biplex if each vertex of the subgraph is adjacent to all but at most $s$ vertices on the opposite set. The enumeration of $s$-biplexes from a given graph is a fundamental problem in bipartite graph analysis. However, in real-world data engineering, finding all $s$-biplexes is neither necessary nor computationally affordable. A more realistic problem is to identify some of the largest $s$-biplexes from the large input graph. We formulate the problem as the {\em top-$k$ $s$-biplex search (TBS) problem}, which aims to find the top-$k$ maximal $s$-biplexes with the most vertices, where $k$ is an input parameter. We prove that the TBS problem is NP-hard for any fixed $k\ge 1$. Then, we propose a branching algorithm, named MVBP, that breaks the simple $2^n$ enumeration algorithm. Furthermore, from a practical perspective, we investigate three techniques to improve the performance of MVBP: 2-hop decomposition, single-side bounds, and progressive search. Complexity analysis shows that the improved algorithm, named FastMVBP, has a running time $O^*(\gamma_s^{d_2})$, where $\gamma_s<2$, and $d_2$ is a parameter much smaller than the number of vertex in the sparse real-world graphs, e.g. $d_2$ is only $67$ in the AmazonRatings dataset which has more than $3$ million vertices. Finally, we conducted extensive experiments on eight real-world and synthetic datasets to demonstrate the empirical efficiency of the proposed algorithms. In particular, FastMVBP outperforms the benchmark algorithms by up to three orders of magnitude in several instances.
♻ ☆ CLASP: Contrastive Language-Speech Pretraining for Multilingual Multimodal Information Retrieval ECIR 2025
This study introduces CLASP (Contrastive Language-Speech Pretraining), a multilingual, multimodal representation tailored for audio-text information retrieval. CLASP leverages the synergy between spoken content and textual data. During training, we utilize our newly introduced speech-text dataset, which encompasses 15 diverse categories ranging from fiction to religion. CLASP's audio component integrates audio spectrograms with a pre-trained self-supervised speech model, while its language encoding counterpart employs a sentence encoder pre-trained on over 100 languages. This unified lightweight model bridges the gap between various modalities and languages, enhancing its effectiveness in handling and retrieving multilingual and multimodal data. Our evaluations across multiple languages demonstrate that CLASP establishes new benchmarks in HITS@1, MRR, and meanR metrics, outperforming traditional ASR-based retrieval methods that rely on transcribing speech into text for subsequent text retrieval, especially in specific scenarios.
comment: accepted at ECIR 2025, 13 pages, 4 figures
Multimedia 6
☆ MAO: Efficient Model-Agnostic Optimization of Prompt Tuning for Vision-Language Models ICME 2025
Though CLIP-based prompt tuning significantly enhances pre-trained Vision-Language Models, existing research focuses on reconstructing the model architecture, e.g., additional loss calculation and meta-networks. These approaches generally lead to increased complexity and extended training cost. To maintain the efficiency of the tuning process, we propose plug-and-play Model-Agnostic Optimization (MAO) for prompt tuning. Without altering any components of the prompt tuning backbone, we introduce a Data-Driven Enhancement framework to optimize the distribution of the initial data, and incorporate an Alterable Regularization module to boost the task-specific feature processing pipeline, thereby improving overall performance while maintaining low computational cost. Extensive experiments on MAO demonstrate its outstanding performance and efficiency. The code of MAO is available at: https://github.com/JREion/M.A.O .
comment: Accepted by the IEEE International Conference on Multimedia & Expo 2025 (ICME 2025); 12 pages, 6 figures, 8 tables
☆ Challenging Dataset and Multi-modal Gated Mixture of Experts Model for Remote Sensing Copy-Move Forgery Understanding
The Remote Sensing Copy-Move Question Answering (RSCMQA) task focuses on interpreting complex tampering scenarios and inferring the relationships between objects. Currently, publicly available datasets often use randomly generated tampered images, which lack spatial logic and do not meet the practical needs of defense security and land resource monitoring. To address this, we propose a high-quality manually annotated RSCMQA dataset, Real-RSCM, which provides more realistic evaluation metrics for the identification and understanding of remote sensing image tampering. The tampered images in the Real-RSCM dataset are subtle, authentic, and challenging, posing significant difficulties for model discrimination capabilities. To overcome these challenges, we introduce a multimodal gated mixture of experts model (CM-MMoE), which guides multi-expert models to discern tampered information in images through multi-level visual semantics and textual joint modeling. Extensive experiments demonstrate that CM-MMoE provides a stronger benchmark for the RSCMQA task compared to general VQA and CMQA models. Our dataset and code are available at https://anonymous.4open.science/r/CM-MMoE.
comment: 6 pages, 6 figures
☆ Anomaly Detection and Localization for Speech Deepfakes via Feature Pyramid Matching
The rise of AI-driven generative models has enabled the creation of highly realistic speech deepfakes - synthetic audio signals that can imitate target speakers' voices - raising critical security concerns. Existing methods for detecting speech deepfakes primarily rely on supervised learning, which suffers from two critical limitations: limited generalization to unseen synthesis techniques and a lack of explainability. In this paper, we address these issues by introducing a novel interpretable one-class detection framework, which reframes speech deepfake detection as an anomaly detection task. Our model is trained exclusively on real speech to characterize its distribution, enabling the classification of out-of-distribution samples as synthetically generated. Additionally, our framework produces interpretable anomaly maps during inference, highlighting anomalous regions across both time and frequency domains. This is done through a Student-Teacher Feature Pyramid Matching system, enhanced with Discrepancy Scaling to improve generalization capabilities across unseen data distributions. Extensive evaluations demonstrate the superior performance of our approach compared to the considered baselines, validating the effectiveness of framing speech deepfake detection as an anomaly detection problem.
☆ Semi-supervised Semantic Segmentation with Multi-Constraint Consistency Learning
Consistency regularization has prevailed in semi-supervised semantic segmentation and achieved promising performance. However, existing methods typically concentrate on enhancing the Image-augmentation based Prediction consistency and optimizing the segmentation network as a whole, resulting in insufficient utilization of potential supervisory information. In this paper, we propose a Multi-Constraint Consistency Learning (MCCL) approach to facilitate the staged enhancement of the encoder and decoder. Specifically, we first design a feature knowledge alignment (FKA) strategy to promote the feature consistency learning of the encoder from image-augmentation. Our FKA encourages the encoder to derive consistent features for strongly and weakly augmented views from the perspectives of point-to-point alignment and prototype-based intra-class compactness. Moreover, we propose a self-adaptive intervention (SAI) module to increase the discrepancy of aligned intermediate feature representations, promoting Feature-perturbation based Prediction consistency learning. Self-adaptive feature masking and noise injection are designed in an instance-specific manner to perturb the features for robust learning of the decoder. Experimental results on Pascal VOC2012 and Cityscapes datasets demonstrate that our proposed MCCL achieves new state-of-the-art performance. The source code and models are made available at https://github.com/NUST-Machine-Intelligence-Laboratory/MCCL.
comment: accepted by IEEE Transactions on Multimedia
♻ ☆ Speech Emotion Recognition with ASR Transcripts: A Comprehensive Study on Word Error Rate and Fusion Techniques
Text data is commonly utilized as a primary input to enhance Speech Emotion Recognition (SER) performance and reliability. However, the reliance on human-transcribed text in most studies impedes the development of practical SER systems, creating a gap between in-lab research and real-world scenarios where Automatic Speech Recognition (ASR) serves as the text source. Hence, this study benchmarks SER performance using ASR transcripts with varying Word Error Rates (WERs) from eleven models on three well-known corpora: IEMOCAP, CMU-MOSI, and MSP-Podcast. Our evaluation includes both text-only and bimodal SER with six fusion techniques, aiming for a comprehensive analysis that uncovers novel findings and challenges faced by current SER research. Additionally, we propose a unified ASR error-robust framework integrating ASR error correction and modality-gated fusion, achieving lower WER and higher SER results compared to the best-performing ASR transcript. These findings provide insights into SER with ASR assistance, especially for real-world applications.
comment: Accepted to IEEE SLT 2024. Fixed table and figure reference mistakes in the IEEE Xplore version
♻ ☆ VideoAutoArena: An Automated Arena for Evaluating Large Multimodal Models in Video Analysis through User Simulation CVPR 2025
Large multimodal models (LMMs) with advanced video analysis capabilities have recently garnered significant attention. However, most evaluations rely on traditional methods like multiple-choice questions in benchmarks such as VideoMME and LongVideoBench, which are prone to lack the depth needed to capture the complex demands of real-world users. To address this limitation-and due to the prohibitive cost and slow pace of human annotation for video tasks-we introduce VideoAutoArena, an arena-style benchmark inspired by LMSYS Chatbot Arena's framework, designed to automatically assess LMMs' video analysis abilities. VideoAutoArena utilizes user simulation to generate open-ended, adaptive questions that rigorously assess model performance in video understanding. The benchmark features an automated, scalable evaluation framework, incorporating a modified ELO Rating System for fair and continuous comparisons across multiple LMMs. To validate our automated judging system, we construct a 'gold standard' using a carefully curated subset of human annotations, demonstrating that our arena strongly aligns with human judgment while maintaining scalability. Additionally, we introduce a fault-driven evolution strategy, progressively increasing question complexity to push models toward handling more challenging video analysis scenarios. Experimental results demonstrate that VideoAutoArena effectively differentiates among state-of-the-art LMMs, providing insights into model strengths and areas for improvement. To further streamline our evaluation, we introduce VideoAutoBench as an auxiliary benchmark, where human annotators label winners in a subset of VideoAutoArena battles. We use GPT-4o as a judge to compare responses against these human-validated answers. Together, VideoAutoArena and VideoAutoBench offer a cost-effective, and scalable framework for evaluating LMMs in user-centric video analysis.
comment: CVPR 2025, Project Page: https://videoautoarena.github.io/
Robotics 24
☆ A Robot-Led Intervention for Emotion Regulation: From Expression to Reappraisal
Emotion regulation is a crucial skill for managing emotions in everyday life, yet finding a constructive and accessible method to support these processes remains challenging due to their cognitive demands. In this study, we explore how regular interactions with a social robot, conducted in a structured yet familiar environment within university halls and departments, can provide effective support for emotion regulation through cognitive reappraisal. Twenty-one students participated in a five-session study at a university hall or department, where the robot facilitated structured conversations, encouraging the students to reinterpret emotionally charged situations that they shared with the robot. Quantitative and qualitative results indicate significant improvements in emotion self-regulation, with participants reporting better understanding and control of their emotions. The intervention led to significant changes in constructive emotion regulation tendencies and positive effects on mood and sentiment after each session. The findings also demonstrate that repeated interactions with the robot encouraged greater emotional expressiveness, including longer speech disclosures, increased use of affective language, and heightened facial arousal. Notably, expressiveness followed structured patterns aligned with the reappraisal process, with expression peaking during key reappraisal moments, particularly when participants were prompted to reinterpret negative experiences. The qualitative feedback further highlighted how the robot fostered introspection and provided a supportive space for discussing emotions, enabling participants to confront long-avoided emotional challenges. These findings demonstrate the potential of robots to effectively assist in emotion regulation in familiar environments, offering both emotional support and cognitive guidance.
☆ Decentralized Navigation of a Cable-Towed Load using Quadrupedal Robot Team via MARL
This work addresses the challenge of enabling a team of quadrupedal robots to collaboratively tow a cable-connected load through cluttered and unstructured environments while avoiding obstacles. Leveraging cables allows the multi-robot system to navigate narrow spaces by maintaining slack when necessary. However, this introduces hybrid physical interactions due to alternating taut and slack states, with computational complexity that scales exponentially as the number of agents increases. To tackle these challenges, we developed a scalable and decentralized system capable of dynamically coordinating a variable number of quadrupedal robots while managing the hybrid physical interactions inherent in the load-towing task. At the core of this system is a novel multi-agent reinforcement learning (MARL)-based planner, designed for decentralized coordination. The MARL-based planner is trained using a centralized training with decentralized execution (CTDE) framework, enabling each robot to make decisions autonomously using only local (ego) observations. To accelerate learning and ensure effective collaboration across varying team sizes, we introduce a tailored training curriculum for MARL. Experimental results highlight the flexibility and scalability of the framework, demonstrating successful deployment with one to four robots in real-world scenarios and up to twelve robots in simulation. The decentralized planner maintains consistent inference times, regardless of the team size. Additionally, the proposed system demonstrates robustness to environment perturbations and adaptability to varying load weights. This work represents a step forward in achieving flexible and efficient multi-legged robotic collaboration in complex and real-world environments.
☆ Extended Visibility of Autonomous Vehicles via Optimized Cooperative Perception under Imperfect Communication
Autonomous Vehicles (AVs) rely on individual perception systems to navigate safely. However, these systems face significant challenges in adverse weather conditions, complex road geometries, and dense traffic scenarios. Cooperative Perception (CP) has emerged as a promising approach to extending the perception quality of AVs by jointly processing shared camera feeds and sensor readings across multiple vehicles. This work presents a novel CP framework designed to optimize vehicle selection and networking resource utilization under imperfect communications. Our optimized CP formation considers critical factors such as the helper vehicles' spatial position, visual range, motion blur, and available communication budgets. Furthermore, our resource optimization module allocates communication channels while adjusting power levels to maximize data flow efficiency between the ego and helper vehicles, considering realistic models of modern vehicular communication systems, such as LTE and 5G NR-V2X. We validate our approach through extensive experiments on pedestrian detection in challenging scenarios, using synthetic data generated by the CARLA simulator. The results demonstrate that our method significantly improves upon the perception quality of individual AVs with about 10% gain in detection accuracy. This substantial gain uncovers the unleashed potential of CP to enhance AV safety and performance in complex situations.
comment: 55 pages, 13 figures, 3 tables, Elsevier Journal
☆ Joint State-Parameter Observer-Based Robust Control of a UAV for Heavy Load Transportation
This paper proposes a joint state-parameter observer-based controller for trajectory tracking of an octocopter unmanned aerial vehicle (OUAV), for transportation of a heavy load with unknown mass and size. The multi-body dynamic model of the OUAV with a rigidly attached load is obtained, effectively considering the effects of the load parameters into the dynamics of the system. A robust nonlinear W-infinity control strategy is designed for optimal trajectory tracking of the OUAV, with information of the states and load parameters provided by a joint estimation unscented Kalman filter. The effectiveness of the proposed strategy is corroborated by numerical results.
comment: 12 pages, 3 figures. This is a preprint of a paper presented at the 2023 Conference on Climbing and Walking Robots (CLAWAR) and published later by Springer Nature Switzerland
☆ Unraveling the Effects of Synthetic Data on End-to-End Autonomous Driving
End-to-end (E2E) autonomous driving (AD) models require diverse, high-quality data to perform well across various driving scenarios. However, collecting large-scale real-world data is expensive and time-consuming, making high-fidelity synthetic data essential for enhancing data diversity and model robustness. Existing driving simulators for synthetic data generation have significant limitations: game-engine-based simulators struggle to produce realistic sensor data, while NeRF-based and diffusion-based methods face efficiency challenges. Additionally, recent simulators designed for closed-loop evaluation provide limited interaction with other vehicles, failing to simulate complex real-world traffic dynamics. To address these issues, we introduce SceneCrafter, a realistic, interactive, and efficient AD simulator based on 3D Gaussian Splatting (3DGS). SceneCrafter not only efficiently generates realistic driving logs across diverse traffic scenarios but also enables robust closed-loop evaluation of end-to-end models. Experimental results demonstrate that SceneCrafter serves as both a reliable evaluation platform and a efficient data generator that significantly improves end-to-end model generalization.
☆ PanopticSplatting: End-to-End Panoptic Gaussian Splatting
Open-vocabulary panoptic reconstruction is a challenging task for simultaneous scene reconstruction and understanding. Recently, methods have been proposed for 3D scene understanding based on Gaussian splatting. However, these methods are multi-staged, suffering from the accumulated errors and the dependence of hand-designed components. To streamline the pipeline and achieve global optimization, we propose PanopticSplatting, an end-to-end system for open-vocabulary panoptic reconstruction. Our method introduces query-guided Gaussian segmentation with local cross attention, lifting 2D instance masks without cross-frame association in an end-to-end way. The local cross attention within view frustum effectively reduces the training memory, making our model more accessible to large scenes with more Gaussians and objects. In addition, to address the challenge of noisy labels in 2D pseudo masks, we propose label blending to promote consistent 3D segmentation with less noisy floaters, as well as label warping on 2D predictions which enhances multi-view coherence and segmentation accuracy. Our method demonstrates strong performances in 3D scene panoptic reconstruction on the ScanNet-V2 and ScanNet++ datasets, compared with both NeRF-based and Gaussian-based panoptic reconstruction methods. Moreover, PanopticSplatting can be easily generalized to numerous variants of Gaussian splatting, and we demonstrate its robustness on different Gaussian base models.
comment: 8 pages, 6 figures
☆ Unseen from Seen: Rewriting Observation-Instruction Using Foundation Models for Augmenting Vision-Language Navigation
Data scarcity is a long-standing challenge in the Vision-Language Navigation (VLN) field, which extremely hinders the generalization of agents to unseen environments. Previous works primarily rely on additional simulator data or web-collected images/videos to improve the generalization. However, the simulator environments still face limited diversity, and the web-collected data often requires extensive labor to remove the noise. In this paper, we propose a Rewriting-driven AugMentation (RAM) paradigm for VLN, which directly creates the unseen observation-instruction pairs via rewriting human-annotated training data. Benefiting from our rewriting mechanism, new observation-instruction can be obtained in both simulator-free and labor-saving manners to promote generalization. Specifically, we first introduce Object-Enriched Observation Rewriting, where we combine Vision-Language Models (VLMs) and Large Language Models (LLMs) to derive rewritten object-enriched scene descriptions, enabling observation synthesis with diverse objects and spatial layouts via Text-to-Image Generation Models (T2IMs). Then, we propose Observation-Contrast Instruction Rewriting, which generates observation-aligned rewritten instructions by requiring LLMs to reason the difference between original and new observations. We further develop a mixing-then-focusing training strategy with a random observation cropping scheme, effectively enhancing data distribution diversity while suppressing augmentation data noise during training. Experiments on both the discrete environments (R2R, REVERIE, and R4R datasets) and continuous environments (R2R-CE dataset) show the superior performance and impressive generalization ability of our method. Code is available at https://github.com/SaDil13/VLN-RAM.
☆ Assist-as-needed Hip Exoskeleton Control for Gait Asymmetry Correction via Human-in-the-loop Optimization
Gait asymmetry is a significant clinical characteristic of hemiplegic gait that most stroke survivors suffer, leading to limited mobility and long-term negative impacts on their quality of life. Although a variety of exoskeleton controls have been developed for robot-assisted gait rehabilitation, little attention has been paid to correcting the gait asymmetry of stroke patients following the assist-as-need (AAN) principle, and it is still challenging to properly share control between the exoskeleton and stroke patients with partial motor control. In view of this, this article proposes an AAN hip exoskeleton control with human-in-the-loop optimization to correct gait asymmetry in stroke patients. To realize the AAN concept, an objective function was designed for real-time evaluation of the subject's gait performance and active participation, which considers the variability of natural human movement and guides the online tuning of control parameters on a subject-specific basis. In this way, patients were stimulated to contribute as much as possible to movement, thus maximizing the efficiency and outcomes of post-stroke gait rehabilitation. Finally, an experimental study was conducted to verify the feasibility and effectiveness of the proposed AAN control on healthy subjects with artificial gait impairment. For the first time, the common hypothesis that AAN controls can improve human active participation was validated from the biomechanics viewpoint.
☆ Optimizing Navigation And Chemical Application in Precision Agriculture With Deep Reinforcement Learning And Conditional Action Tree
This paper presents a novel reinforcement learning (RL)-based planning scheme for optimized robotic management of biotic stresses in precision agriculture. The framework employs a hierarchical decision-making structure with conditional action masking, where high-level actions direct the robot's exploration, while low-level actions optimize its navigation and efficient chemical spraying in affected areas. The key objectives of optimization include improving the coverage of infected areas with limited battery power and reducing chemical usage, thus preventing unnecessary spraying of healthy areas of the field. Our numerical experimental results demonstrate that the proposed method, Hierarchical Action Masking Proximal Policy Optimization (HAM-PPO), significantly outperforms baseline practices, such as LawnMower navigation + indiscriminate spraying (Carpet Spray), in terms of yield recovery and resource efficiency. HAM-PPO consistently achieves higher yield recovery percentages and lower chemical costs across a range of infection scenarios. The framework also exhibits robustness to observation noise and generalizability under diverse environmental conditions, adapting to varying infection ranges and spatial distribution patterns.
comment: 32 pages, 9 figures
☆ PhysTwin: Physics-Informed Reconstruction and Simulation of Deformable Objects from Videos
Creating a physical digital twin of a real-world object has immense potential in robotics, content creation, and XR. In this paper, we present PhysTwin, a novel framework that uses sparse videos of dynamic objects under interaction to produce a photo- and physically realistic, real-time interactive virtual replica. Our approach centers on two key components: (1) a physics-informed representation that combines spring-mass models for realistic physical simulation, generative shape models for geometry, and Gaussian splats for rendering; and (2) a novel multi-stage, optimization-based inverse modeling framework that reconstructs complete geometry, infers dense physical properties, and replicates realistic appearance from videos. Our method integrates an inverse physics framework with visual perception cues, enabling high-fidelity reconstruction even from partial, occluded, and limited viewpoints. PhysTwin supports modeling various deformable objects, including ropes, stuffed animals, cloth, and delivery packages. Experiments show that PhysTwin outperforms competing methods in reconstruction, rendering, future prediction, and simulation under novel interactions. We further demonstrate its applications in interactive real-time simulation and model-based robotic motion planning.
comment: Project Page: https://jianghanxiao.github.io/phystwin-web/
☆ Adaptive Koopman Model Predictive Control of Simple Serial Robots IROS 2025
Approximating nonlinear systems as linear ones is a common workaround to apply control tools tailored for linear systems. This motivates our present work where we developed a data-driven model predictive controller (MPC) based on the Koopman operator framework, allowing the embedding of nonlinear dynamics in a higher dimensional, but linear function space. The controller, termed adaptive Koopman model predictive control (KMPC), uses online closed-loop feedback to learn and incrementally update a linear representation of nonlinear system dynamics, without the prior knowledge of a model. Adaptive KMPC differs from most other Koopman-based control frameworks that aim to identify high-validity-range models in advance and then enter closed-loop control without further model adaptations. To validate the controller, trajectory tracking experiments are conducted with 1R and 2R robots under force disturbances and changing model parameters. We compare the controller to classical linearization MPC and Koopman-based MPC without model updates, denoted static KMPC. The results show that adaptive KMPC can, opposed to static KMPC, generalize over unforeseen force disturbances and can, opposed to linearization MPC, handle varying dynamic parameters, while using a small set of basis functions to approximate the Koopman operator.
comment: Preprint submitted to IROS 2025; See supplementary material at https://github.com/adrianodelr/adaptive-koopman-mpc
☆ SG-Tailor: Inter-Object Commonsense Relationship Reasoning for Scene Graph Manipulation
Scene graphs capture complex relationships among objects, serving as strong priors for content generation and manipulation. Yet, reasonably manipulating scene graphs -- whether by adding nodes or modifying edges -- remains a challenging and untouched task. Tasks such as adding a node to the graph or reasoning about a node's relationships with all others are computationally intractable, as even a single edge modification can trigger conflicts due to the intricate interdependencies within the graph. To address these challenges, we introduce SG-Tailor, an autoregressive model that predicts the conflict-free relationship between any two nodes. SG-Tailor not only infers inter-object relationships, including generating commonsense edges for newly added nodes but also resolves conflicts arising from edge modifications to produce coherent, manipulated graphs for downstream tasks. For node addition, the model queries the target node and other nodes from the graph to predict the appropriate relationships. For edge modification, SG-Tailor employs a Cut-And-Stitch strategy to solve the conflicts and globally adjust the graph. Extensive experiments demonstrate that SG-Tailor outperforms competing methods by a large margin and can be seamlessly integrated as a plug-in module for scene generation and robotic manipulation tasks.
comment: The code will be available at https://github.com/josef5838/SG-Tailor
♻ ☆ Distributed Bayesian Estimation in Sensor Networks: Consensus on Marginal Densities
In this paper, we aim to design and analyze distributed Bayesian estimation algorithms for sensor networks. The challenges we address are to (i) derive a distributed provably-correct algorithm in the functional space of probability distributions over continuous variables, and (ii) leverage these results to obtain new distributed estimators restricted to subsets of variables observed by individual agents. This relates to applications such as cooperative localization and federated learning, where the data collected at any agent depends on a subset of all variables of interest. We present Bayesian density estimation algorithms using data from non-linear likelihoods at agents in centralized, distributed, and marginal distributed settings. After setting up a distributed estimation objective, we prove almost-sure convergence to the optimal set of pdfs at each agent. Then, we prove the same for a storage-aware algorithm estimating densities only over relevant variables at each agent. Finally, we present a Gaussian version of these algorithms and implement it in a mapping problem using variational inference to handle non-linear likelihood models associated with LiDAR sensing.
♻ ☆ Socially-Aware Robot Navigation Enhanced by Bidirectional Natural Language Conversations Using Large Language Models
Robot navigation is crucial across various domains, yet traditional methods focus on efficiency and obstacle avoidance, often overlooking human behavior in shared spaces. With the rise of service robots, socially aware navigation has gained prominence. However, existing approaches primarily predict pedestrian movements or issue alerts, lacking true human-robot interaction. We introduce Hybrid Soft Actor-Critic with Large Language Model (HSAC-LLM), a novel framework for socially aware navigation. By integrating deep reinforcement learning with large language models, HSAC-LLM enables bidirectional natural language interactions, predicting both continuous and discrete navigation actions. When potential collisions arise, the robot proactively communicates with pedestrians to determine avoidance strategies. Experiments in 2D simulation, Gazebo, and real-world environments demonstrate that HSAC-LLM outperforms state-of-the-art DRL methods in interaction, navigation, and obstacle avoidance. This paradigm advances effective human-robot interactions in dynamic settings. Videos are available at https://hsacllm.github.io/.
♻ ☆ Towards Safe Mid-Air Drone Interception: Strategies for Tracking & Capture
A unique approach for the mid-air autonomous aerial interception of non-cooperating UAV by a flying robot equipped with a net is presented in this paper. A novel interception guidance method dubbed EPN is proposed, designed to catch agile maneuvering targets while relying on onboard state estimation and tracking. The proposed method is compared with state-of-the-art approaches in simulations using 100 different trajectories of the target with varying complexity comprising almost 14 hours of flight data, and EPN demonstrates the shortest response time and the highest number of interceptions, which are key parameters of agile interception. To enable robust transfer from theory and simulation to a real-world implementation, we aim to avoid overfitting to specific assumptions about the target, and to tackle interception of a target following an unknown general trajectory. Furthermore, we identify several often overlooked problems related to tracking and estimation of the target's state that can have a significant influence on the overall performance of the system. We propose the use of a novel state estimation filter based on the IMM filter and a new measurement model. Simulated experiments show that the proposed solution provides significant improvements in estimation accuracy over the commonly employed KF approaches when considering general trajectories. Based on these results, we employ the proposed filtering and guidance methods to implement a complete autonomous interception system, which is thoroughly evaluated in realistic simulations and tested in real-world experiments with a maneuvering target going far beyond the performance of any state-of-the-art solution.
♻ ☆ ModeSeq: Taming Sparse Multimodal Motion Prediction with Sequential Mode Modeling CVPR 2025
Anticipating the multimodality of future events lays the foundation for safe autonomous driving. However, multimodal motion prediction for traffic agents has been clouded by the lack of multimodal ground truth. Existing works predominantly adopt the winner-take-all training strategy to tackle this challenge, yet still suffer from limited trajectory diversity and uncalibrated mode confidence. While some approaches address these limitations by generating excessive trajectory candidates, they necessitate a post-processing stage to identify the most representative modes, a process lacking universal principles and compromising trajectory accuracy. We are thus motivated to introduce ModeSeq, a new multimodal prediction paradigm that models modes as sequences. Unlike the common practice of decoding multiple plausible trajectories in one shot, ModeSeq requires motion decoders to infer the next mode step by step, thereby more explicitly capturing the correlation between modes and significantly enhancing the ability to reason about multimodality. Leveraging the inductive bias of sequential mode prediction, we also propose the Early-Match-Take-All (EMTA) training strategy to diversify the trajectories further. Without relying on dense mode prediction or heuristic post-processing, ModeSeq considerably improves the diversity of multimodal output while attaining satisfactory trajectory accuracy, resulting in balanced performance on motion prediction benchmarks. Moreover, ModeSeq naturally emerges with the capability of mode extrapolation, which supports forecasting more behavior modes when the future is highly uncertain.
comment: CVPR 2025
♻ ☆ Omnidirectional Multi-Object Tracking CVPR 2025
Panoramic imagery, with its 360{\deg} field of view, offers comprehensive information to support Multi-Object Tracking (MOT) in capturing spatial and temporal relationships of surrounding objects. However, most MOT algorithms are tailored for pinhole images with limited views, impairing their effectiveness in panoramic settings. Additionally, panoramic image distortions, such as resolution loss, geometric deformation, and uneven lighting, hinder direct adaptation of existing MOT methods, leading to significant performance degradation. To address these challenges, we propose OmniTrack, an omnidirectional MOT framework that incorporates Tracklet Management to introduce temporal cues, FlexiTrack Instances for object localization and association, and the CircularStatE Module to alleviate image and geometric distortions. This integration enables tracking in panoramic field-of-view scenarios, even under rapid sensor motion. To mitigate the lack of panoramic MOT datasets, we introduce the QuadTrack dataset--a comprehensive panoramic dataset collected by a quadruped robot, featuring diverse challenges such as panoramic fields of view, intense motion, and complex environments. Extensive experiments on the public JRDB dataset and the newly introduced QuadTrack benchmark demonstrate the state-of-the-art performance of the proposed framework. OmniTrack achieves a HOTA score of 26.92% on JRDB, representing an improvement of 3.43%, and further achieves 23.45% on QuadTrack, surpassing the baseline by 6.81%. The established dataset and source code are available at https://github.com/xifen523/OmniTrack.
comment: Accepted to CVPR 2025. The established dataset and source code are available at https://github.com/xifen523/OmniTrack
♻ ☆ A Data-Centric Revisit of Pre-Trained Vision Models for Robot Learning CVPR 2025
Pre-trained vision models (PVMs) are fundamental to modern robotics, yet their optimal configuration remains unclear. Through systematic evaluation, we find that while DINO and iBOT outperform MAE across visuomotor control and perception tasks, they struggle when trained on non-(single-)object-centric (NOC) data--a limitation strongly correlated with their diminished ability to learn object-centric representations. This investigation indicates that the ability to form object-centric representations from the non-object-centric robotics dataset is the key to success for PVMs. Motivated by this discovery, we designed SlotMIM, a method that induces object-centric representations by introducing a semantic bottleneck to reduce the number of prototypes to encourage the emergence of objectness as well as cross-view consistency regularization for encouraging multiview invariance. Our experiments encompass pre-training on object-centric, scene-centric, web-crawled, and ego-centric data. Across all settings, our approach learns transferrable representations and achieves significant improvements over prior work in image recognition, scene understanding, and robot learning evaluations. When scaled up with million-scale datasets, our method also demonstrates superior data efficiency and scalability. Our code and models are publicly available at https://github.com/CVMI-Lab/SlotMIM.
comment: Accepted by CVPR 2025
♻ ☆ One-Shot Manipulation Strategy Learning by Making Contact Analogies ICRA 2025
We present a novel approach, MAGIC (manipulation analogies for generalizable intelligent contacts), for one-shot learning of manipulation strategies with fast and extensive generalization to novel objects. By leveraging a reference action trajectory, MAGIC effectively identifies similar contact points and sequences of actions on novel objects to replicate a demonstrated strategy, such as using different hooks to retrieve distant objects of different shapes and sizes. Our method is based on a two-stage contact-point matching process that combines global shape matching using pretrained neural features with local curvature analysis to ensure precise and physically plausible contact points. We experiment with three tasks including scooping, hanging, and hooking objects. MAGIC demonstrates superior performance over existing methods, achieving significant improvements in runtime speed and generalization to different object categories. Website: https://magic-2024.github.io/ .
comment: ICRA 2025; CoRL LEAP Workshop, 2024
♻ ☆ Intelligent LiDAR Navigation: Leveraging External Information and Semantic Maps with LLM as Copilot
Traditional robot navigation systems primarily utilize occupancy grid maps and laser-based sensing technologies, as demonstrated by the popular move_base package in ROS. Unlike robots, humans navigate not only through spatial awareness and physical distances but also by integrating external information, such as elevator maintenance updates from public notification boards and experiential knowledge, like the need for special access through certain doors. With the development of Large Language Models (LLMs), which possesses text understanding and intelligence close to human performance, there is now an opportunity to infuse robot navigation systems with a level of understanding akin to human cognition. In this study, we propose using osmAG (Area Graph in OpensStreetMap textual format), an innovative semantic topometric hierarchical map representation, to bridge the gap between the capabilities of ROS move_base and the contextual understanding offered by LLMs. Our methodology employs LLMs as an actual copilot in robot navigation, enabling the integration of a broader range of informational inputs while maintaining the robustness of traditional robotic navigation systems. Our code, demo, map, experiment results can be accessed at https://github.com/xiexiexiaoxiexie/Intelligent-LiDAR-Navigation-LLM-as-Copilot.
♻ ☆ Physics-Aware Combinatorial Assembly Sequence Planning using Data-free Action Masking
Combinatorial assembly uses standardized unit primitives to build objects that satisfy user specifications. This paper studies assembly sequence planning (ASP) for physical combinatorial assembly. Given the shape of the desired object, the goal is to find a sequence of actions for placing unit primitives to build the target object. In particular, we aim to ensure the planned assembly sequence is physically executable. However, ASP for combinatorial assembly is particularly challenging due to its combinatorial nature. To address the challenge, we employ deep reinforcement learning to learn a construction policy for placing unit primitives sequentially to build the desired object. Specifically, we design an online physics-aware action mask that filters out invalid actions, which effectively guides policy learning and ensures violation-free deployment. In the end, we apply the proposed method to Lego assembly with more than 250 3D structures. The experiment results demonstrate that the proposed method plans physically valid assembly sequences to build all structures, achieving a $100\%$ success rate, whereas the best comparable baseline fails more than $40$ structures. Our implementation is available at \url{https://github.com/intelligent-control-lab/PhysicsAwareCombinatorialASP}.
♻ ☆ Diffusion Transformer Policy
Recent large vision-language-action models pretrained on diverse robot datasets have demonstrated the potential for generalizing to new environments with a few in-domain data. However, those approaches usually predict individual discretized or continuous action by a small action head, which limits the ability in handling diverse action spaces. In contrast, we model the continuous action sequence with a large multi-modal diffusion transformer, dubbed as Diffusion Transformer Policy, in which we directly denoise action chunks by a large transformer model rather than a small action head for action embedding. By leveraging the scaling capability of transformers, the proposed approach can effectively model continuous end-effector actions across large diverse robot datasets, and achieve better generalization performance. Extensive experiments demonstrate the effectiveness and generalization of Diffusion Transformer Policy on Maniskill2, Libero, Calvin and SimplerEnv, as well as the real-world Franka arm, achieving consistent better performance on Real-to-Sim benchmark SimplerEnv, real-world Franka Arm and Libero compared to OpenVLA and Octo. Specifically, without bells and whistles, the proposed approach achieves state-of-the-art performance with only a single third-view camera stream in the Calvin task ABC->D, improving the average number of tasks completed in a row of 5 to 3.6, and the pretraining stage significantly facilitates the success sequence length on the Calvin by over 1.2.
comment: preprint; New Project Page: https://robodita.github.io; revert unsuitable replacement
♻ ☆ Intent Prediction-Driven Model Predictive Control for UAV Planning and Navigation in Dynamic Environments
Aerial robots can enhance construction site productivity by autonomously handling inspection and mapping tasks. However, ensuring safe navigation near human workers remains challenging. While navigation in static environments has been well studied, navigating dynamic environments remains open due to challenges in perception and planning. Payload limitations restrict the robots to using cameras with limited fields of view, resulting in unreliable perception and tracking during collision avoidance. Moreover, the rapidly changing conditions of dynamic environments can quickly make the generated optimal trajectory outdated.To address these challenges, this paper presents a comprehensive navigation framework that integrates perception, intent prediction, and planning. Our perception module detects and tracks dynamic obstacles efficiently and handles tracking loss and occlusion during collision avoidance. The proposed intent prediction module employs a Markov Decision Process (MDP) to forecast potential actions of dynamic obstacles with the possible future trajectories. Finally, a novel intent-based planning algorithm, leveraging model predictive control (MPC), is applied to generate navigation trajectories. Simulation and physical experiments demonstrate that our method improves the safety of navigation by achieving the fewest collisions compared to benchmarks.
comment: 8 pages, 7 figures, 2 tables, experiment video: https://youtu.be/4xsEeMB9WPY, GitHub: https://github.com/Zhefan-Xu/Intent-MPC
♻ ☆ DexDiffuser: Interaction-aware Diffusion Planning for Adaptive Dexterous Manipulation
Dexterous manipulation with contact-rich interactions is crucial for advanced robotics. While recent diffusion-based planning approaches show promise for simple manipulation tasks, they often produce unrealistic ghost states (e.g., the object automatically moves without hand contact) or lack adaptability when handling complex sequential interactions. In this work, we introduce DexDiffuser, an interaction-aware diffusion planning framework for adaptive dexterous manipulation. DexDiffuser models joint state-action dynamics through a dual-phase diffusion process which consists of pre-interaction contact alignment and post-contact goal-directed control, enabling goal-adaptive generalizable dexterous manipulation. Additionally, we incorporate dynamics model-based dual guidance and leverage large language models for automated guidance function generation, enhancing generalizability for physical interactions and facilitating diverse goal adaptation through language cues. Experiments on physical interaction tasks such as door opening, pen and block re-orientation, object relocation, and hammer striking demonstrate DexDiffuser's effectiveness on goals outside training distributions, achieving over twice the average success rate (59.2% vs. 29.5%) compared to existing methods. Our framework achieves an average of 70.7% success rate on goal adaptive dexterous tasks, highlighting its robustness and flexibility in contact-rich manipulation.
comment: Camera ready version. 27 pages. Project page: https://dexdiffuser.github.io/
Artificial Intelligence 93
☆ ShED-HD: A Shannon Entropy Distribution Framework for Lightweight Hallucination Detection on Edge Devices
Large Language Models (LLMs) have demonstrated impressive capabilities on a broad array of NLP tasks, but their tendency to produce hallucinations$\unicode{x2013}$plausible-sounding but factually incorrect content$\unicode{x2013}$poses severe challenges in high-stakes domains. Existing hallucination detection methods either bear the computational cost of multiple inference passes or sacrifice accuracy for efficiency with single-pass approaches, neither of which is ideal in resource-constrained environments such as edge devices. We propose the Shannon Entropy Distribution Hallucination Detector (ShED-HD), a novel hallucination detection framework that bridges this gap by classifying sequence-level entropy patterns using a lightweight BiLSTM architecture with single-headed attention. In contrast to prior approaches, ShED-HD efficiently detects distinctive uncertainty patterns across entire output sequences, preserving contextual awareness. Through in-depth evaluation on three datasets (BioASQ, TriviaQA, and Jeopardy Questions), we show that ShED-HD significantly outperforms other computationally efficient approaches in the out-of-distribution setting, while achieving comparable performance in the in-distribution setting. ShED-HD facilitates hallucination detection that is low-cost, accurate, and generalizable, improving the credibility of content generated by LLMs in resource-constrained environments where trustworthy AI functionality is crucial.
☆ Collaborating with AI Agents: Field Experiments on Teamwork, Productivity, and Performance
To uncover how AI agents change productivity, performance, and work processes, we introduce MindMeld: an experimentation platform enabling humans and AI agents to collaborate in integrative workspaces. In a large-scale marketing experiment on the platform, 2310 participants were randomly assigned to human-human and human-AI teams, with randomized AI personality traits. The teams exchanged 183,691 messages, and created 63,656 image edits, 1,960,095 ad copy edits, and 10,375 AI-generated images while producing 11,138 ads for a large think tank. Analysis of fine-grained communication, collaboration, and workflow logs revealed that collaborating with AI agents increased communication by 137% and allowed humans to focus 23% more on text and image content generation messaging and 20% less on direct text editing. Humans on Human-AI teams sent 23% fewer social messages, creating 60% greater productivity per worker and higher-quality ad copy. In contrast, human-human teams produced higher-quality images, suggesting that AI agents require fine-tuning for multimodal workflows. AI personality prompt randomization revealed that AI traits can complement human personalities to enhance collaboration. For example, conscientious humans paired with open AI agents improved image quality, while extroverted humans paired with conscientious AI agents reduced the quality of text, images, and clicks. In field tests of ad campaigns with ~5M impressions, ads with higher image quality produced by human collaborations and higher text quality produced by AI collaborations performed significantly better on click-through rate and cost per click metrics. Overall, ads created by human-AI teams performed similarly to those created by human-human teams. Together, these results suggest AI agents can improve teamwork and productivity, especially when tuned to complement human traits.
comment: 56 pages, 8 figures
☆ Adaptive Multi-Fidelity Reinforcement Learning for Variance Reduction in Engineering Design Optimization
Multi-fidelity Reinforcement Learning (RL) frameworks efficiently utilize computational resources by integrating analysis models of varying accuracy and costs. The prevailing methodologies, characterized by transfer learning, human-inspired strategies, control variate techniques, and adaptive sampling, predominantly depend on a structured hierarchy of models. However, this reliance on a model hierarchy can exacerbate variance in policy learning when the underlying models exhibit heterogeneous error distributions across the design space. To address this challenge, this work proposes a novel adaptive multi-fidelity RL framework, in which multiple heterogeneous, non-hierarchical low-fidelity models are dynamically leveraged alongside a high-fidelity model to efficiently learn a high-fidelity policy. Specifically, low-fidelity policies and their experience data are adaptively used for efficient targeted learning, guided by their alignment with the high-fidelity policy. The effectiveness of the approach is demonstrated in an octocopter design optimization problem, utilizing two low-fidelity models alongside a high-fidelity simulator. The results demonstrate that the proposed approach substantially reduces variance in policy learning, leading to improved convergence and consistent high-quality solutions relative to traditional hierarchical multi-fidelity RL methods. Moreover, the framework eliminates the need for manually tuning model usage schedules, which can otherwise introduce significant computational overhead. This positions the framework as an effective variance-reduction strategy for multi-fidelity RL, while also mitigating the computational and operational burden of manual fidelity scheduling.
☆ PG-SAM: Prior-Guided SAM with Medical for Multi-organ Segmentation
Segment Anything Model (SAM) demonstrates powerful zero-shot capabilities; however, its accuracy and robustness significantly decrease when applied to medical image segmentation. Existing methods address this issue through modality fusion, integrating textual and image information to provide more detailed priors. In this study, we argue that the granularity of text and the domain gap affect the accuracy of the priors. Furthermore, the discrepancy between high-level abstract semantics and pixel-level boundary details in images can introduce noise into the fusion process. To address this, we propose Prior-Guided SAM (PG-SAM), which employs a fine-grained modality prior aligner to leverage specialized medical knowledge for better modality alignment. The core of our method lies in efficiently addressing the domain gap with fine-grained text from a medical LLM. Meanwhile, it also enhances the priors' quality after modality alignment, ensuring more accurate segmentation. In addition, our decoder enhances the model's expressive capabilities through multi-level feature fusion and iterative mask optimizer operations, supporting unprompted learning. We also propose a unified pipeline that effectively supplies high-quality semantic information to SAM. Extensive experiments on the Synapse dataset demonstrate that the proposed PG-SAM achieves state-of-the-art performance. Our anonymous code is released at https://github.com/logan-0623/PG-SAM.
☆ A Study on Neuro-Symbolic Artificial Intelligence: Healthcare Perspectives
Over the last few decades, Artificial Intelligence (AI) scientists have been conducting investigations to attain human-level performance by a machine in accomplishing a cognitive task. Within machine learning, the ultimate aspiration is to attain Artificial General Intelligence (AGI) through a machine. This pursuit has led to the exploration of two distinct AI paradigms. Symbolic AI, also known as classical or GOFAI (Good Old-Fashioned AI) and Connectionist (Sub-symbolic) AI, represented by Neural Systems, are two mutually exclusive paradigms. Symbolic AI excels in reasoning, explainability, and knowledge representation but faces challenges in processing complex real-world data with noise. Conversely, deep learning (Black-Box systems) research breakthroughs in neural networks are notable, yet they lack reasoning and interpretability. Neuro-symbolic AI (NeSy), an emerging area of AI research, attempts to bridge this gap by integrating logical reasoning into neural networks, enabling them to learn and reason with symbolic representations. While a long path, this strategy has made significant progress towards achieving common sense reasoning by systems. This article conducts an extensive review of over 977 studies from prominent scientific databases (DBLP, ACL, IEEExplore, Scopus, PubMed, ICML, ICLR), thoroughly examining the multifaceted capabilities of Neuro-Symbolic AI, with a particular focus on its healthcare applications, particularly in drug discovery, and Protein engineering research. The survey addresses vital themes, including reasoning, explainability, integration strategies, 41 healthcare-related use cases, benchmarking, datasets, current approach limitations from both healthcare and broader perspectives, and proposed novel approaches for future experiments.
comment: 18 pages
ViVa: Video-Trained Value Functions for Guiding Online RL from Diverse Data
Online reinforcement learning (RL) with sparse rewards poses a challenge partly because of the lack of feedback on states leading to the goal. Furthermore, expert offline data with reward signal is rarely available to provide this feedback and bootstrap online learning. How can we guide online agents to the right solution without this on-task data? Reward shaping offers a solution by providing fine-grained signal to nudge the policy towards the optimal solution. However, reward shaping often requires domain knowledge to hand-engineer heuristics for a specific goal. To enable more general and inexpensive guidance, we propose and analyze a data-driven methodology that automatically guides RL by learning from widely available video data such as Internet recordings, off-task demonstrations, task failures, and undirected environment interaction. By learning a model of optimal goal-conditioned value from diverse passive data, we open the floor to scaling up and using various data sources to model general goal-reaching behaviors relevant to guiding online RL. Specifically, we use intent-conditioned value functions to learn from diverse videos and incorporate these goal-conditioned values into the reward. Our experiments show that video-trained value functions work well with a variety of data sources, exhibit positive transfer from human video pre-training, can generalize to unseen goals, and scale with dataset size.
☆ FROG: Fair Removal on Graphs
As compliance with privacy regulations becomes increasingly critical, the growing demand for data privacy has highlighted the significance of machine unlearning in many real world applications, such as social network and recommender systems, many of which can be represented as graph-structured data. However, existing graph unlearning algorithms indiscriminately modify edges or nodes from well-trained models without considering the potential impact of such structural modifications on fairness. For example, forgetting links between nodes with different genders in a social network may exacerbate group disparities, leading to significant fairness concerns. To address these challenges, we propose a novel approach that jointly optimizes the graph structure and the corresponding model for fair unlearning tasks. Specifically,our approach rewires the graph to enhance unlearning efficiency by removing redundant edges that hinder forgetting while preserving fairness through targeted edge augmentation. Additionally, we introduce a worst-case evaluation mechanism to assess the reliability of fair unlearning performance. Extensive experiments on real-world datasets demonstrate the effectiveness of the proposed approach in achieving superior unlearning outcomes.
☆ Exploring Energy Landscapes for Minimal Counterfactual Explanations: Applications in Cybersecurity and Beyond
Counterfactual explanations have emerged as a prominent method in Explainable Artificial Intelligence (XAI), providing intuitive and actionable insights into Machine Learning model decisions. In contrast to other traditional feature attribution methods that assess the importance of input variables, counterfactual explanations focus on identifying the minimal changes required to alter a model's prediction, offering a ``what-if'' analysis that is close to human reasoning. In the context of XAI, counterfactuals enhance transparency, trustworthiness and fairness, offering explanations that are not just interpretable but directly applicable in the decision-making processes. In this paper, we present a novel framework that integrates perturbation theory and statistical mechanics to generate minimal counterfactual explanations in explainable AI. We employ a local Taylor expansion of a Machine Learning model's predictive function and reformulate the counterfactual search as an energy minimization problem over a complex landscape. In sequence, we model the probability of candidate perturbations leveraging the Boltzmann distribution and use simulated annealing for iterative refinement. Our approach systematically identifies the smallest modifications required to change a model's prediction while maintaining plausibility. Experimental results on benchmark datasets for cybersecurity in Internet of Things environments, demonstrate that our method provides actionable, interpretable counterfactuals and offers deeper insights into model sensitivity and decision boundaries in high-dimensional spaces.
☆ Adaptive Physics-informed Neural Networks: A Survey
Physics-informed neural networks (PINNs) have emerged as a promising approach to solving partial differential equations (PDEs) using neural networks, particularly in data-scarce scenarios, due to their unsupervised training capability. However, limitations related to convergence and the need for re-optimization with each change in PDE parameters hinder their widespread adoption across scientific and engineering applications. This survey reviews existing research that addresses these limitations through transfer learning and meta-learning. The covered methods improve the training efficiency, allowing faster adaptation to new PDEs with fewer data and computational resources. While traditional numerical methods solve systems of differential equations directly, neural networks learn solutions implicitly by adjusting their parameters. One notable advantage of neural networks is their ability to abstract away from specific problem domains, allowing them to retain, discard, or adapt learned representations to efficiently address similar problems. By exploring the application of these techniques to PINNs, this survey identifies promising directions for future research to facilitate the broader adoption of PINNs in a wide range of scientific and engineering applications.
comment: https://openreview.net/forum?id=vz5P1Kbt6t
☆ Unmasking Deceptive Visuals: Benchmarking Multimodal Large Language Models on Misleading Chart Question Answering
Misleading chart visualizations, which intentionally manipulate data representations to support specific claims, can distort perceptions and lead to incorrect conclusions. Despite decades of research, misleading visualizations remain a widespread and pressing issue. Recent advances in multimodal large language models (MLLMs) have demonstrated strong chart comprehension capabilities, yet no existing work has systematically evaluated their ability to detect and interpret misleading charts. This paper introduces the Misleading Chart Question Answering (Misleading ChartQA) Benchmark, a large-scale multimodal dataset designed to assess MLLMs in identifying and reasoning about misleading charts. It contains over 3,000 curated examples, covering 21 types of misleaders and 10 chart types. Each example includes standardized chart code, CSV data, and multiple-choice questions with labeled explanations, validated through multi-round MLLM checks and exhausted expert human review. We benchmark 16 state-of-the-art MLLMs on our dataset, revealing their limitations in identifying visually deceptive practices. We also propose a novel pipeline that detects and localizes misleaders, enhancing MLLMs' accuracy in misleading chart interpretation. Our work establishes a foundation for advancing MLLM-driven misleading chart comprehension. We publicly release the sample dataset to support further research in this critical area.
comment: 31 pages in total. Under Review For ARR
☆ Self-Attention Diffusion Models for Zero-Shot Biomedical Image Segmentation: Unlocking New Frontiers in Medical Imaging
Producing high-quality segmentation masks for medical images is a fundamental challenge in biomedical image analysis. Recent research has explored large-scale supervised training to enable segmentation across various medical imaging modalities and unsupervised training to facilitate segmentation without dense annotations. However, constructing a model capable of segmenting diverse medical images in a zero-shot manner without any annotations remains a significant hurdle. This paper introduces the Attention Diffusion Zero-shot Unsupervised System (ADZUS), a novel approach that leverages self-attention diffusion models for zero-shot biomedical image segmentation. ADZUS harnesses the intrinsic capabilities of pre-trained diffusion models, utilizing their generative and discriminative potentials to segment medical images without requiring annotated training data or prior domain-specific knowledge. The ADZUS architecture is detailed, with its integration of self-attention mechanisms that facilitate context-aware and detail-sensitive segmentations being highlighted. Experimental results across various medical imaging datasets, including skin lesion segmentation, chest X-ray infection segmentation, and white blood cell segmentation, reveal that ADZUS achieves state-of-the-art performance. Notably, ADZUS reached Dice scores ranging from 88.7\% to 92.9\% and IoU scores from 66.3\% to 93.3\% across different segmentation tasks, demonstrating significant improvements in handling novel, unseen medical imagery. It is noteworthy that while ADZUS demonstrates high effectiveness, it demands substantial computational resources and extended processing times. The model's efficacy in zero-shot settings underscores its potential to reduce reliance on costly annotations and seamlessly adapt to new medical imaging tasks, thereby expanding the diagnostic capabilities of AI-driven medical imaging technologies.
comment: 15 pages, 5 figures
☆ Strategic Prompt Pricing for AIGC Services: A User-Centric Approach
The rapid growth of AI-generated content (AIGC) services has created an urgent need for effective prompt pricing strategies, yet current approaches overlook users' strategic two-step decision-making process in selecting and utilizing generative AI models. This oversight creates two key technical challenges: quantifying the relationship between user prompt capabilities and generation outcomes, and optimizing platform payoff while accounting for heterogeneous user behaviors. We address these challenges by introducing prompt ambiguity, a theoretical framework that captures users' varying abilities in prompt engineering, and developing an Optimal Prompt Pricing (OPP) algorithm. Our analysis reveals a counterintuitive insight: users with higher prompt ambiguity (i.e., lower capability) exhibit non-monotonic prompt usage patterns, first increasing then decreasing with ambiguity levels, reflecting complex changes in marginal utility. Experimental evaluation using a character-level GPT-like model demonstrates that our OPP algorithm achieves up to 31.72% improvement in platform payoff compared to existing pricing mechanisms, validating the importance of user-centric prompt pricing in AIGC services.
comment: accepted in WiOpt 2025
☆ Evaluating Negative Sampling Approaches for Neural Topic Models
Negative sampling has emerged as an effective technique that enables deep learning models to learn better representations by introducing the paradigm of learn-to-compare. The goal of this approach is to add robustness to deep learning models to learn better representation by comparing the positive samples against the negative ones. Despite its numerous demonstrations in various areas of computer vision and natural language processing, a comprehensive study of the effect of negative sampling in an unsupervised domain like topic modeling has not been well explored. In this paper, we present a comprehensive analysis of the impact of different negative sampling strategies on neural topic models. We compare the performance of several popular neural topic models by incorporating a negative sampling technique in the decoder of variational autoencoder-based neural topic models. Experiments on four publicly available datasets demonstrate that integrating negative sampling into topic models results in significant enhancements across multiple aspects, including improved topic coherence, richer topic diversity, and more accurate document classification. Manual evaluations also indicate that the inclusion of negative sampling into neural topic models enhances the quality of the generated topics. These findings highlight the potential of negative sampling as a valuable tool for advancing the effectiveness of neural topic models.
comment: Code is available at: https://github.com/AdhyaSuman/Eval_NegTM
☆ SNRAware: Improved Deep Learning MRI Denoising with SNR Unit Training and G-factor Map Augmentation
To develop and evaluate a new deep learning MR denoising method that leverages quantitative noise distribution information from the reconstruction process to improve denoising performance and generalization. This retrospective study trained 14 different transformer and convolutional models with two backbone architectures on a large dataset of 2,885,236 images from 96,605 cardiac retro-gated cine complex series acquired at 3T. The proposed training scheme, termed SNRAware, leverages knowledge of the MRI reconstruction process to improve denoising performance by simulating large, high quality, and diverse synthetic datasets, and providing quantitative information about the noise distribution to the model. In-distribution testing was performed on a hold-out dataset of 3000 samples with performance measured using PSNR and SSIM, with ablation comparison without the noise augmentation. Out-of-distribution tests were conducted on cardiac real-time cine, first-pass cardiac perfusion, and neuro and spine MRI, all acquired at 1.5T, to test model generalization across imaging sequences, dynamically changing contrast, different anatomies, and field strengths. The best model found in the in-distribution test generalized well to out-of-distribution samples, delivering 6.5x and 2.9x CNR improvement for real-time cine and perfusion imaging, respectively. Further, a model trained with 100% cardiac cine data generalized well to a T1 MPRAGE neuro 3D scan and T2 TSE spine MRI.
☆ Active Inference for Energy Control and Planning in Smart Buildings and Communities
Active Inference (AIF) is emerging as a powerful framework for decision-making under uncertainty, yet its potential in engineering applications remains largely unexplored. In this work, we propose a novel dual-layer AIF architecture that addresses both building-level and community-level energy management. By leveraging the free energy principle, each layer adapts to evolving conditions and handles partial observability without extensive sensor information and respecting data privacy. We validate the continuous AIF model against both a perfect optimization baseline and a reinforcement learning-based approach. We also test the community AIF framework under extreme pricing scenarios. The results highlight the model's robustness in handling abrupt changes. This study is the first to show how a distributed AIF works in engineering. It also highlights new opportunities for privacy-preserving and uncertainty-aware control strategies in engineering applications.
comment: Submitted to IEEE CASE 2025 (IEEE 21st International Conference on Automation Science and Engineering)
DiffusionTalker: Efficient and Compact Speech-Driven 3D Talking Head via Personalizer-Guided Distillation ICME2025
Real-time speech-driven 3D facial animation has been attractive in academia and industry. Traditional methods mainly focus on learning a deterministic mapping from speech to animation. Recent approaches start to consider the nondeterministic fact of speech-driven 3D face animation and employ the diffusion model for the task. Existing diffusion-based methods can improve the diversity of facial animation. However, personalized speaking styles conveying accurate lip language is still lacking, besides, efficiency and compactness still need to be improved. In this work, we propose DiffusionTalker to address the above limitations via personalizer-guided distillation. In terms of personalization, we introduce a contrastive personalizer that learns identity and emotion embeddings to capture speaking styles from audio. We further propose a personalizer enhancer during distillation to enhance the influence of embeddings on facial animation. For efficiency, we use iterative distillation to reduce the steps required for animation generation and achieve more than 8x speedup in inference. To achieve compactness, we distill the large teacher model into a smaller student model, reducing our model's storage by 86.4\% while minimizing performance loss. After distillation, users can derive their identity and emotion embeddings from audio to quickly create personalized animations that reflect specific speaking styles. Extensive experiments are conducted to demonstrate that our method outperforms state-of-the-art methods. The code will be released at: https://github.com/ChenVoid/DiffusionTalker.
comment: Accepted by ICME2025
☆ Adoption of Watermarking for Generative AI Systems in Practice and Implications under the new EU AI Act
AI-generated images have become so good in recent years that individuals cannot distinguish them any more from "real" images. This development creates a series of societal risks, and challenges our perception of what is true and what is not, particularly with the emergence of "deep fakes" that impersonate real individuals. Watermarking, a technique that involves embedding identifying information within images to indicate their AI-generated nature, has emerged as a primary mechanism to address the risks posed by AI-generated images. The implementation of watermarking techniques is now becoming a legal requirement in many jurisdictions, including under the new 2024 EU AI Act. Despite the widespread use of AI image generation systems, the current status of watermarking implementation remains largely unexamined. Moreover, the practical implications of the AI Act's watermarking requirements have not previously been studied. The present paper therefore both provides an empirical analysis of 50 of the most widely used AI systems for image generation, and embeds this empirical analysis into a legal analysis of the AI Act. We identify four categories of generative AI image systems relevant under the AI Act, outline the legal obligations for each category, and find that only a minority number of providers currently implement adequate watermarking practices.
comment: 12 pages, 7 figures, note that this work has not been published in a peer reviewed venue yet. While we have made our best effort to ensure the validity of our findings, it is therefore still work in progress and potentially subject to change
☆ Efficient Deep Learning Approaches for Processing Ultra-Widefield Retinal Imaging
Deep learning has emerged as the predominant solution for classifying medical images. We intend to apply these developments to the ultra-widefield (UWF) retinal imaging dataset. Since UWF images can accurately diagnose various retina diseases, it is very important to clas sify them accurately and prevent them with early treatment. However, processing images manually is time-consuming and labor-intensive, and there are two challenges to automating this process. First, high perfor mance usually requires high computational resources. Artificial intelli gence medical technology is better suited for places with limited medical resources, but using high-performance processing units in such environ ments is challenging. Second, the problem of the accuracy of colour fun dus photography (CFP) methods. In general, the UWF method provides more information for retinal diagnosis than the CFP method, but most of the research has been conducted based on the CFP method. Thus, we demonstrate that these problems can be efficiently addressed in low performance units using methods such as strategic data augmentation and model ensembles, which balance performance and computational re sources while utilizing UWF images.
☆ LocDiffusion: Identifying Locations on Earth by Diffusing in the Hilbert Space
Image geolocalization is a fundamental yet challenging task, aiming at inferring the geolocation on Earth where an image is taken. Existing methods approach it either via grid-based classification or via image retrieval. Their performance significantly suffers when the spatial distribution of test images does not align with such choices. To address these limitations, we propose to leverage diffusion as a mechanism for image geolocalization. To avoid the problematic manifold reprojection step in diffusion, we developed a novel spherical positional encoding-decoding framework, which encodes points on a spherical surface (e.g., geolocations on Earth) into a Hilbert space of Spherical Harmonics coefficients and decodes points (geolocations) by mode-seeking. We call this type of position encoding Spherical Harmonics Dirac Delta (SHDD) Representation. We also propose a novel SirenNet-based architecture called CS-UNet to learn the conditional backward process in the latent SHDD space by minimizing a latent KL-divergence loss. We train a conditional latent diffusion model called LocDiffusion that generates geolocations under the guidance of images -- to the best of our knowledge, the first generative model for image geolocalization by diffusing geolocation information in a hidden location embedding space. We evaluate our method against SOTA image geolocalization baselines. LocDiffusion achieves competitive geolocalization performance and demonstrates significantly stronger generalizability to unseen geolocations.
☆ Mitigating Reward Over-Optimization in RLHF via Behavior-Supported Regularization ICLR 2025
Reinforcement learning from human feedback (RLHF) is an effective method for aligning large language models (LLMs) with human values. However, reward over-optimization remains an open challenge leading to discrepancies between the performance of LLMs under the reward model and the true human objectives. A primary contributor to reward over-optimization is the extrapolation error that arises when the reward model evaluates out-of-distribution (OOD) responses. However, current methods still fail to prevent the increasing frequency of OOD response generation during the reinforcement learning (RL) process and are not effective at handling extrapolation errors from OOD responses. In this work, we propose the Behavior-Supported Policy Optimization (BSPO) method to mitigate the reward over-optimization issue. Specifically, we define behavior policy as the next token distribution of the reward training dataset to model the in-distribution (ID) region of the reward model. Building on this, we introduce the behavior-supported Bellman operator to regularize the value function, penalizing all OOD values without impacting the ID ones. Consequently, BSPO reduces the generation of OOD responses during the RL process, thereby avoiding overestimation caused by the reward model's extrapolation errors. Theoretically, we prove that BSPO guarantees a monotonic improvement of the supported policy until convergence to the optimal behavior-supported policy. Empirical results from extensive experiments show that BSPO outperforms baselines in preventing reward over-optimization due to OOD evaluation and finding the optimal ID policy.
comment: Published as a conference paper at ICLR 2025
☆ GeoBenchX: Benchmarking LLMs for Multistep Geospatial Tasks
In this paper, we establish a benchmark for evaluating large language models (LLMs) on multi-step geospatial tasks relevant to commercial GIS practitioners. We assess seven leading commercial LLMs (Sonnet 3.5 and 3.7, Haiku 3.5, Gemini 2.0, GPT-4o, GPT-4o mini, and o3-mini) using a simple tool-calling agent equipped with 23 geospatial functions. Our benchmark comprises tasks across four categories of increasing complexity, with both solvable and intentionally unsolvable tasks to test hallucination rejection. We develop an LLM-as-Judge evaluation framework to compare agent solutions against reference implementations. Results show Sonnet 3.5 and GPT-4o achieve the best overall performance, with Claude models excelling on solvable tasks while OpenAI models better identify unsolvable scenarios. We observe significant differences in token usage, with Anthropic models consuming substantially more tokens than competitors. Common errors include misunderstanding geometrical relationships, relying on outdated knowledge, and inefficient data manipulation. The resulting benchmark set, evaluation framework, and data generation pipeline are released as open-source resources, providing one more standardized method for ongoing evaluation of LLMs for GeoAI.
comment: Github with code and benchmark set: https://github.com/Solirinai/GeoBenchX
☆ AgentRxiv: Towards Collaborative Autonomous Research
Progress in scientific discovery is rarely the result of a single "Eureka" moment, but is rather the product of hundreds of scientists incrementally working together toward a common goal. While existing agent workflows are capable of producing research autonomously, they do so in isolation, without the ability to continuously improve upon prior research results. To address these challenges, we introduce AgentRxiv-a framework that lets LLM agent laboratories upload and retrieve reports from a shared preprint server in order to collaborate, share insights, and iteratively build on each other's research. We task agent laboratories to develop new reasoning and prompting techniques and find that agents with access to their prior research achieve higher performance improvements compared to agents operating in isolation (11.4% relative improvement over baseline on MATH-500). We find that the best performing strategy generalizes to benchmarks in other domains (improving on average by 3.3%). Multiple agent laboratories sharing research through AgentRxiv are able to work together towards a common goal, progressing more rapidly than isolated laboratories, achieving higher overall accuracy (13.7% relative improvement over baseline on MATH-500). These findings suggest that autonomous agents may play a role in designing future AI systems alongside humans. We hope that AgentRxiv allows agents to collaborate toward research goals and enables researchers to accelerate discovery.
☆ Temporal Relation Extraction in Clinical Texts: A Span-based Graph Transformer Approach
Temporal information extraction from unstructured text is essential for contextualizing events and deriving actionable insights, particularly in the medical domain. We address the task of extracting clinical events and their temporal relations using the well-studied I2B2 2012 Temporal Relations Challenge corpus. This task is inherently challenging due to complex clinical language, long documents, and sparse annotations. We introduce GRAPHTREX, a novel method integrating span-based entity-relation extraction, clinical large pre-trained language models (LPLMs), and Heterogeneous Graph Transformers (HGT) to capture local and global dependencies. Our HGT component facilitates information propagation across the document through innovative global landmarks that bridge distant entities. Our method improves the state-of-the-art with 5.5% improvement in the tempeval $F_1$ score over the previous best and up to 8.9% improvement on long-range relations, which presents a formidable challenge. This work not only advances temporal information extraction but also lays the groundwork for improved diagnostic and prognostic models through enhanced temporal reasoning.
comment: Introducing a novel method for joint extraction of medical events and temporal relations from free-text, leveraging clinical LPLMs and Heterogeneous Graph Transformers, achieving a 5.5% improvement over the previous state-of-the-art and up to 8.9% on long-range relations
☆ On the effectiveness of LLMs for automatic grading of open-ended questions in Spanish
Grading is a time-consuming and laborious task that educators must face. It is an important task since it provides feedback signals to learners, and it has been demonstrated that timely feedback improves the learning process. In recent years, the irruption of LLMs has shed light on the effectiveness of automatic grading. In this paper, we explore the performance of different LLMs and prompting techniques in automatically grading short-text answers to open-ended questions. Unlike most of the literature, our study focuses on a use case where the questions, answers, and prompts are all in Spanish. Experimental results comparing automatic scores to those of human-expert evaluators show good outcomes in terms of accuracy, precision and consistency for advanced LLMs, both open and proprietary. Results are notably sensitive to prompt styles, suggesting biases toward certain words or content in the prompt. However, the best combinations of models and prompt strategies, consistently surpasses an accuracy of 95% in a three-level grading task, which even rises up to more than 98% when the it is simplified to a binary right or wrong rating problem, which demonstrates the potential that LLMs have to implement this type of automation in education applications.
☆ Unseen from Seen: Rewriting Observation-Instruction Using Foundation Models for Augmenting Vision-Language Navigation
Data scarcity is a long-standing challenge in the Vision-Language Navigation (VLN) field, which extremely hinders the generalization of agents to unseen environments. Previous works primarily rely on additional simulator data or web-collected images/videos to improve the generalization. However, the simulator environments still face limited diversity, and the web-collected data often requires extensive labor to remove the noise. In this paper, we propose a Rewriting-driven AugMentation (RAM) paradigm for VLN, which directly creates the unseen observation-instruction pairs via rewriting human-annotated training data. Benefiting from our rewriting mechanism, new observation-instruction can be obtained in both simulator-free and labor-saving manners to promote generalization. Specifically, we first introduce Object-Enriched Observation Rewriting, where we combine Vision-Language Models (VLMs) and Large Language Models (LLMs) to derive rewritten object-enriched scene descriptions, enabling observation synthesis with diverse objects and spatial layouts via Text-to-Image Generation Models (T2IMs). Then, we propose Observation-Contrast Instruction Rewriting, which generates observation-aligned rewritten instructions by requiring LLMs to reason the difference between original and new observations. We further develop a mixing-then-focusing training strategy with a random observation cropping scheme, effectively enhancing data distribution diversity while suppressing augmentation data noise during training. Experiments on both the discrete environments (R2R, REVERIE, and R4R datasets) and continuous environments (R2R-CE dataset) show the superior performance and impressive generalization ability of our method. Code is available at https://github.com/SaDil13/VLN-RAM.
☆ Dynamic Task Vector Grouping for Efficient Multi-Task Prompt Tuning
Multi-task prompt tuning utilizes multiple high-resource source tasks to improve performance on low-source target tasks. Existing approaches transfer the soft prompt trained by combining all source tasks or a single ``high-similar'' source task one-time-only. However, we find that the optimal transfer performance often comes from a combination of source tasks, which is neither one nor all. Further, we find that the similarity between source and target tasks also changes dynamically during fine-tuning after transfering, making similarity calculation in the initiation stage inadequate. To address these issues, we propose a method called Dynamic Task Vector Grouping (DTVG), whose core ideas contain (1) measuring the task similarity with task vectors instead of soft prompt, (2) grouping the optimal source task combination based on two metrics: {\it target similarity} and {\it knowledge consistency}; (3) dynamically updating the combination in each iteration step. Extensive experiments on the 26 NLP datasets under different settings demonstrate that DTVG effectively groups similar source tasks while reducing negative transfer, achieving the start-of-art performance.
comment: Work in progress
☆ Decision from Suboptimal Classifiers: Excess Risk Pre- and Post-Calibration
Probabilistic classifiers are central for making informed decisions under uncertainty. Based on the maximum expected utility principle, optimal decision rules can be derived using the posterior class probabilities and misclassification costs. Yet, in practice only learned approximations of the oracle posterior probabilities are available. In this work, we quantify the excess risk (a.k.a. regret) incurred using approximate posterior probabilities in batch binary decision-making. We provide analytical expressions for miscalibration-induced regret ($R^{\mathrm{CL}}$), as well as tight and informative upper and lower bounds on the regret of calibrated classifiers ($R^{\mathrm{GL}}$). These expressions allow us to identify regimes where recalibration alone addresses most of the regret, and regimes where the regret is dominated by the grouping loss, which calls for post-training beyond recalibration. Crucially, both $R^{\mathrm{CL}}$ and $R^{\mathrm{GL}}$ can be estimated in practice using a calibration curve and a recent grouping loss estimator. On NLP experiments, we show that these quantities identify when the expected gain of more advanced post-training is worth the operational cost. Finally, we highlight the potential of multicalibration approaches as efficient alternatives to costlier fine-tuning approaches.
☆ Lost in Cultural Translation: Do LLMs Struggle with Math Across Cultural Contexts?
Large Language Models (LLMs) have significantly advanced various fields, particularly coding, mathematical reasoning, and logical problem solving. However, a critical question remains: Do these mathematical reasoning abilities persist when LLMs are presented with culturally adapted math problems? Specifically, how do LLMs perform when faced with math problems embedded in cultural contexts that have no significant representation in main stream web-scale AI training data? To explore this, we generated six synthetic cultural datasets from GSM8K, a widely used benchmark for assessing LLMs' mathematical reasoning skills. While preserving the mathematical logic and numerical values of the original GSM8K test set, we modify cultural elements such as personal names, food items, place names, etc. These culturally adapted datasets provide a more reliable framework for evaluating LLMs' mathematical reasoning under shifting cultural contexts. Our findings reveal that LLMs struggle with math problems when cultural references change, even though the underlying mathematical structure remains constant. Smaller models exhibit greater performance drops compared to larger models. Interestingly, our results also suggest that cultural familiarity can enhance mathematical reasoning. Even models with no explicit mathematical training but exposure to relevant cultural contexts sometimes outperform larger, mathematically proficient models on culturally embedded math problems. This study highlights the impact of cultural context on the mathematical reasoning abilities of LLMs, underscoring the need for more diverse and representative training data to improve robustness in real-world applications. The benchmark data sets and script for reproducing the results are available at https://github.com/akarim23131/Lost_in_Cultural_Translation
☆ Vision-R1: Evolving Human-Free Alignment in Large Vision-Language Models via Vision-Guided Reinforcement Learning
Large Vision-Language Models (LVLMs) typically follow a two-stage training paradigm-pretraining and supervised fine-tuning. Recently, preference optimization, derived from the language domain, has emerged as an effective post-training reinforcement strategy to enhance capabilities of LVLMs. However, constructing high-quality human-annotated preference data and developing robust reward models to mimic these preferences are both costly and challenging. Motivated by this observation, we propose Vision-R1, a novel vision-guided R1-like reinforcement learning algorithm for LVLMs that rewards models with definitive vision feedback. It only leverages curated instruction data, eliminating the need for specialized reward models and handcrafted preference datasets. We incorporate a criterion-driven reward function that further integrates multi-dimensional feedback to evaluate model completions comprehensively based on the vision task logic. Furthermore, we introduce a progressive rule refinement strategy that dynamically adjusts the reward criteria during training, enabling continuous model improvement and mitigating reward hacking. Extensive experiments on both in-distribution and out-of-distribution benchmarks demonstrate that fine-tuning the 7B LVLMs with Vision-R1 achieves consistent performance gains, with even up to 50% improvement and surpassing the state-of-the-art 10x size model.
comment: Project in development. Github: https://github.com/jefferyZhan/Griffon/tree/master/Vision-R1
☆ Instructing the Architecture Search for Spatial-temporal Sequence Forecasting with LLM
Spatial-temporal sequence forecasting (STSF) is a long-standing research problem with widespread real-world applications. Neural architecture search (NAS), which automates the neural network design, has been shown effective in tackling the STSF problem. However, the existing NAS methods for STSF focus on generating architectures in a time-consuming data-driven fashion, which heavily limits their ability to use background knowledge and explore the complicated search trajectory. Large language models (LLMs) have shown remarkable ability in decision-making with comprehensive internal world knowledge, but how it could benefit NAS for STSF remains unexplored. In this paper, we propose a novel NAS method for STSF based on LLM. Instead of directly generate architectures with LLM, We inspire the LLM's capability with a multi-level enhancement mechanism. Specifically, on the step-level, we decompose the generation task into decision steps with powerful prompt engineering and inspire LLM to serve as instructor for architecture search based on its internal knowledge. On the instance-level, we utilize a one-step tuning framework to quickly evaluate the architecture instance and a memory bank to cumulate knowledge to improve LLM's search ability. On the task-level, we propose a two-stage architecture search, balancing the exploration stage and optimization stage, to reduce the possibility of being trapped in local optima. Extensive experimental results demonstrate that our method can achieve competitive effectiveness with superior efficiency against existing NAS methods for STSF.
☆ Predicting Multitasking in Manual and Automated Driving with Optimal Supervisory Control
Modern driving involves interactive technologies that can divert attention, increasing the risk of accidents. This paper presents a computational cognitive model that simulates human multitasking while driving. Based on optimal supervisory control theory, the model predicts how multitasking adapts to variations in driving demands, interactive tasks, and automation levels. Unlike previous models, it accounts for context-dependent multitasking across different degrees of driving automation. The model predicts longer in-car glances on straight roads and shorter glances during curves. It also anticipates increased glance durations with driver aids such as lane-centering assistance and their interaction with environmental demands. Validated against two empirical datasets, the model offers insights into driver multitasking amid evolving in-car technologies and automation.
☆ Metaphor-based Jailbreaking Attacks on Text-to-Image Models
To mitigate misuse, text-to-image~(T2I) models commonly incorporate safety filters to prevent the generation of sensitive images. Unfortunately, recent jailbreaking attack methods use LLMs to generate adversarial prompts that effectively bypass safety filters while generating sensitive images, revealing the safety vulnerabilities within the T2I model. However, existing LLM-based attack methods lack explicit guidance, relying on substantial queries to achieve a successful attack, which limits their practicality in real-world scenarios. In this work, we introduce \textbf{MJA}, a \textbf{m}etaphor-based \textbf{j}ailbreaking \textbf{a}ttack method inspired by the Taboo game, aiming to balance the attack effectiveness and query efficiency by generating metaphor-based adversarial prompts. Specifically, MJA consists of two modules: an LLM-based multi-agent generation module~(MLAG) and an adversarial prompt optimization module~(APO). MLAG decomposes the generation of metaphor-based adversarial prompts into three subtasks: metaphor retrieval, context matching, and adversarial prompt generation. Subsequently, MLAG coordinates three LLM-based agents to generate diverse adversarial prompts by exploring various metaphors and contexts. To enhance the attack efficiency, APO first trains a surrogate model to predict the attack results of adversarial prompts and then designs an acquisition strategy to adaptively identify optimal adversarial prompts. Experiments demonstrate that MJA achieves better attack effectiveness while requiring fewer queries compared to baseline methods. Moreover, our adversarial prompts exhibit strong transferability across various open-source and commercial T2I models. \textcolor{red}{This paper includes model-generated content that may contain offensive or distressing material.}
comment: 13 page3, 4 figures. This paper includes model-generated content that may contain offensive or distressing material
☆ Optimizing Navigation And Chemical Application in Precision Agriculture With Deep Reinforcement Learning And Conditional Action Tree
This paper presents a novel reinforcement learning (RL)-based planning scheme for optimized robotic management of biotic stresses in precision agriculture. The framework employs a hierarchical decision-making structure with conditional action masking, where high-level actions direct the robot's exploration, while low-level actions optimize its navigation and efficient chemical spraying in affected areas. The key objectives of optimization include improving the coverage of infected areas with limited battery power and reducing chemical usage, thus preventing unnecessary spraying of healthy areas of the field. Our numerical experimental results demonstrate that the proposed method, Hierarchical Action Masking Proximal Policy Optimization (HAM-PPO), significantly outperforms baseline practices, such as LawnMower navigation + indiscriminate spraying (Carpet Spray), in terms of yield recovery and resource efficiency. HAM-PPO consistently achieves higher yield recovery percentages and lower chemical costs across a range of infection scenarios. The framework also exhibits robustness to observation noise and generalizability under diverse environmental conditions, adapting to varying infection ranges and spatial distribution patterns.
comment: 32 pages, 9 figures
☆ Taste More, Taste Better: Diverse Data and Strong Model Boost Semi-Supervised Crowd Counting CVPR 2025
Semi-supervised crowd counting is crucial for addressing the high annotation costs of densely populated scenes. Although several methods based on pseudo-labeling have been proposed, it remains challenging to effectively and accurately utilize unlabeled data. In this paper, we propose a novel framework called Taste More Taste Better (TMTB), which emphasizes both data and model aspects. Firstly, we explore a data augmentation technique well-suited for the crowd counting task. By inpainting the background regions, this technique can effectively enhance data diversity while preserving the fidelity of the entire scenes. Secondly, we introduce the Visual State Space Model as backbone to capture the global context information from crowd scenes, which is crucial for extremely crowded, low-light, and adverse weather scenarios. In addition to the traditional regression head for exact prediction, we employ an Anti-Noise classification head to provide less exact but more accurate supervision, since the regression head is sensitive to noise in manual annotations. We conduct extensive experiments on four benchmark datasets and show that our method outperforms state-of-the-art methods by a large margin. Code is publicly available on https://github.com/syhien/taste_more_taste_better.
comment: Accepted by CVPR 2025
☆ Co-SemDepth: Fast Joint Semantic Segmentation and Depth Estimation on Aerial Images
Understanding the geometric and semantic properties of the scene is crucial in autonomous navigation and particularly challenging in the case of Unmanned Aerial Vehicle (UAV) navigation. Such information may be by obtained by estimating depth and semantic segmentation maps of the surrounding environment and for their practical use in autonomous navigation, the procedure must be performed as close to real-time as possible. In this paper, we leverage monocular cameras on aerial robots to predict depth and semantic maps in low-altitude unstructured environments. We propose a joint deep-learning architecture that can perform the two tasks accurately and rapidly, and validate its effectiveness on MidAir and Aeroscapes benchmark datasets. Our joint-architecture proves to be competitive or superior to the other single and joint architecture methods while performing its task fast predicting 20.2 FPS on a single NVIDIA quadro p5000 GPU and it has a low memory footprint. All codes for training and prediction can be found on this link: https://github.com/Malga-Vision/Co-SemDepth
☆ Trade-offs in Large Reasoning Models: An Empirical Analysis of Deliberative and Adaptive Reasoning over Foundational Capabilities
Recent advancements in Large Reasoning Models (LRMs), such as OpenAI's o1/o3 and DeepSeek-R1, have demonstrated remarkable performance in specialized reasoning tasks through human-like deliberative thinking and long chain-of-thought reasoning. However, our systematic evaluation across various model families (DeepSeek, Qwen, and LLaMA) and scales (7B to 671B) reveals that acquiring these deliberative reasoning capabilities significantly reduces the foundational capabilities of LRMs, including notable declines in helpfulness and harmlessness, alongside substantially increased inference costs. Importantly, we demonstrate that adaptive reasoning -- employing modes like Zero-Thinking, Less-Thinking, and Summary-Thinking -- can effectively alleviate these drawbacks. Our empirical insights underline the critical need for developing more versatile LRMs capable of dynamically allocating inference-time compute according to specific task characteristics.
comment: 23 pages. Work in progress
☆ PIM: Physics-Informed Multi-task Pre-training for Improving Inertial Sensor-Based Human Activity Recognition
Human activity recognition (HAR) with deep learning models relies on large amounts of labeled data, often challenging to obtain due to associated cost, time, and labor. Self-supervised learning (SSL) has emerged as an effective approach to leverage unlabeled data through pretext tasks, such as masked reconstruction and multitask learning with signal processing-based data augmentations, to pre-train encoder models. However, such methods are often derived from computer vision approaches that disregard physical mechanisms and constraints that govern wearable sensor data and the phenomena they reflect. In this paper, we propose a physics-informed multi-task pre-training (PIM) framework for IMU-based HAR. PIM generates pre-text tasks based on the understanding of basic physical aspects of human motion: including movement speed, angles of movement, and symmetry between sensor placements. Given a sensor signal, we calculate corresponding features using physics-based equations and use them as pretext tasks for SSL. This enables the model to capture fundamental physical characteristics of human activities, which is especially relevant for multi-sensor systems. Experimental evaluations on four HAR benchmark datasets demonstrate that the proposed method outperforms existing state-of-the-art methods, including data augmentation and masked reconstruction, in terms of accuracy and F1 score. We have observed gains of almost 10\% in macro f1 score and accuracy with only 2 to 8 labeled examples per class and up to 3% when there is no reduction in the amount of training data.
☆ Shot Sequence Ordering for Video Editing: Benchmarks, Metrics, and Cinematology-Inspired Computing Methods
With the rising popularity of short video platforms, the demand for video production has increased substantially. However, high-quality video creation continues to rely heavily on professional editing skills and a nuanced understanding of visual language. To address this challenge, the Shot Sequence Ordering (SSO) task in AI-assisted video editing has emerged as a pivotal approach for enhancing video storytelling and the overall viewing experience. Nevertheless, the progress in this field has been impeded by a lack of publicly available benchmark datasets. In response, this paper introduces two novel benchmark datasets, AVE-Order and ActivityNet-Order. Additionally, we employ the Kendall Tau distance as an evaluation metric for the SSO task and propose the Kendall Tau Distance-Cross Entropy Loss. We further introduce the concept of Cinematology Embedding, which incorporates movie metadata and shot labels as prior knowledge into the SSO model, and constructs the AVE-Meta dataset to validate the method's effectiveness. Experimental results indicate that the proposed loss function and method substantially enhance SSO task accuracy. All datasets are publicly accessible at https://github.com/litchiar/ShotSeqBench.
☆ PhysTwin: Physics-Informed Reconstruction and Simulation of Deformable Objects from Videos
Creating a physical digital twin of a real-world object has immense potential in robotics, content creation, and XR. In this paper, we present PhysTwin, a novel framework that uses sparse videos of dynamic objects under interaction to produce a photo- and physically realistic, real-time interactive virtual replica. Our approach centers on two key components: (1) a physics-informed representation that combines spring-mass models for realistic physical simulation, generative shape models for geometry, and Gaussian splats for rendering; and (2) a novel multi-stage, optimization-based inverse modeling framework that reconstructs complete geometry, infers dense physical properties, and replicates realistic appearance from videos. Our method integrates an inverse physics framework with visual perception cues, enabling high-fidelity reconstruction even from partial, occluded, and limited viewpoints. PhysTwin supports modeling various deformable objects, including ropes, stuffed animals, cloth, and delivery packages. Experiments show that PhysTwin outperforms competing methods in reconstruction, rendering, future prediction, and simulation under novel interactions. We further demonstrate its applications in interactive real-time simulation and model-based robotic motion planning.
comment: Project Page: https://jianghanxiao.github.io/phystwin-web/
☆ Understanding the Effects of RLHF on the Quality and Detectability of LLM-Generated Texts
Large Language Models (LLMs) have demonstrated exceptional performance on a range of downstream NLP tasks by generating text that closely resembles human writing. However, the ease of achieving this similarity raises concerns from potential malicious uses at scale by bad actors, as LLM-generated text becomes increasingly difficult to discern from human text. Although detection methods have been developed to address this issue, bad actors can further manipulate LLM-generated texts to make them less detectable. In this work, we study how further editing texts with Reinforcement Learning from Human Feedback (RLHF), which aligns model outputs with human preferences, affects (a) the quality of generated texts for two tasks, and (b) the performance of LLM-generated text detectors, looking at both training-based and zero-shot detection methods. Although RLHF improves the quality of LLM-generated texts, we find that it also tends to produce more detectable, lengthy, and repetitive outputs. Additionally, we observe that training-based detectors are vulnerable to short texts and to texts that incorporate code, whereas zero-shot detectors exhibit greater robustness.
comment: 14 pages, 3 figures
☆ Dynamic Gradient Sparse Update for Edge Training ISCA
Training on edge devices enables personalized model fine-tuning to enhance real-world performance and maintain data privacy. However, the gradient computation for backpropagation in the training requires significant memory buffers to store intermediate features and compute losses. This is unacceptable for memory-constrained edge devices such as microcontrollers. To tackle this issue, we propose a training acceleration method using dynamic gradient sparse updates. This method updates the important channels and layers only and skips gradient computation for the less important channels and layers to reduce memory usage for each update iteration. In addition, the channel selection is dynamic for different iterations to traverse most of the parameters in the update layers along the time dimension for better performance. The experimental result shows that the proposed method enables an ImageNet pre-trained MobileNetV2 trained on CIFAR-10 to achieve an accuracy of 85.77\% while updating only 2\% of convolution weights within 256KB on-chip memory. This results in a remarkable 98\% reduction in feature memory usage compared to dense model training.
comment: published in IEEE International Symposium on Circuits and Systems (IEEE ISCAS 2024)
☆ Human-AI Interaction and User Satisfaction: Empirical Evidence from Online Reviews of AI Products
Human-AI Interaction (HAI) guidelines and design principles have become increasingly important in both industry and academia to guide the development of AI systems that align with user needs and expectations. However, large-scale empirical evidence on how HAI principles shape user satisfaction in practice remains limited. This study addresses that gap by analyzing over 100,000 user reviews of AI-related products from G2.com, a leading review platform for business software and services. Based on widely adopted industry guidelines, we identify seven core HAI dimensions and examine their coverage and sentiment within the reviews. We find that the sentiment on four HAI dimensions-adaptability, customization, error recovery, and security-is positively associated with overall user satisfaction. Moreover, we show that engagement with HAI dimensions varies by professional background: Users with technical job roles are more likely to discuss system-focused aspects, such as reliability, while non-technical users emphasize interaction-focused features like customization and feedback. Interestingly, the relationship between HAI sentiment and overall satisfaction is not moderated by job role, suggesting that once an HAI dimension has been identified by users, its effect on satisfaction is consistent across job roles.
☆ Physics-Guided Multi-Fidelity DeepONet for Data-Efficient Flow Field Prediction
This study presents an enhanced multi-fidelity deep operator network (DeepONet) framework for efficient spatio-temporal flow field prediction, with particular emphasis on practical scenarios where high-fidelity data is scarce. We introduce several key innovations to improve the framework's efficiency and accuracy. First, we enhance the DeepONet architecture by incorporating a merge network that enables more complex feature interactions between operator and coordinate spaces, achieving a 50.4% reduction in prediction error compared to traditional dot-product operations. We further optimize the architecture through temporal positional encoding and point-based sampling strategies, achieving a 7.57% improvement in prediction accuracy while reducing training time by 96% through efficient sampling and automatic mixed precision training. Building upon this foundation, we develop a transfer learning-based multi-fidelity framework that leverages knowledge from pre-trained low-fidelity models to guide high-fidelity predictions. Our approach freezes the pre-trained branch and trunk networks while making only the merge network trainable during high-fidelity training, preserving valuable low-fidelity representations while efficiently adapting to high-fidelity features. Through systematic investigation, we demonstrate that this fine-tuning strategy not only significantly outperforms linear probing and full-tuning alternatives but also surpasses conventional multi-fidelity frameworks by up to 76%, while achieving up to 43.7% improvement in prediction accuracy compared to single-fidelity training. The core contribution lies in our novel time-derivative guided sampling approach: it maintains prediction accuracy equivalent to models trained with the full dataset while requiring only 60% of the original high-fidelity samples.
☆ An Empirical Study of the Role of Incompleteness and Ambiguity in Interactions with Large Language Models
Natural language as a medium for human-computer interaction has long been anticipated, has been undergoing a sea-change with the advent of Large Language Models (LLMs) with startling capacities for processing and generating language. Many of us now treat LLMs as modern-day oracles, asking it almost any kind of question. Unlike its Delphic predecessor, consulting an LLM does not have to be a single-turn activity (ask a question, receive an answer, leave); and -- also unlike the Pythia -- it is widely acknowledged that answers from LLMs can be improved with additional context. In this paper, we aim to study when we need multi-turn interactions with LLMs to successfully get a question answered; or conclude that a question is unanswerable. We present a neural symbolic framework that models the interactions between human and LLM agents. Through the proposed framework, we define incompleteness and ambiguity in the questions as properties deducible from the messages exchanged in the interaction, and provide results from benchmark problems, in which the answer-correctness is shown to depend on whether or not questions demonstrate the presence of incompleteness or ambiguity (according to the properties we identify). Our results show multi-turn interactions are usually required for datasets which have a high proportion of incompleteness or ambiguous questions; and that that increasing interaction length has the effect of reducing incompleteness or ambiguity. The results also suggest that our measures of incompleteness and ambiguity can be useful tools for characterising interactions with an LLM on question-answeringproblems
☆ Experience Retrieval-Augmentation with Electronic Health Records Enables Accurate Discharge QA
To improve the reliability of Large Language Models (LLMs) in clinical applications, retrieval-augmented generation (RAG) is extensively applied to provide factual medical knowledge. However, beyond general medical knowledge from open-ended datasets, clinical case-based knowledge is also critical for effective medical reasoning, as it provides context grounded in real-world patient experiences. Motivated by this, we propose Experience Retrieval Augmentation - ExpRAG framework based on Electronic Health Record (EHR), aiming to offer the relevant context from other patients' discharge reports. ExpRAG performs retrieval through a coarse-to-fine process, utilizing an EHR-based report ranker to efficiently identify similar patients, followed by an experience retriever to extract task-relevant content for enhanced medical reasoning. To evaluate ExpRAG, we introduce DischargeQA, a clinical QA dataset with 1,280 discharge-related questions across diagnosis, medication, and instruction tasks. Each problem is generated using EHR data to ensure realistic and challenging scenarios. Experimental results demonstrate that ExpRAG consistently outperforms a text-based ranker, achieving an average relative improvement of 5.2%, highlighting the importance of case-based knowledge for medical reasoning.
☆ STShield: Single-Token Sentinel for Real-Time Jailbreak Detection in Large Language Models
Large Language Models (LLMs) have become increasingly vulnerable to jailbreak attacks that circumvent their safety mechanisms. While existing defense methods either suffer from adaptive attacks or require computationally expensive auxiliary models, we present STShield, a lightweight framework for real-time jailbroken judgement. STShield introduces a novel single-token sentinel mechanism that appends a binary safety indicator to the model's response sequence, leveraging the LLM's own alignment capabilities for detection. Our framework combines supervised fine-tuning on normal prompts with adversarial training using embedding-space perturbations, achieving robust detection while preserving model utility. Extensive experiments demonstrate that STShield successfully defends against various jailbreak attacks, while maintaining the model's performance on legitimate queries. Compared to existing approaches, STShield achieves superior defense performance with minimal computational overhead, making it a practical solution for real-world LLM deployment.
comment: 11 pages
☆ WLB-LLM: Workload-Balanced 4D Parallelism for Large Language Model Training
In this work, we present WLB-LLM, a workLoad-balanced 4D parallelism for large language model training. We first thoroughly analyze the workload imbalance issue in LLM training and identify two primary sources of imbalance at the pipeline parallelism and context parallelism levels. Then, to address the imbalance issue, at the pipeline parallelism level, WLB-LLM incorporates a workload-aware variable-length document packing method to balance the computation and communication workload across micro-batches. Additionally, at the context parallelism level, WLB-LLM introduces a novel fine-grained per-document sharding strategy, ensuring each worker within a context parallelism group has an identical workload. Comprehensive experiments under different model scales demonstrate that WLB-LLM significantly mitigates the workload imbalance during 4D parallelism LLM training and achieves an average speedup of 1.23x when applying WLB-LLM in our internal LLM training framework.
comment: 12 pages, 16 figures
☆ Cat-AIR: Content and Task-Aware All-in-One Image Restoration
All-in-one image restoration seeks to recover high-quality images from various types of degradation using a single model, without prior knowledge of the corruption source. However, existing methods often struggle to effectively and efficiently handle multiple degradation types. We present Cat-AIR, a novel \textbf{C}ontent \textbf{A}nd \textbf{T}ask-aware framework for \textbf{A}ll-in-one \textbf{I}mage \textbf{R}estoration. Cat-AIR incorporates an alternating spatial-channel attention mechanism that adaptively balances the local and global information for different tasks. Specifically, we introduce cross-layer channel attentions and cross-feature spatial attentions that allocate computations based on content and task complexity. Furthermore, we propose a smooth learning strategy that allows for seamless adaptation to new restoration tasks while maintaining performance on existing ones. Extensive experiments demonstrate that Cat-AIR achieves state-of-the-art results across a wide range of restoration tasks, requiring fewer FLOPs than previous methods, establishing new benchmarks for efficient all-in-one image restoration.
☆ GLADMamba: Unsupervised Graph-Level Anomaly Detection Powered by Selective State Space Model
Unsupervised graph-level anomaly detection (UGLAD) is a critical and challenging task across various domains, such as social network analysis, anti-cancer drug discovery, and toxic molecule identification. However, existing methods often struggle to capture the long-range dependencies efficiently and neglect the spectral information. Recently, selective State Space Models (SSMs), particularly Mamba, have demonstrated remarkable advantages in capturing long-range dependencies with linear complexity and a selection mechanism. Motivated by their success across various domains, we propose GLADMamba, a novel framework that adapts the selective state space model into UGLAD field. We design View-Fused Mamba (VFM) with a Mamba-Transformer-style architecture to efficiently fuse information from different views with a selective state mechanism. We also design Spectrum-Guided Mamba (SGM) with a Mamba-Transformer-style architecture to leverage the Rayleigh quotient to guide the embedding refining process. GLADMamba can dynamically focus on anomaly-related information while discarding irrelevant information for anomaly detection. To the best of our knowledge, this is the first work to introduce Mamba and explicit spectral information to UGLAD. Extensive experiments on 12 real-world datasets demonstrate that GLADMamba outperforms existing state-of-the-art methods, achieving superior performance in UGLAD. The code is available at https://github.com/Yali-F/GLADMamba.
☆ Generative AI for Validating Physics Laws
We present generative artificial intelligence (AI) to empirically validate fundamental laws of physics, focusing on the Stefan-Boltzmann law linking stellar temperature and luminosity. Our approach simulates counterfactual luminosities under hypothetical temperature regimes for each individual star and iteratively refines the temperature-luminosity relationship in a deep learning architecture. We use Gaia DR3 data and find that, on average, temperature's effect on luminosity increases with stellar radius and decreases with absolute magnitude, consistent with theoretical predictions. By framing physics laws as causal problems, our method offers a novel, data-driven approach to refine theoretical understanding and inform evidence-based policy and practice.
☆ Enhanced prediction of spine surgery outcomes using advanced machine learning techniques and oversampling methods
The study proposes an advanced machine learning approach to predict spine surgery outcomes by incorporating oversampling techniques and grid search optimization. A variety of models including GaussianNB, ComplementNB, KNN, Decision Tree, and optimized versions with RandomOverSampler and SMOTE were tested on a dataset of 244 patients, which included pre-surgical, psychometric, socioeconomic, and analytical variables. The enhanced KNN models achieved up to 76% accuracy and a 67% F1-score, while grid-search optimization further improved performance. The findings underscore the potential of these advanced techniques to aid healthcare professionals in decision-making, with future research needed to refine these models on larger and more diverse datasets.
☆ LLMs in the Classroom: Outcomes and Perceptions of Questions Written with the Aid of AI AAAI 2025
We randomly deploy questions constructed with and without use of the LLM tool and gauge the ability of the students to correctly answer, as well as their ability to correctly perceive the difference between human-authored and LLM-authored questions. In determining whether the questions written with the aid of ChatGPT were consistent with the instructor's questions and source text, we computed representative vectors of both the human and ChatGPT questions using SBERT and compared cosine similarity to the course textbook. A non-significant Mann-Whitney U test (z = 1.018, p = .309) suggests that students were unable to perceive whether questions were written with or without the aid of ChatGPT. However, student scores on LLM-authored questions were almost 9% lower (z = 2.702, p < .01). This result may indicate that either the AI questions were more difficult or that the students were more familiar with the instructor's style of questions. Overall, the study suggests that while there is potential for using LLM tools to aid in the construction of assessments, care must be taken to ensure that the questions are fair, well-composed, and relevant to the course material.
comment: Accepted to AAAI 2025 Technical Track on AI Alignment
☆ HH4AI: A methodological Framework for AI Human Rights impact assessment under the EUAI ACT
This paper introduces the HH4AI Methodology, a structured approach to assessing the impact of AI systems on human rights, focusing on compliance with the EU AI Act and addressing technical, ethical, and regulatory challenges. The paper highlights AIs transformative nature, driven by autonomy, data, and goal-oriented design, and how the EU AI Act promotes transparency, accountability, and safety. A key challenge is defining and assessing "high-risk" AI systems across industries, complicated by the lack of universally accepted standards and AIs rapid evolution. To address these challenges, the paper explores the relevance of ISO/IEC and IEEE standards, focusing on risk management, data quality, bias mitigation, and governance. It proposes a Fundamental Rights Impact Assessment (FRIA) methodology, a gate-based framework designed to isolate and assess risks through phases including an AI system overview, a human rights checklist, an impact assessment, and a final output phase. A filtering mechanism tailors the assessment to the system's characteristics, targeting areas like accountability, AI literacy, data governance, and transparency. The paper illustrates the FRIA methodology through a fictional case study of an automated healthcare triage service. The structured approach enables systematic filtering, comprehensive risk assessment, and mitigation planning, effectively prioritizing critical risks and providing clear remediation strategies. This promotes better alignment with human rights principles and enhances regulatory compliance.
comment: 19 pages, 7 figures, 1 table
♻ ☆ SFO: Piloting VLM Feedback for Offline RL
While internet-scale image and textual data have enabled strong generalization in Vision-Language Models (VLMs), the absence of internet-scale control data has impeded the development of similar generalization in standard reinforcement learning (RL) agents. Although VLMs are fundamentally limited in their ability to solve control tasks due to their lack of action-conditioned training data, their capacity for image understanding allows them to provide valuable feedback in RL tasks by recognizing successful outcomes. A key challenge in Reinforcement Learning from AI Feedback (RLAIF) is determining how best to integrate VLM-derived signals into the learning process. We explore this question in the context of offline RL and introduce a class of methods called sub-trajectory filtered optimization. We identify three key insights. First, trajectory length plays a crucial role in offline RL, as full-trajectory preference learning exacerbates the stitching problem, necessitating the use of sub-trajectories. Second, even in Markovian environments, a non-Markovian reward signal from a sequence of images is required to assess trajectory improvement, as VLMs do not interpret control actions and must rely on visual cues over time. Third, a simple yet effective approach--filtered and weighted behavior cloning--consistently outperforms more complex reinforcement learning from human feedback-based methods. We propose sub-trajectory filtered behavior cloning, a method that leverages VLM feedback on sub-trajectories while incorporating a retrospective filtering mechanism that removes sub-trajectories preceding failures to improve robustness and prevent turbulence. This study is preliminary; we provide initial evidence through evaluations on a toy control domain. Please enjoy our airport puns.
comment: Code is provided at https://github.com/jacooba/OfflineRLAIF
♻ ☆ Computation-Efficient and Recognition-Friendly 3D Point Cloud Privacy Protection CVPR2025
3D point cloud has been widely used in applications such as self-driving cars, robotics, CAD models, etc. To the best of our knowledge, these applications raised the issue of privacy leakage in 3D point clouds, which has not been studied well. Different from the 2D image privacy, which is related to texture and 2D geometric structure, the 3D point cloud is texture-less and only relevant to 3D geometric structure. In this work, we defined the 3D point cloud privacy problem and proposed an efficient privacy-preserving framework named PointFlowGMM that can support downstream classification and segmentation tasks without seeing the original data. Using a flow-based generative model, the point cloud is projected into a latent Gaussian mixture distributed subspace. We further designed a novel angular similarity loss to obfuscate the original geometric structure and reduce the model size from 767MB to 120MB without a decrease in recognition performance. The projected point cloud in the latent space is orthogonally rotated randomly to further protect the original geometric structure, the class-to-class relationship is preserved after rotation, thus, the protected point cloud can support the recognition task. We evaluated our model on multiple datasets and achieved comparable recognition results on encrypted point clouds compared to the original point clouds.
comment: Accepted by CVPR2025
♻ ☆ When neural implant meets multimodal LLM: A dual-loop system for neuromodulation and naturalistic neuralbehavioral research
We propose a novel dual-loop system that synergistically combines responsive neurostimulation (RNS) implants with artificial intelligence-driven wearable devices for treating post-traumatic stress disorder (PTSD) and enabling naturalistic brain research. In PTSD Therapy Mode, an implanted closed-loop neural device monitors amygdala activity and provides on-demand stimulation upon detecting pathological theta oscillations, while an ensemble of wearables (smart glasses, smartwatches, smartphones) uses multimodal large language model (LLM) analysis of sensory data to detect environmental or physiological PTSD triggers and deliver timely audiovisual interventions. Logged events from both the neural and wearable loops are analyzed to personalize trigger detection and progressively transition patients to non-invasive interventions. In Neuroscience Research Mode, the same platform is adapted for real-world brain activity capture. Wearable-LLM systems recognize naturalistic events (social interactions, emotional situations, compulsive behaviors, decision making) and signal implanted RNS devices (via wireless triggers) to record synchronized intracranial data during these moments. This approach builds on recent advances in mobile intracranial EEG recording and closed-loop neuromodulation in humans (BRAIN Initiative, 2023) (Mobbs et al., 2021). We discuss how our interdisciplinary system could revolutionize PTSD therapy and cognitive neuroscience by enabling 24/7 monitoring, context-aware intervention, and rich data collection outside traditional labs. The vision is a future where AI-enhanced devices continuously collaborate with the human brain, offering therapeutic support and deep insights into neural function, with the resulting real-world context rich neural data, in turn, accelerating the development of more biologically-grounded and human-centric AI.
♻ ☆ Constructing and explaining machine learning models for chemistry: example of the exploration and design of boron-based Lewis acids
The integration of machine learning (ML) into chemistry offers transformative potential in the design of molecules with targeted properties. However, the focus has often been on creating highly efficient predictive models, sometimes at the expense of interpretability. In this study, we leverage explainable AI techniques to explore the rational design of boron-based Lewis acids, which play a pivotal role in organic reactions due to their electron-ccepting properties. Using Fluoride Ion Affinity as a proxy for Lewis acidity, we developed interpretable ML models based on chemically meaningful descriptors, including ab initio computed features and substituent-based parameters derived from the Hammett linear free-energy relationship. By constraining the chemical space to well-defined molecular scaffolds, we achieved highly accurate predictions (mean absolute error < 6 kJ/mol), surpassing conventional black-box deep learning models in low-data regimes. Interpretability analyses of the models shed light on the origin of Lewis acidity in these compounds and identified actionable levers to modulate it through the nature and positioning of substituents on the molecular scaffold. This work bridges ML and chemist's way of thinking, demonstrating how explainable models can inspire molecular design and enhance scientific understanding of chemical reactivity.
comment: Main text is 14 pages, 7 figures, 1 scheme. Supporting information is 25 pages. For associated code and datasets, see https://github.com/jfenogli/XAI_boron_LA
♻ ☆ GhostWriter: Augmenting Collaborative Human-AI Writing Experiences Through Personalization and Agency
Writing is a well-established practice to support ideation and creativity. While Large Language Models (LLMs) have become ubiquitous in providing different kinds of writing assistance to different writers, LLM-powered writing systems often fall short in capturing the nuanced personalization and control necessary for effective support and creative exploration. To address these challenges, we introduce GhostWriter, an AI-enhanced writing design probe that enables users to exercise enhanced agency and personalization. GhostWriter leverages LLMs to implicitly learn the user's intended writing style for seamless personalization, while exposing explicit teaching moments for style refinement and reflection. We study 18 participants who use GhostWriter for editing and creative tasks, observing that it helps users craft personalized text and empowers them by providing multiple ways to steer system output. Based on this study, we present insights on people's relationships with AI-assisted writing and offer design recommendations to promote user agency in similar co-creative systems.
comment: 23 pages, 12 figures
♻ ☆ Atom of Thoughts for Markov LLM Test-Time Scaling
Large Language Models (LLMs) achieve superior performance through training-time scaling, and test-time scaling further enhances their capabilities by conducting effective reasoning during inference. However, as the scale of reasoning increases, existing test-time scaling methods suffer from accumulated historical information, which not only wastes computational resources but also interferes with effective reasoning. To address this issue, we observe that complex reasoning can be achieved by solving a series of independent and self-contained subquestions. These subquestions are essentially \textit{atomic questions}, exhibiting the memoryless property similar to Markov processes. Based on this observation, we propose Atom of Thoughts (\our), where each state transition consists of decomposing the current question into a dependency-based directed acyclic graph and contracting its subquestions, forming a simplified question that maintains answer equivalence with the original problem. This answer preservation enables the iterative \textit{decomposition-contraction} process to naturally form a meaningful Markov reasoning process. Furthermore, these atomic states can be seamlessly integrated into existing test-time scaling methods, enabling \our to serve as a plug-in enhancement for improving reasoning capabilities. Experiments across six benchmarks demonstrate the effectiveness of \our both as a standalone framework and a plug-in enhancement. Notably, on HotpotQA, when applied to gpt-4o-mini, \our achieves an \textbf{80.6\%} F1 score, surpassing o3-mini by \textbf{3.4\%} and DeepSeek-R1 by \textbf{10.6\%}. The code is available at \href{https://github.com/qixucen/atom}{https://github.com/qixucen/atom}.
♻ ☆ The Representational Status of Deep Learning Models
This paper aims to clarify the representational status of Deep Learning Models (DLMs). While commonly referred to as 'representations', what this entails is ambiguous due to a conflation of functional and relational conceptions of representation. This paper argues that while DLMs represent their targets in a relational sense, in general, we have no good reason to believe that DLMs encode locally semantically decomposable representations of their targets. That is, the representational capacity these models have is largely global, rather than decomposable into stable, local subrepresentations. This result has immediate implications for explainable AI (XAI) and directs attention toward exploring the global relational nature of deep learning representations and their relationship both to models more generally to understand their potential role in future scientific inquiry.
comment: 18 pages, 1 figure
♻ ☆ Balancing Content Size in RAG-Text2SQL System
Large Language Models (LLMs) have emerged as a promising solution for converting natural language queries into SQL commands, enabling seamless database interaction. However, these Text-to-SQL (Text2SQL) systems face inherent limitations, hallucinations, outdated knowledge, and untraceable reasoning. To address these challenges, the integration of retrieval-augmented generation (RAG) with Text2SQL models has gained traction. RAG serves as a retrieval mechanism, providing essential contextual information, such as table schemas and metadata, to enhance the query generation process. Despite their potential, RAG + Text2SQL systems are susceptible to the quality and size of retrieved documents. While richer document content can improve schema relevance and retrieval accuracy, it also introduces noise, increasing the risk of hallucinations and reducing query fidelity as the prompt size of the Text2SQL model increases. This research investigates the nuanced trade-off between document size and quality, aiming to strike a balance that optimizes system performance. Key thresholds are identified where performance degradation occurs, along with actionable strategies to mitigate these challenges. Additionally, we explore the phenomenon of hallucinations in Text2SQL models, emphasizing the critical role of curated document presentation in minimizing errors. Our findings provide a roadmap for enhancing the robustness of RAG + Text2SQL systems, offering practical insights for real-world applications.
♻ ☆ TSPRank: Bridging Pairwise and Listwise Methods with a Bilinear Travelling Salesman Model KDD 2025
Traditional Learning-To-Rank (LETOR) approaches, including pairwise methods like RankNet and LambdaMART, often fall short by solely focusing on pairwise comparisons, leading to sub-optimal global rankings. Conversely, deep learning based listwise methods, while aiming to optimise entire lists, require complex tuning and yield only marginal improvements over robust pairwise models. To overcome these limitations, we introduce Travelling Salesman Problem Rank (TSPRank), a hybrid pairwise-listwise ranking method. TSPRank reframes the ranking problem as a Travelling Salesman Problem (TSP), a well-known combinatorial optimisation challenge that has been extensively studied for its numerous solution algorithms and applications. This approach enables the modelling of pairwise relationships and leverages combinatorial optimisation to determine the listwise ranking. This approach can be directly integrated as an additional component into embeddings generated by existing backbone models to enhance ranking performance. Our extensive experiments across three backbone models on diverse tasks, including stock ranking, information retrieval, and historical events ordering, demonstrate that TSPRank significantly outperforms both pure pairwise and listwise methods. Our qualitative analysis reveals that TSPRank's main advantage over existing methods is its ability to harness global information better while ranking. TSPRank's robustness and superior performance across different domains highlight its potential as a versatile and effective LETOR solution.
comment: Accepted to ACM SIGKDD 2025 Research Track. The code and preprocessed data are available at https://github.com/waylonli/TSPRank-KDD2025
♻ ☆ COVID 19 Diagnosis Analysis using Transfer Learning
Coronaviruses, including SARS-CoV-2, are responsible for COVID-19, a highly transmissible disease that emerged in December 2019 in Wuhan, China. During the past five years, significant advancements have been made in understanding and mitigating the virus. Although the initial outbreak led to global health crises, improved vaccination strategies, antiviral treatments, and AI-driven diagnostic tools have contributed to better disease management. However, COVID-19 continues to pose risks, particularly for immuno-compromised individuals and those with pre-existing conditions. This study explores the use of deep learning for a rapid and accurate diagnosis of COVID-19, addressing ongoing challenges in healthcare infrastructure and testing accessibility. We propose an enhanced automated detection system leveraging state-of-the-art convolutional neural networks (CNNs), including updated versions of VGG16, VGG19, and ResNet50, to classify COVID-19 infections from chest radiographs and computerized tomography (CT) scans. Our results, based on an expanded dataset of over 6000 medical images, demonstrate that the optimized ResNet50 model achieves the highest classification performance, with 97.77% accuracy, 100% sensitivity, 93.33% specificity, and a 98.0% F1-score. These findings reinforce the potential of AI-assisted diagnostic tools in improving early detection and pandemic preparedness.
♻ ☆ A semantic approach to mapping the Provenance Ontology to Basic Formal Ontology
The Provenance Ontology (PROV-O) is a World Wide Web Consortium (W3C) recommended ontology used to structure data about provenance across a wide variety of domains. Basic Formal Ontology (BFO) is a top-level ontology ISO/IEC standard used to structure a wide variety of ontologies, such as the OBO Foundry ontologies and the Common Core Ontologies (CCO). To enhance interoperability between these two ontologies, their extensions, and data organized by them, a mapping methodology and set of alignments are presented according to specific criteria which prioritize semantic and logical principles. The ontology alignments are evaluated by checking their logical consistency with canonical examples of PROV-O instances and querying terms that do not satisfy the alignment criteria as formalized in SPARQL. A variety of semantic web technologies are used in support of FAIR (Findable, Accessible, Interoperable, Reusable) principles.
comment: 31 pages, 12 figures. This version of the article has been accepted for publication, after peer review (when applicable) but is not the Version of Record and does not reflect post-acceptance improvements, or any corrections. The Version of Record is available online at: https://doi.org/10.1038/s41597-025-04580-1
♻ ☆ FALCON: Feedback-driven Adaptive Long/short-term memory reinforced Coding Optimization system
Recently, large language models (LLMs) have achieved significant progress in automated code generation. Despite their strong instruction-following capabilities, these models frequently struggled to align with user intent in coding scenarios. In particular, they were hampered by datasets that lacked diversity and failed to address specialized tasks or edge cases. Furthermore, challenges in supervised fine-tuning (SFT) and reinforcement learning from human feedback (RLHF) led to failures in generating precise, human-intent-aligned code. To tackle these challenges and improve the code generation performance for automated programming systems, we propose Feedback-driven Adaptive Long/short-term memory reinforced Coding Optimization (i.e., FALCON). FALCON is structured into two hierarchical levels. From the global level, long-term memory improves code quality by retaining and applying learned knowledge. At the local level, short-term memory allows for the incorporation of immediate feedback from compilers and AI systems. Additionally, we introduce meta-reinforcement learning with feedback rewards to solve the global-local bi-level optimization problem and enhance the model's adaptability across diverse code generation tasks. Extensive experiments demonstrate that our technique achieves state-of-the-art performance, leading other reinforcement learning methods by more than 4.5 percentage points on the MBPP benchmark and 6.1 percentage points on the Humaneval benchmark. The open-sourced code is publicly available at https://github.com/titurte/FALCON.
comment: 20 pages, 7 figures
♻ ☆ Trajectory Imputation in Multi-Agent Sports with Derivative-Accumulating Self-Ensemble
Multi-agent trajectory data collected from domains such as team sports often suffer from missing values due to various factors. While many imputation methods have been proposed for spatiotemporal data, they are not well-suited for multi-agent sports scenarios where player movements are highly dynamic and inter-agent interactions continuously evolve. To address these challenges, we propose MIDAS (Multi-agent Imputer with Derivative-Accumulating Self-ensemble), a framework that imputes multi-agent trajectories with high accuracy and physical plausibility. It jointly predicts positions, velocities, and accelerations through a Set Transformer-based neural network and generates alternative estimates by recursively accumulating predicted velocity and acceleration values. These predictions are then combined using a learnable weighted ensemble to produce final imputed trajectories. Experiments on three sports datasets demonstrate that MIDAS significantly outperforms existing baselines in both positional accuracy and physical plausibility. Lastly, we showcase use cases of MIDAS, such as approximating total distance and pass success probability, to highlight its applicability to practical downstream tasks that require complete tracking data.
♻ ☆ Adaptive Intelligence: leveraging insights from adaptive behavior in animals to build flexible AI systems
Biological intelligence is inherently adaptive -- animals continually adjust their actions based on environmental feedback. However, creating adaptive artificial intelligence (AI) remains a major challenge. The next frontier is to go beyond traditional AI to develop "adaptive intelligence," defined here as harnessing insights from biological intelligence to build agents that can learn online, generalize, and rapidly adapt to changes in their environment. Recent advances in neuroscience offer inspiration through studies that increasingly focus on how animals naturally learn and adapt their world models. In this Perspective, I will review the behavioral and neural foundations of adaptive biological intelligence, the parallel progress in AI, and explore brain-inspired approaches for building more adaptive algorithms.
comment: 10 pages, 4 figures
♻ ☆ Inductive Moment Matching
Diffusion models and Flow Matching generate high-quality samples but are slow at inference, and distilling them into few-step models often leads to instability and extensive tuning. To resolve these trade-offs, we propose Inductive Moment Matching (IMM), a new class of generative models for one- or few-step sampling with a single-stage training procedure. Unlike distillation, IMM does not require pre-training initialization and optimization of two networks; and unlike Consistency Models, IMM guarantees distribution-level convergence and remains stable under various hyperparameters and standard model architectures. IMM surpasses diffusion models on ImageNet-256x256 with 1.99 FID using only 8 inference steps and achieves state-of-the-art 2-step FID of 1.98 on CIFAR-10 for a model trained from scratch.
♻ ☆ From Knowledge Generation to Knowledge Verification: Examining the BioMedical Generative Capabilities of ChatGPT
The generative capabilities of LLM models offer opportunities for accelerating tasks but raise concerns about the authenticity of the knowledge they produce. To address these concerns, we present a computational approach that evaluates the factual accuracy of biomedical knowledge generated by an LLM. Our approach consists of two processes: generating disease-centric associations and verifying these associations using the semantic framework of biomedical ontologies. Using ChatGPT as the selected LLM, we designed prompt-engineering processes to establish linkages between diseases and related drugs, symptoms, and genes, and assessed consistency across multiple ChatGPT models (e.g., GPT-turbo, GPT-4, etc.). Experimental results demonstrate high accuracy in identifying disease terms (88%-97%), drug names (90%-91%), and genetic information (88%-98%). However, symptom term identification was notably lower (49%-61%), due to the informal and verbose nature of symptom descriptions, which hindered effective semantic matching with the formal language of specialized ontologies. Verification of associations reveals literature coverage rates of 89%-91% for disease-drug and disease-gene pairs, while symptom-related associations exhibit lower coverage (49%-62%).
comment: 28 pages, 6 figures, In Review with a Cell Press Journal
♻ ☆ DeepInnovation AI: A Global Dataset Mapping the AI innovation from Academic Research to Industrial Patents
In the rapidly evolving field of artificial intelligence (AI), mapping innovation patterns and understanding effective technology transfer from research to applications are essential for economic growth. However, existing data infrastructures suffer from fragmentation, incomplete coverage, and insufficient evaluative capacity. Here, we present DeepInnovationAI, a comprehensive global dataset containing three structured files. DeepPatentAI.csv: Contains 2,356,204 patent records with 8 field-specific attributes. DeepDiveAI.csv: Encompasses 3,511,929 academic publications with 13 metadata fields. These two datasets leverage large language models, multilingual text analysis and dual-layer BERT classifiers to accurately identify AI-related content, while utilizing hypergraph analysis to create robust innovation metrics. Additionally, DeepCosineAI.csv: By applying semantic vector proximity analysis, this file presents approximately one hundred million calculated paper-patent similarity pairs to enhance understanding of how theoretical advancements translate into commercial technologies. DeepInnovationAI enables researchers, policymakers, and industry leaders to anticipate trends and identify collaboration opportunities. With extensive temporal and geographical scope, it supports detailed analysis of technological development patterns and international competition dynamics, establishing a foundation for modeling AI innovation and technology transfer processes.
comment: 32 pages and 8 figures
♻ ☆ MV-MATH: Evaluating Multimodal Math Reasoning in Multi-Visual Contexts
Multimodal Large Language Models (MLLMs) have shown promising capabilities in mathematical reasoning within visual contexts across various datasets. However, most existing multimodal math benchmarks are limited to single-visual contexts, which diverges from the multi-visual scenarios commonly encountered in real-world mathematical applications. To address this gap, we introduce MV-MATH: a meticulously curated dataset of 2,009 high-quality mathematical problems. Each problem integrates multiple images interleaved with text, derived from authentic K-12 scenarios, and enriched with detailed annotations. MV-MATH includes multiple-choice, free-form, and multi-step questions, covering 11 subject areas across 3 difficulty levels, and serves as a comprehensive and rigorous benchmark for assessing MLLMs' mathematical reasoning in multi-visual contexts. Through extensive experimentation, we observe that MLLMs encounter substantial challenges in multi-visual math tasks, with a considerable performance gap relative to human capabilities on MV-MATH. Furthermore, we analyze the performance and error patterns of various models, providing insights into MLLMs' mathematical reasoning capabilities within multi-visual settings.
comment: 47 pages
♻ ☆ Large Language Models: A Survey
Large Language Models (LLMs) have drawn a lot of attention due to their strong performance on a wide range of natural language tasks, since the release of ChatGPT in November 2022. LLMs' ability of general-purpose language understanding and generation is acquired by training billions of model's parameters on massive amounts of text data, as predicted by scaling laws \cite{kaplan2020scaling,hoffmann2022training}. The research area of LLMs, while very recent, is evolving rapidly in many different ways. In this paper, we review some of the most prominent LLMs, including three popular LLM families (GPT, LLaMA, PaLM), and discuss their characteristics, contributions and limitations. We also give an overview of techniques developed to build, and augment LLMs. We then survey popular datasets prepared for LLM training, fine-tuning, and evaluation, review widely used LLM evaluation metrics, and compare the performance of several popular LLMs on a set of representative benchmarks. Finally, we conclude the paper by discussing open challenges and future research directions.
♻ ☆ ModeSeq: Taming Sparse Multimodal Motion Prediction with Sequential Mode Modeling CVPR 2025
Anticipating the multimodality of future events lays the foundation for safe autonomous driving. However, multimodal motion prediction for traffic agents has been clouded by the lack of multimodal ground truth. Existing works predominantly adopt the winner-take-all training strategy to tackle this challenge, yet still suffer from limited trajectory diversity and uncalibrated mode confidence. While some approaches address these limitations by generating excessive trajectory candidates, they necessitate a post-processing stage to identify the most representative modes, a process lacking universal principles and compromising trajectory accuracy. We are thus motivated to introduce ModeSeq, a new multimodal prediction paradigm that models modes as sequences. Unlike the common practice of decoding multiple plausible trajectories in one shot, ModeSeq requires motion decoders to infer the next mode step by step, thereby more explicitly capturing the correlation between modes and significantly enhancing the ability to reason about multimodality. Leveraging the inductive bias of sequential mode prediction, we also propose the Early-Match-Take-All (EMTA) training strategy to diversify the trajectories further. Without relying on dense mode prediction or heuristic post-processing, ModeSeq considerably improves the diversity of multimodal output while attaining satisfactory trajectory accuracy, resulting in balanced performance on motion prediction benchmarks. Moreover, ModeSeq naturally emerges with the capability of mode extrapolation, which supports forecasting more behavior modes when the future is highly uncertain.
comment: CVPR 2025
♻ ☆ Beautimeter: Harnessing GPT for Assessing Architectural and Urban Beauty based on the 15 Properties of Living Structure
Beautimeter is a new tool powered by generative pre-trained transformer (GPT) technology, designed to evaluate architectural and urban beauty. Rooted in Christopher Alexander's theory of centers, this work builds on the idea that all environments possess, to varying degrees, an innate sense of life. Alexander identified 15 fundamental properties, such as levels of scale and thick boundaries, that characterize living structure, which Beautimeter uses as a basis for its analysis. By integrating GPT's advanced natural language processing capabilities, Beautimeter assesses the extent to which a structure embodies these 15 properties, enabling a nuanced evaluation of architectural and urban aesthetics. Using ChatGPT, the tool helps users generate insights into the perceived beauty and coherence of spaces. We conducted a series of case studies, evaluating images of architectural and urban environments, as well as carpets, paintings, and other artifacts. The results demonstrate Beautimeter's effectiveness in analyzing aesthetic qualities across diverse contexts. Our findings suggest that by leveraging GPT technology, Beautimeter offers architects, urban planners, and designers a powerful tool to create spaces that resonate deeply with people. This paper also explores the implications of such technology for architecture and urban design, highlighting its potential to enhance both the design process and the assessment of built environments. Keywords: Living structure, structural beauty, Christopher Alexander, AI in Design, human centered design
comment: 12 pages, 6 figure, and 3 tables
♻ ☆ Open-Sora 2.0: Training a Commercial-Level Video Generation Model in $200k
Video generation models have achieved remarkable progress in the past year. The quality of AI video continues to improve, but at the cost of larger model size, increased data quantity, and greater demand for training compute. In this report, we present Open-Sora 2.0, a commercial-level video generation model trained for only $200k. With this model, we demonstrate that the cost of training a top-performing video generation model is highly controllable. We detail all techniques that contribute to this efficiency breakthrough, including data curation, model architecture, training strategy, and system optimization. According to human evaluation results and VBench scores, Open-Sora 2.0 is comparable to global leading video generation models including the open-source HunyuanVideo and the closed-source Runway Gen-3 Alpha. By making Open-Sora 2.0 fully open-source, we aim to democratize access to advanced video generation technology, fostering broader innovation and creativity in content creation. All resources are publicly available at: https://github.com/hpcaitech/Open-Sora.
♻ ☆ MMAC-Copilot: Multi-modal Agent Collaboration Operating Copilot
Large language model agents that interact with PC applications often face limitations due to their singular mode of interaction with real-world environments, leading to restricted versatility and frequent hallucinations. To address this, we propose the Multi-Modal Agent Collaboration framework (MMAC-Copilot), a framework utilizes the collective expertise of diverse agents to enhance interaction ability with application. The framework introduces a team collaboration chain, enabling each participating agent to contribute insights based on their specific domain knowledge, effectively reducing the hallucination associated with knowledge domain gaps. We evaluate MMAC-Copilot using the GAIA benchmark and our newly introduced Visual Interaction Benchmark (VIBench). MMAC-Copilot achieved exceptional performance on GAIA, with an average improvement of 6.8\% over existing leading systems. VIBench focuses on non-API-interactable applications across various domains, including 3D gaming, recreation, and office scenarios. It also demonstrated remarkable capability on VIBench. We hope this work can inspire in this field and provide a more comprehensive assessment of Autonomous agents. The anonymous Github is available at \href{https://anonymous.4open.science/r/ComputerAgentWithVision-3C12}{Anonymous Github}
comment: Technical Reports
♻ ☆ Certifiably Robust Policies for Uncertain Parametric Environments
We present a data-driven approach for producing policies that are provably robust across unknown stochastic environments. Existing approaches can learn models of a single environment as an interval Markov decision processes (IMDP) and produce a robust policy with a probably approximately correct (PAC) guarantee on its performance. However these are unable to reason about the impact of environmental parameters underlying the uncertainty. We propose a framework based on parametric Markov decision processes (MDPs) with unknown distributions over parameters. We learn and analyse IMDPs for a set of unknown sample environments induced by parameters. The key challenge is then to produce meaningful performance guarantees that combine the two layers of uncertainty: (1) multiple environments induced by parameters with an unknown distribution; (2) unknown induced environments which are approximated by IMDPs. We present a novel approach based on scenario optimisation that yields a single PAC guarantee quantifying the risk level for which a specified performance level can be assured in unseen environments, plus a means to trade-off risk and performance. We implement and evaluate our framework using multiple robust policy generation methods on a range of benchmarks. We show that our approach produces tight bounds on a policy's performance with high confidence.
♻ ☆ A new approach for encoding code and assisting code understanding
Some companies (e.g., Microsoft Research and Google DeepMind) have discovered some of the limitations of GPTs' autoregressive paradigm next-word prediction, manifested in the model's lack of planning, working memory, backtracking, and reasoning skills. GPTs rely on a local and greedy process of generating the next word, without a global understanding of the task or the output. We have confirmed the above limitations through specialized empirical studies of code comprehension. Although GPT-4 is good at producing fluent and coherent text, it cannot handle complex logic and generate new code that hasn't been seen, and it relies too much on the formatting of the prompt to generate the correct code. We propose a new paradigm for code understanding that goes beyond the next-word prediction paradigm, inspired by the successful application of diffusion techniques to image generation (Dalle-2, Sora) and protein structure generation (AlphaFold-3), which have no autoregressive constraints. Instead of encoding the code in a form that mimics natural language, we encode the code as a heterogeneous image paradigm with a memory of global information that mimics both images and protein structures. We then refer to Sora's CLIP upstream text-to-image encoder model to design a text-to-code encoder model that can be applied to various downstream code understanding tasks. The model learns the global understanding of code under the new paradigm heterogeneous image, connects the encoding space of text and code, and encodes the input of text into the vector of code most similar to it. Using self-supervised comparative learning on 456,360 text-code pairs, the model achieved a zero-shot prediction of new data. This work is the basis for future work on code generation using diffusion techniques under a new paradigm to avoid autoregressive limitations.
comment: 10 page, 14 figures
♻ ☆ VideoAutoArena: An Automated Arena for Evaluating Large Multimodal Models in Video Analysis through User Simulation CVPR 2025
Large multimodal models (LMMs) with advanced video analysis capabilities have recently garnered significant attention. However, most evaluations rely on traditional methods like multiple-choice questions in benchmarks such as VideoMME and LongVideoBench, which are prone to lack the depth needed to capture the complex demands of real-world users. To address this limitation-and due to the prohibitive cost and slow pace of human annotation for video tasks-we introduce VideoAutoArena, an arena-style benchmark inspired by LMSYS Chatbot Arena's framework, designed to automatically assess LMMs' video analysis abilities. VideoAutoArena utilizes user simulation to generate open-ended, adaptive questions that rigorously assess model performance in video understanding. The benchmark features an automated, scalable evaluation framework, incorporating a modified ELO Rating System for fair and continuous comparisons across multiple LMMs. To validate our automated judging system, we construct a 'gold standard' using a carefully curated subset of human annotations, demonstrating that our arena strongly aligns with human judgment while maintaining scalability. Additionally, we introduce a fault-driven evolution strategy, progressively increasing question complexity to push models toward handling more challenging video analysis scenarios. Experimental results demonstrate that VideoAutoArena effectively differentiates among state-of-the-art LMMs, providing insights into model strengths and areas for improvement. To further streamline our evaluation, we introduce VideoAutoBench as an auxiliary benchmark, where human annotators label winners in a subset of VideoAutoArena battles. We use GPT-4o as a judge to compare responses against these human-validated answers. Together, VideoAutoArena and VideoAutoBench offer a cost-effective, and scalable framework for evaluating LMMs in user-centric video analysis.
comment: CVPR 2025, Project Page: https://videoautoarena.github.io/
♻ ☆ Credit Risk Meets Large Language Models: Building a Risk Indicator from Loan Descriptions in P2P Lending
Peer-to-peer (P2P) lending connects borrowers and lenders through online platforms but suffers from significant information asymmetry, as lenders often lack sufficient data to assess borrowers' creditworthiness. This paper addresses this challenge by leveraging BERT, a Large Language Model (LLM) known for its ability to capture contextual nuances in text, to generate a risk score based on borrowers' loan descriptions using a dataset from the Lending Club platform. We fine-tune BERT to distinguish between defaulted and non-defaulted loans using the loan descriptions provided by the borrowers. The resulting BERT-generated risk score is then integrated as an additional feature into an XGBoost classifier used at the loan granting stage, where decision-makers have limited information available to guide their decisions. This integration enhances predictive performance, with improvements in balanced accuracy and AUC, highlighting the value of textual features in complementing traditional inputs. Moreover, we find that the incorporation of the BERT score alters how classification models utilize traditional input variables, with these changes varying by loan purpose. These findings suggest that BERT discerns meaningful patterns in loan descriptions, encompassing borrower-specific features, specific purposes, and linguistic characteristics. However, the inherent opacity of LLMs and their potential biases underscore the need for transparent frameworks to ensure regulatory compliance and foster trust. Overall, this study demonstrates how LLM-derived insights interact with traditional features in credit risk modeling, opening new avenues to enhance the explainability and fairness of these models.
♻ ☆ Diffusion-Driven Semantic Communication for Generative Models with Bandwidth Constraints
Diffusion models have been extensively utilized in AI-generated content (AIGC) in recent years, thanks to the superior generation capabilities. Combining with semantic communications, diffusion models are used for tasks such as denoising, data reconstruction, and content generation. However, existing diffusion-based generative models do not consider the stringent bandwidth limitation, which limits its application in wireless communication. This paper introduces a diffusion-driven semantic communication framework with advanced VAE-based compression for bandwidth-constrained generative model. Our designed architecture utilizes the diffusion model, where the signal transmission process through the wireless channel acts as the forward process in diffusion. To reduce bandwidth requirements, we incorporate a downsampling module and a paired upsampling module based on a variational auto-encoder with reparameterization at the receiver to ensure that the recovered features conform to the Gaussian distribution. Furthermore, we derive the loss function for our proposed system and evaluate its performance through comprehensive experiments. Our experimental results demonstrate significant improvements in pixel-level metrics such as peak signal to noise ratio (PSNR) and semantic metrics like learned perceptual image patch similarity (LPIPS). These enhancements are more profound regarding the compression rates and SNR compared to deep joint source-channel coding (DJSCC).
comment: accepted to IEEE for possible publication
♻ ☆ Gazing at Rewards: Eye Movements as a Lens into Human and AI Decision-Making in Hybrid Visual Foraging
Imagine searching a collection of coins for quarters ($0.25$), dimes ($0.10$), nickels ($0.05$), and pennies ($0.01$)-a hybrid foraging task where observers look for multiple instances of multiple target types. In such tasks, how do target values and their prevalence influence foraging and eye movement behaviors (e.g., should you prioritize rare quarters or common nickels)? To explore this, we conducted human psychophysics experiments, revealing that humans are proficient reward foragers. Their eye fixations are drawn to regions with higher average rewards, fixation durations are longer on more valuable targets, and their cumulative rewards exceed chance, approaching the upper bound of optimal foragers. To probe these decision-making processes of humans, we developed a transformer-based Visual Forager (VF) model trained via reinforcement learning. Our VF model takes a series of targets, their corresponding values, and the search image as inputs, processes the images using foveated vision, and produces a sequence of eye movements along with decisions on whether to collect each fixated item. Our model outperforms all baselines, achieves cumulative rewards comparable to those of humans, and approximates human foraging behavior in eye movements and foraging biases within time-limited environments. Furthermore, stress tests on out-of-distribution tasks with novel targets, unseen values, and varying set sizes demonstrate the VF model's effective generalization. Our work offers valuable insights into the relationship between eye movements and decision-making, with our model serving as a powerful tool for further exploration of this connection. All data, code, and models are available at https://github.com/ZhangLab-DeepNeuroCogLab/visual-forager.
♻ ☆ SupReMix: Supervised Contrastive Learning for Medical Imaging Regression with Mixup
In medical image analysis, regression plays a critical role in computer-aided diagnosis. It enables quantitative measurements such as age prediction from structural imaging, cardiac function quantification, and molecular measurement from PET scans. While deep learning has shown promise for these tasks, most approaches focus solely on optimizing regression loss or model architecture, neglecting the quality of learned feature representations which are crucial for robust clinical predictions. Directly applying representation learning techniques designed for classification to regression often results in fragmented representations in the latent space, yielding sub-optimal performance. In this paper, we argue that the potential of contrastive learning for medical image regression has been overshadowed due to the neglect of two crucial aspects: ordinality-awareness and hardness. To address these challenges, we propose Supervised Contrastive Learning for Medical Imaging Regression with Mixup (SupReMix). It takes anchor-inclusive mixtures (mixup of the anchor and a distinct negative sample) as hard negative pairs and anchor-exclusive mixtures (mixup of two distinct negative samples) as hard positive pairs at the embedding level. This strategy formulates harder contrastive pairs by integrating richer ordinal information. Through theoretical analysis and extensive experiments on six datasets spanning MRI, X-ray, ultrasound, and PET modalities, we demonstrate that SupReMix fosters continuous ordered representations, significantly improving regression performance.
comment: The first two authors equally contributed to this work. Previously titled "Mixup Your Own Pair", content extended and revised
♻ ☆ Heterogeneous bimodal attention fusion for speech emotion recognition
Multi-modal emotion recognition in conversations is a challenging problem due to the complex and complementary interactions between different modalities. Audio and textual cues are particularly important for understanding emotions from a human perspective. Most existing studies focus on exploring interactions between audio and text modalities at the same representation level. However, a critical issue is often overlooked: the heterogeneous modality gap between low-level audio representations and high-level text representations. To address this problem, we propose a novel framework called Heterogeneous Bimodal Attention Fusion (HBAF) for multi-level multi-modal interaction in conversational emotion recognition. The proposed method comprises three key modules: the uni-modal representation module, the multi-modal fusion module, and the inter-modal contrastive learning module. The uni-modal representation module incorporates contextual content into low-level audio representations to bridge the heterogeneous multi-modal gap, enabling more effective fusion. The multi-modal fusion module uses dynamic bimodal attention and a dynamic gating mechanism to filter incorrect cross-modal relationships and fully exploit both intra-modal and inter-modal interactions. Finally, the inter-modal contrastive learning module captures complex absolute and relative interactions between audio and text modalities. Experiments on the MELD and IEMOCAP datasets demonstrate that the proposed HBAF method outperforms existing state-of-the-art baselines.
♻ ☆ One-Shot Manipulation Strategy Learning by Making Contact Analogies ICRA 2025
We present a novel approach, MAGIC (manipulation analogies for generalizable intelligent contacts), for one-shot learning of manipulation strategies with fast and extensive generalization to novel objects. By leveraging a reference action trajectory, MAGIC effectively identifies similar contact points and sequences of actions on novel objects to replicate a demonstrated strategy, such as using different hooks to retrieve distant objects of different shapes and sizes. Our method is based on a two-stage contact-point matching process that combines global shape matching using pretrained neural features with local curvature analysis to ensure precise and physically plausible contact points. We experiment with three tasks including scooping, hanging, and hooking objects. MAGIC demonstrates superior performance over existing methods, achieving significant improvements in runtime speed and generalization to different object categories. Website: https://magic-2024.github.io/ .
comment: ICRA 2025; CoRL LEAP Workshop, 2024
♻ ☆ Optical-Flow Guided Prompt Optimization for Coherent Video Generation CVPR 2025
While text-to-video diffusion models have made significant strides, many still face challenges in generating videos with temporal consistency. Within diffusion frameworks, guidance techniques have proven effective in enhancing output quality during inference; however, applying these methods to video diffusion models introduces additional complexity of handling computations across entire sequences. To address this, we propose a novel framework called MotionPrompt that guides the video generation process via optical flow. Specifically, we train a discriminator to distinguish optical flow between random pairs of frames from real videos and generated ones. Given that prompts can influence the entire video, we optimize learnable token embeddings during reverse sampling steps by using gradients from a trained discriminator applied to random frame pairs. This approach allows our method to generate visually coherent video sequences that closely reflect natural motion dynamics, without compromising the fidelity of the generated content. We demonstrate the effectiveness of our approach across various models.
comment: CVPR 2025 (poster); project page: https://motionprompt.github.io/
♻ ☆ Concept-as-Tree: Synthetic Data is All You Need for VLM Personalization
Vision-Language Models (VLMs) have demonstrated exceptional performance in various multi-modal tasks. Recently, there has been an increasing interest in improving the personalization capabilities of VLMs. To better integrate user-provided concepts into VLMs, many methods use positive and negative samples to fine-tune these models. However, the scarcity of user-provided positive samples and the low quality of retrieved negative samples pose challenges for fine-tuning. To reveal the relationship between sample and model performance, we systematically investigate the impact of positive and negative samples (easy and hard) and their diversity on VLM personalization tasks. Based on the detailed analysis, we introduce Concept-as-Tree (CaT), which represents a concept as a tree structure, thereby enabling the data generation of positive and negative samples with varying difficulty and diversity for VLM personalization. With a well-designed data filtering strategy, our CaT framework can ensure the quality of generated data, constituting a powerful pipeline. We perform thorough experiments with various VLM personalization baselines to assess the effectiveness of the pipeline, alleviating the lack of positive samples and the low quality of negative samples. Our results demonstrate that CaT equipped with the proposed data filter significantly enhances the personalization capabilities of VLMs across the MyVLM, Yo'LLaVA, and MC-LLaVA datasets. To our knowledge, this work is the first controllable synthetic data pipeline for VLM personalization. The code is released at $\href{https://github.com/zengkaiya/CaT}{\text{https://github.com/zengkaiya/CaT}}$.
comment: The code is released at $\href{https://github.com/zengkaiya/CaT}{\text{https://github.com/zengkaiya/CaT}}$
♻ ☆ Nuclear Deployed: Analyzing Catastrophic Risks in Decision-making of Autonomous LLM Agents
Large language models (LLMs) are evolving into autonomous decision-makers, raising concerns about catastrophic risks in high-stakes scenarios, particularly in Chemical, Biological, Radiological and Nuclear (CBRN) domains. Based on the insight that such risks can originate from trade-offs between the agent's Helpful, Harmlessness and Honest (HHH) goals, we build a novel three-stage evaluation framework, which is carefully constructed to effectively and naturally expose such risks. We conduct 14,400 agentic simulations across 12 advanced LLMs, with extensive experiments and analysis. Results reveal that LLM agents can autonomously engage in catastrophic behaviors and deception, without being deliberately induced. Furthermore, stronger reasoning abilities often increase, rather than mitigate, these risks. We also show that these agents can violate instructions and superior commands. On the whole, we empirically prove the existence of catastrophic risks in autonomous LLM agents. We release our code to foster further research.
comment: Please visit https://llm-catastrophic-risks.github.io for a quick tour of our research. Our code is available at https://github.com/pillowsofwind/LLM-CBRN-Risks
♻ ☆ GEMeX: A Large-Scale, Groundable, and Explainable Medical VQA Benchmark for Chest X-ray Diagnosis
Medical Visual Question Answering (Med-VQA) combines computer vision and natural language processing to automatically answer clinical inquiries about medical images. However, current Med-VQA datasets exhibit two significant limitations: (1) they often lack visual and textual explanations for answers, hindering comprehension for patients and junior doctors; (2) they typically offer a narrow range of question formats, inadequately reflecting the diverse requirements in practical scenarios. These limitations pose significant challenges to the development of a reliable and user-friendly Med-VQA system. To address these challenges, we introduce a large-scale, Groundable, and Explainable Medical VQA benchmark for chest X-ray diagnosis (GEMeX), featuring several innovative components: (1) a multi-modal explainability mechanism that offers detailed visual and textual explanations for each question-answer pair, thereby enhancing answer comprehensibility; (2) four question types, open-ended, closed-ended, single-choice, and multiple-choice, to better reflect practical needs. With 151,025 images and 1,605,575 questions, GEMeX is the currently largest chest X-ray VQA dataset. Evaluation of 12 representative large vision language models (LVLMs) on GEMeX reveals suboptimal performance, underscoring the dataset's complexity. Meanwhile, we propose a strong model by fine-tuning an existing LVLM on the GEMeX training set. The substantial performance improvement showcases the dataset's effectiveness. The benchmark is available at https://www.med-vqa.com/GEMeX.
comment: This project is available at https://www.med-vqa.com/GEMeX
♻ ☆ Autonomous Data Selection with Zero-shot Generative Classifiers for Mathematical Texts
We present Autonomous Data Selection (AutoDS), a method that leverages base language models themselves as zero-shot "generative classifiers" to automatically curate high-quality mathematical texts. Unlike prior approaches that require human annotations or training a dedicated data filter, AutoDS relies solely on a model's logits to determine whether a given passage is mathematically informative and educational. By integrating AutoDS into a continual pretraining pipeline, we substantially boost downstream performance on challenging math benchmarks (MATH, GSM8K, and BBH) while using far fewer tokens than previous methods. Empirically, our approach achieves roughly a twofold improvement in pretraining token efficiency over strong baselines, underscoring the potential of self-directed data selection in enhancing mathematical reasoning. We release our curated AutoMathText dataset to facilitate future research in automated domain-specific data curation. The AutoMathText dataset is available at https://huggingface.co/datasets/math-ai/AutoMathText. The code is available at https://github.com/yifanzhang-pro/AutoMathText.
comment: 24 pages, 8 figures. arXiv admin note: text overlap with arXiv:0808.2664, arXiv:0806.2159, arXiv:1703.08834, arXiv:math/0610707 by other authors
♻ ☆ Few-Shot Joint Multimodal Entity-Relation Extraction via Knowledge-Enhanced Cross-modal Prompt Model ACM MM 2024
Joint Multimodal Entity-Relation Extraction (JMERE) is a challenging task that aims to extract entities and their relations from text-image pairs in social media posts. Existing methods for JMERE require large amounts of labeled data. However, gathering and annotating fine-grained multimodal data for JMERE poses significant challenges. Initially, we construct diverse and comprehensive multimodal few-shot datasets fitted to the original data distribution. To address the insufficient information in the few-shot setting, we introduce the \textbf{K}nowledge-\textbf{E}nhanced \textbf{C}ross-modal \textbf{P}rompt \textbf{M}odel (KECPM) for JMERE. This method can effectively address the problem of insufficient information in the few-shot setting by guiding a large language model to generate supplementary background knowledge. Our proposed method comprises two stages: (1) a knowledge ingestion stage that dynamically formulates prompts based on semantic similarity guide ChatGPT generating relevant knowledge and employs self-reflection to refine the knowledge; (2) a knowledge-enhanced language model stage that merges the auxiliary knowledge with the original input and utilizes a transformer-based model to align with JMERE's required output format. We extensively evaluate our approach on a few-shot dataset derived from the JMERE dataset, demonstrating its superiority over strong baselines in terms of both micro and macro F$_1$ scores. Additionally, we present qualitative analyses and case studies to elucidate the effectiveness of our model.
comment: accepted by ACM MM 2024
♻ ☆ MotiF: Making Text Count in Image Animation with Motion Focal Loss CVPR 2025
Text-Image-to-Video (TI2V) generation aims to generate a video from an image following a text description, which is also referred to as text-guided image animation. Most existing methods struggle to generate videos that align well with the text prompts, particularly when motion is specified. To overcome this limitation, we introduce MotiF, a simple yet effective approach that directs the model's learning to the regions with more motion, thereby improving the text alignment and motion generation. We use optical flow to generate a motion heatmap and weight the loss according to the intensity of the motion. This modified objective leads to noticeable improvements and complements existing methods that utilize motion priors as model inputs. Additionally, due to the lack of a diverse benchmark for evaluating TI2V generation, we propose TI2V Bench, a dataset consists of 320 image-text pairs for robust evaluation. We present a human evaluation protocol that asks the annotators to select an overall preference between two videos followed by their justifications. Through a comprehensive evaluation on TI2V Bench, MotiF outperforms nine open-sourced models, achieving an average preference of 72%. The TI2V Bench and additional results are released in https://wang-sj16.github.io/motif/.
comment: Accepted by CVPR 2025. Project page: https://wang-sj16.github.io/motif/
♻ ☆ Toward Conditional Distribution Calibration in Survival Prediction NeurIPS 2024
Survival prediction often involves estimating the time-to-event distribution from censored datasets. Previous approaches have focused on enhancing discrimination and marginal calibration. In this paper, we highlight the significance of conditional calibration for real-world applications -- especially its role in individual decision-making. We propose a method based on conformal prediction that uses the model's predicted individual survival probability at that instance's observed time. This method effectively improves the model's marginal and conditional calibration, without compromising discrimination. We provide asymptotic theoretical guarantees for both marginal and conditional calibration and test it extensively across 15 diverse real-world datasets, demonstrating the method's practical effectiveness and versatility in various settings.
comment: Accepted to NeurIPS 2024. 41 pages, 23 figures
Computer Vision and Pattern Recognition 45
☆ CustomKD: Customizing Large Vision Foundation for Edge Model Improvement via Knowledge Distillation CVPR 2025
We propose a novel knowledge distillation approach, CustomKD, that effectively leverages large vision foundation models (LVFMs) to enhance the performance of edge models (e.g., MobileNetV3). Despite recent advancements in LVFMs, such as DINOv2 and CLIP, their potential in knowledge distillation for enhancing edge models remains underexplored. While knowledge distillation is a promising approach for improving the performance of edge models, the discrepancy in model capacities and heterogeneous architectures between LVFMs and edge models poses a significant challenge. Our observation indicates that although utilizing larger backbones (e.g., ViT-S to ViT-L) in teacher models improves their downstream task performances, the knowledge distillation from the large teacher models fails to bring as much performance gain for student models as for teacher models due to the large model discrepancy. Our simple yet effective CustomKD customizes the well-generalized features inherent in LVFMs to a given student model in order to reduce model discrepancies. Specifically, beyond providing well-generalized original knowledge from teachers, CustomKD aligns the features of teachers to those of students, making it easy for students to understand and overcome the large model discrepancy overall. CustomKD significantly improves the performances of edge models in scenarios with unlabeled data such as unsupervised domain adaptation (e.g., OfficeHome and DomainNet) and semi-supervised learning (e.g., CIFAR-100 with 400 labeled samples and ImageNet with 1% labeled samples), achieving the new state-of-the-art performances.
comment: Accepted to CVPR 2025
☆ PG-SAM: Prior-Guided SAM with Medical for Multi-organ Segmentation
Segment Anything Model (SAM) demonstrates powerful zero-shot capabilities; however, its accuracy and robustness significantly decrease when applied to medical image segmentation. Existing methods address this issue through modality fusion, integrating textual and image information to provide more detailed priors. In this study, we argue that the granularity of text and the domain gap affect the accuracy of the priors. Furthermore, the discrepancy between high-level abstract semantics and pixel-level boundary details in images can introduce noise into the fusion process. To address this, we propose Prior-Guided SAM (PG-SAM), which employs a fine-grained modality prior aligner to leverage specialized medical knowledge for better modality alignment. The core of our method lies in efficiently addressing the domain gap with fine-grained text from a medical LLM. Meanwhile, it also enhances the priors' quality after modality alignment, ensuring more accurate segmentation. In addition, our decoder enhances the model's expressive capabilities through multi-level feature fusion and iterative mask optimizer operations, supporting unprompted learning. We also propose a unified pipeline that effectively supplies high-quality semantic information to SAM. Extensive experiments on the Synapse dataset demonstrate that the proposed PG-SAM achieves state-of-the-art performance. Our anonymous code is released at https://github.com/logan-0623/PG-SAM.
☆ Decoupling Angles and Strength in Low-rank Adaptation ICLR 2025
Parameter-Efficient FineTuning (PEFT) methods have recently gained significant popularity thanks to the widespread availability of large-scale pretrained models. These methods allow for quick adaptation to downstream tasks with minimal computational cost. However, popular finetuning methods such as LoRA exhibit limited robustness when it comes to hyperparameter choices or extended training regimes, preventing optimal out-of-the-box performance. In contrast, bounded approaches, such as ETHER, provide greater robustness but are limited to extremely low-rank adaptations and fixed-strength transformations, reducing their adaptation expressive power. In this work, we propose Decoupled Low-rank Adaptation (DeLoRA), a novel finetuning method that normalizes and scales learnable low-rank matrices. By bounding the distance of the transformation, DeLoRA effectively decouples the angular learning from the adaptation strength, enhancing robustness without compromising performance. Through evaluations on subject-driven image generation, natural language understanding, and instruction tuning, we show that DeLoRA matches or surpasses performance of competing PEFT methods, while exhibiting stronger robustness. Code is available at https://github.com/ExplainableML/DeLoRA.
comment: ICLR 2025
☆ MammAlps: A multi-view video behavior monitoring dataset of wild mammals in the Swiss Alps CVPR 2025
Monitoring wildlife is essential for ecology and ethology, especially in light of the increasing human impact on ecosystems. Camera traps have emerged as habitat-centric sensors enabling the study of wildlife populations at scale with minimal disturbance. However, the lack of annotated video datasets limits the development of powerful video understanding models needed to process the vast amount of fieldwork data collected. To advance research in wild animal behavior monitoring we present MammAlps, a multimodal and multi-view dataset of wildlife behavior monitoring from 9 camera-traps in the Swiss National Park. MammAlps contains over 14 hours of video with audio, 2D segmentation maps and 8.5 hours of individual tracks densely labeled for species and behavior. Based on 6135 single animal clips, we propose the first hierarchical and multimodal animal behavior recognition benchmark using audio, video and reference scene segmentation maps as inputs. Furthermore, we also propose a second ecology-oriented benchmark aiming at identifying activities, species, number of individuals and meteorological conditions from 397 multi-view and long-term ecological events, including false positive triggers. We advocate that both tasks are complementary and contribute to bridging the gap between machine learning and ecology. Code and data are available at: https://github.com/eceo-epfl/MammAlps
comment: CVPR 2025; Benchmark and code at: https://github.com/eceo-epfl/MammAlps
☆ SimMotionEdit: Text-Based Human Motion Editing with Motion Similarity Prediction
Text-based 3D human motion editing is a critical yet challenging task in computer vision and graphics. While training-free approaches have been explored, the recent release of the MotionFix dataset, which includes source-text-motion triplets, has opened new avenues for training, yielding promising results. However, existing methods struggle with precise control, often leading to misalignment between motion semantics and language instructions. In this paper, we introduce a related task, motion similarity prediction, and propose a multi-task training paradigm, where we train the model jointly on motion editing and motion similarity prediction to foster the learning of semantically meaningful representations. To complement this task, we design an advanced Diffusion-Transformer-based architecture that separately handles motion similarity prediction and motion editing. Extensive experiments demonstrate the state-of-the-art performance of our approach in both editing alignment and fidelity.
comment: Project URL: https://github.com/lzhyu/SimMotionEdit
☆ Training A Neural Network For Partially Occluded Road Sign Identification In The Context Of Autonomous Vehicles
The increasing number of autonomous vehicles and the rapid development of computer vision technologies underscore the particular importance of conducting research on the accuracy of traffic sign recognition. Numerous studies in this field have already achieved significant results, demonstrating high effectiveness in addressing traffic sign recognition tasks. However, the task becomes considerably more complex when a sign is partially obscured by surrounding objects, such as tree branches, billboards, or other elements of the urban environment. In our study, we investigated how partial occlusion of traffic signs affects their recognition. For this purpose, we collected a dataset comprising 5,746 images, including both fully visible and partially occluded signs, and made it publicly available. Using this dataset, we compared the performance of our custom convolutional neural network (CNN), which achieved 96% accuracy, with models trained using transfer learning. The best result was obtained by VGG16 with full layer unfreezing, reaching 99% accuracy. Additional experiments revealed that models trained solely on fully visible signs lose effectiveness when recognizing occluded signs. This highlights the critical importance of incorporating real-world data with partial occlusion into training sets to ensure robust model performance in complex practical scenarios and to enhance the safety of autonomous driving.
☆ Self-Attention Diffusion Models for Zero-Shot Biomedical Image Segmentation: Unlocking New Frontiers in Medical Imaging
Producing high-quality segmentation masks for medical images is a fundamental challenge in biomedical image analysis. Recent research has explored large-scale supervised training to enable segmentation across various medical imaging modalities and unsupervised training to facilitate segmentation without dense annotations. However, constructing a model capable of segmenting diverse medical images in a zero-shot manner without any annotations remains a significant hurdle. This paper introduces the Attention Diffusion Zero-shot Unsupervised System (ADZUS), a novel approach that leverages self-attention diffusion models for zero-shot biomedical image segmentation. ADZUS harnesses the intrinsic capabilities of pre-trained diffusion models, utilizing their generative and discriminative potentials to segment medical images without requiring annotated training data or prior domain-specific knowledge. The ADZUS architecture is detailed, with its integration of self-attention mechanisms that facilitate context-aware and detail-sensitive segmentations being highlighted. Experimental results across various medical imaging datasets, including skin lesion segmentation, chest X-ray infection segmentation, and white blood cell segmentation, reveal that ADZUS achieves state-of-the-art performance. Notably, ADZUS reached Dice scores ranging from 88.7\% to 92.9\% and IoU scores from 66.3\% to 93.3\% across different segmentation tasks, demonstrating significant improvements in handling novel, unseen medical imagery. It is noteworthy that while ADZUS demonstrates high effectiveness, it demands substantial computational resources and extended processing times. The model's efficacy in zero-shot settings underscores its potential to reduce reliance on costly annotations and seamlessly adapt to new medical imaging tasks, thereby expanding the diagnostic capabilities of AI-driven medical imaging technologies.
comment: 15 pages, 5 figures
☆ SNRAware: Improved Deep Learning MRI Denoising with SNR Unit Training and G-factor Map Augmentation
To develop and evaluate a new deep learning MR denoising method that leverages quantitative noise distribution information from the reconstruction process to improve denoising performance and generalization. This retrospective study trained 14 different transformer and convolutional models with two backbone architectures on a large dataset of 2,885,236 images from 96,605 cardiac retro-gated cine complex series acquired at 3T. The proposed training scheme, termed SNRAware, leverages knowledge of the MRI reconstruction process to improve denoising performance by simulating large, high quality, and diverse synthetic datasets, and providing quantitative information about the noise distribution to the model. In-distribution testing was performed on a hold-out dataset of 3000 samples with performance measured using PSNR and SSIM, with ablation comparison without the noise augmentation. Out-of-distribution tests were conducted on cardiac real-time cine, first-pass cardiac perfusion, and neuro and spine MRI, all acquired at 1.5T, to test model generalization across imaging sequences, dynamically changing contrast, different anatomies, and field strengths. The best model found in the in-distribution test generalized well to out-of-distribution samples, delivering 6.5x and 2.9x CNR improvement for real-time cine and perfusion imaging, respectively. Further, a model trained with 100% cardiac cine data generalized well to a T1 MPRAGE neuro 3D scan and T2 TSE spine MRI.
☆ MAO: Efficient Model-Agnostic Optimization of Prompt Tuning for Vision-Language Models ICME 2025
Though CLIP-based prompt tuning significantly enhances pre-trained Vision-Language Models, existing research focuses on reconstructing the model architecture, e.g., additional loss calculation and meta-networks. These approaches generally lead to increased complexity and extended training cost. To maintain the efficiency of the tuning process, we propose plug-and-play Model-Agnostic Optimization (MAO) for prompt tuning. Without altering any components of the prompt tuning backbone, we introduce a Data-Driven Enhancement framework to optimize the distribution of the initial data, and incorporate an Alterable Regularization module to boost the task-specific feature processing pipeline, thereby improving overall performance while maintaining low computational cost. Extensive experiments on MAO demonstrate its outstanding performance and efficiency. The code of MAO is available at: https://github.com/JREion/M.A.O .
comment: Accepted by the IEEE International Conference on Multimedia & Expo 2025 (ICME 2025); 12 pages, 6 figures, 8 tables
DiffusionTalker: Efficient and Compact Speech-Driven 3D Talking Head via Personalizer-Guided Distillation ICME2025
Real-time speech-driven 3D facial animation has been attractive in academia and industry. Traditional methods mainly focus on learning a deterministic mapping from speech to animation. Recent approaches start to consider the nondeterministic fact of speech-driven 3D face animation and employ the diffusion model for the task. Existing diffusion-based methods can improve the diversity of facial animation. However, personalized speaking styles conveying accurate lip language is still lacking, besides, efficiency and compactness still need to be improved. In this work, we propose DiffusionTalker to address the above limitations via personalizer-guided distillation. In terms of personalization, we introduce a contrastive personalizer that learns identity and emotion embeddings to capture speaking styles from audio. We further propose a personalizer enhancer during distillation to enhance the influence of embeddings on facial animation. For efficiency, we use iterative distillation to reduce the steps required for animation generation and achieve more than 8x speedup in inference. To achieve compactness, we distill the large teacher model into a smaller student model, reducing our model's storage by 86.4\% while minimizing performance loss. After distillation, users can derive their identity and emotion embeddings from audio to quickly create personalized animations that reflect specific speaking styles. Extensive experiments are conducted to demonstrate that our method outperforms state-of-the-art methods. The code will be released at: https://github.com/ChenVoid/DiffusionTalker.
comment: Accepted by ICME2025
☆ Decorum: A Language-Based Approach For Style-Conditioned Synthesis of Indoor 3D Scenes
3D indoor scene generation is an important problem for the design of digital and real-world environments. To automate this process, a scene generation model should be able to not only generate plausible scene layouts, but also take into consideration visual features and style preferences. Existing methods for this task exhibit very limited control over these attributes, only allowing text inputs in the form of simple object-level descriptions or pairwise spatial relationships. Our proposed method Decorum enables users to control the scene generation process with natural language by adopting language-based representations at each stage. This enables us to harness recent advancements in Large Language Models (LLMs) to model language-to-language mappings. In addition, we show that using a text-based representation allows us to select furniture for our scenes using a novel object retrieval method based on multimodal LLMs. Evaluations on the benchmark 3D-FRONT dataset show that our methods achieve improvements over existing work in text-conditioned scene synthesis and object retrieval.
☆ Efficient Deep Learning Approaches for Processing Ultra-Widefield Retinal Imaging
Deep learning has emerged as the predominant solution for classifying medical images. We intend to apply these developments to the ultra-widefield (UWF) retinal imaging dataset. Since UWF images can accurately diagnose various retina diseases, it is very important to clas sify them accurately and prevent them with early treatment. However, processing images manually is time-consuming and labor-intensive, and there are two challenges to automating this process. First, high perfor mance usually requires high computational resources. Artificial intelli gence medical technology is better suited for places with limited medical resources, but using high-performance processing units in such environ ments is challenging. Second, the problem of the accuracy of colour fun dus photography (CFP) methods. In general, the UWF method provides more information for retinal diagnosis than the CFP method, but most of the research has been conducted based on the CFP method. Thus, we demonstrate that these problems can be efficiently addressed in low performance units using methods such as strategic data augmentation and model ensembles, which balance performance and computational re sources while utilizing UWF images.
☆ LongDiff: Training-Free Long Video Generation in One Go
Video diffusion models have recently achieved remarkable results in video generation. Despite their encouraging performance, most of these models are mainly designed and trained for short video generation, leading to challenges in maintaining temporal consistency and visual details in long video generation. In this paper, we propose LongDiff, a novel training-free method consisting of carefully designed components \ -- Position Mapping (PM) and Informative Frame Selection (IFS) \ -- to tackle two key challenges that hinder short-to-long video generation generalization: temporal position ambiguity and information dilution. Our LongDiff unlocks the potential of off-the-shelf video diffusion models to achieve high-quality long video generation in one go. Extensive experiments demonstrate the efficacy of our method.
☆ PHT-CAD: Efficient CAD Parametric Primitive Analysis with Progressive Hierarchical Tuning
Computer-Aided Design (CAD) plays a pivotal role in industrial manufacturing, yet 2D Parametric Primitive Analysis (PPA) remains underexplored due to two key challenges: structural constraint reasoning and advanced semantic understanding. To tackle these challenges, we first propose an Efficient Hybrid Parametrization (EHP) for better representing 2D engineering drawings. EHP contains four types of atomic component i.e., point, line, circle, and arc). Additionally, we propose PHT-CAD, a novel 2D PPA framework that harnesses the modality alignment and reasoning capabilities of Vision-Language Models (VLMs) for precise engineering drawing analysis. In PHT-CAD, we introduce four dedicated regression heads to predict corresponding atomic components. To train PHT-CAD, a three-stage training paradigm Progressive Hierarchical Tuning (PHT) is proposed to progressively enhance PHT-CAD's capability to perceive individual primitives, infer structural constraints, and align annotation layers with their corresponding geometric representations. Considering that existing datasets lack complete annotation layers and real-world engineering drawings, we introduce ParaCAD, the first large-scale benchmark that explicitly integrates both the geometric and annotation layers. ParaCAD comprises over 10 million annotated drawings for training and 3,000 real-world industrial drawings with complex topological structures and physical constraints for test. Extensive experiments demonstrate the effectiveness of PHT-CAD and highlight the practical significance of ParaCAD in advancing 2D PPA research.
☆ LocDiffusion: Identifying Locations on Earth by Diffusing in the Hilbert Space
Image geolocalization is a fundamental yet challenging task, aiming at inferring the geolocation on Earth where an image is taken. Existing methods approach it either via grid-based classification or via image retrieval. Their performance significantly suffers when the spatial distribution of test images does not align with such choices. To address these limitations, we propose to leverage diffusion as a mechanism for image geolocalization. To avoid the problematic manifold reprojection step in diffusion, we developed a novel spherical positional encoding-decoding framework, which encodes points on a spherical surface (e.g., geolocations on Earth) into a Hilbert space of Spherical Harmonics coefficients and decodes points (geolocations) by mode-seeking. We call this type of position encoding Spherical Harmonics Dirac Delta (SHDD) Representation. We also propose a novel SirenNet-based architecture called CS-UNet to learn the conditional backward process in the latent SHDD space by minimizing a latent KL-divergence loss. We train a conditional latent diffusion model called LocDiffusion that generates geolocations under the guidance of images -- to the best of our knowledge, the first generative model for image geolocalization by diffusing geolocation information in a hidden location embedding space. We evaluate our method against SOTA image geolocalization baselines. LocDiffusion achieves competitive geolocalization performance and demonstrates significantly stronger generalizability to unseen geolocations.
☆ AGIR: Assessing 3D Gait Impairment with Reasoning based on LLMs
Assessing gait impairment plays an important role in early diagnosis, disease monitoring, and treatment evaluation for neurodegenerative diseases. Despite its widespread use in clinical practice, it is limited by subjectivity and a lack of precision. While recent deep learning-based approaches have consistently improved classification accuracies, they often lack interpretability, hindering their utility in clinical decision-making. To overcome these challenges, we introduce AGIR, a novel pipeline consisting of a pre-trained VQ-VAE motion tokenizer and a subsequent Large Language Model (LLM) fine-tuned over pairs of motion tokens and Chain-of-Thought (CoT) reasonings. To fine-tune an LLM for pathological gait analysis, we first introduce a multimodal dataset by adding rationales dedicated to MDS-UPDRS gait score assessment to an existing PD gait dataset. We then introduce a two-stage supervised fine-tuning (SFT) strategy to enhance the LLM's motion comprehension with pathology-specific knowledge. This strategy includes: 1) a generative stage that aligns gait motions with analytic descriptions through bidirectional motion-description generation, 2) a reasoning stage that integrates logical Chain-of-Thought (CoT) reasoning for impairment assessment with UPDRS gait score. Validation on an existing dataset and comparisons with state-of-the-art methods confirm the robustness and accuracy of our pipeline, demonstrating its ability to assign gait impairment scores from motion input with clinically meaningful rationales.
☆ TCFG: Tangential Damping Classifier-free Guidance CVPR2025
Diffusion models have achieved remarkable success in text-to-image synthesis, largely attributed to the use of classifier-free guidance (CFG), which enables high-quality, condition-aligned image generation. CFG combines the conditional score (e.g., text-conditioned) with the unconditional score to control the output. However, the unconditional score is in charge of estimating the transition between manifolds of adjacent timesteps from $x_t$ to $x_{t-1}$, which may inadvertently interfere with the trajectory toward the specific condition. In this work, we introduce a novel approach that leverages a geometric perspective on the unconditional score to enhance CFG performance when conditional scores are available. Specifically, we propose a method that filters the singular vectors of both conditional and unconditional scores using singular value decomposition. This filtering process aligns the unconditional score with the conditional score, thereby refining the sampling trajectory to stay closer to the manifold. Our approach improves image quality with negligible additional computation. We provide deeper insights into the score function behavior in diffusion models and present a practical technique for achieving more accurate and contextually coherent image synthesis.
comment: CVPR2025
☆ MLLM-For3D: Adapting Multimodal Large Language Model for 3D Reasoning Segmentation
Reasoning segmentation aims to segment target objects in complex scenes based on human intent and spatial reasoning. While recent multimodal large language models (MLLMs) have demonstrated impressive 2D image reasoning segmentation, adapting these capabilities to 3D scenes remains underexplored. In this paper, we introduce MLLM-For3D, a simple yet effective framework that transfers knowledge from 2D MLLMs to 3D scene understanding. Specifically, we utilize MLLMs to generate multi-view pseudo segmentation masks and corresponding text embeddings, then unproject 2D masks into 3D space and align them with the text embeddings. The primary challenge lies in the absence of 3D context and spatial consistency across multiple views, causing the model to hallucinate objects that do not exist and fail to target objects consistently. Training the 3D model with such irrelevant objects leads to performance degradation. To address this, we introduce a spatial consistency strategy to enforce that segmentation masks remain coherent in the 3D space, effectively capturing the geometry of the scene. Moreover, we develop a Token-for-Query approach for multimodal semantic alignment, enabling consistent identification of the same object across different views. Extensive evaluations on various challenging indoor scene benchmarks demonstrate that, even without any labeled 3D training data, MLLM-For3D outperforms existing 3D reasoning segmentation methods, effectively interpreting user intent, understanding 3D scenes, and reasoning about spatial relationships.
☆ An Image-like Diffusion Method for Human-Object Interaction Detection CVPR 2025
Human-object interaction (HOI) detection often faces high levels of ambiguity and indeterminacy, as the same interaction can appear vastly different across different human-object pairs. Additionally, the indeterminacy can be further exacerbated by issues such as occlusions and cluttered backgrounds. To handle such a challenging task, in this work, we begin with a key observation: the output of HOI detection for each human-object pair can be recast as an image. Thus, inspired by the strong image generation capabilities of image diffusion models, we propose a new framework, HOI-IDiff. In HOI-IDiff, we tackle HOI detection from a novel perspective, using an Image-like Diffusion process to generate HOI detection outputs as images. Furthermore, recognizing that our recast images differ in certain properties from natural images, we enhance our framework with a customized HOI diffusion process and a slice patchification model architecture, which are specifically tailored to generate our recast ``HOI images''. Extensive experiments demonstrate the efficacy of our framework.
comment: CVPR 2025
☆ End-to-End Implicit Neural Representations for Classification CVPR 2025
Implicit neural representations (INRs) such as NeRF and SIREN encode a signal in neural network parameters and show excellent results for signal reconstruction. Using INRs for downstream tasks, such as classification, is however not straightforward. Inherent symmetries in the parameters pose challenges and current works primarily focus on designing architectures that are equivariant to these symmetries. However, INR-based classification still significantly under-performs compared to pixel-based methods like CNNs. This work presents an end-to-end strategy for initializing SIRENs together with a learned learning-rate scheme, to yield representations that improve classification accuracy. We show that a simple, straightforward, Transformer model applied to a meta-learned SIREN, without incorporating explicit symmetry equivariances, outperforms the current state-of-the-art. On the CIFAR-10 SIREN classification task, we improve the state-of-the-art without augmentations from 38.8% to 59.6%, and from 63.4% to 64.7% with augmentations. We demonstrate scalability on the high-resolution Imagenette dataset achieving reasonable reconstruction quality with a classification accuracy of 60.8% and are the first to do INR classification on the full ImageNet-1K dataset where we achieve a SIREN classification performance of 23.6%. To the best of our knowledge, no other SIREN classification approach has managed to set a classification baseline for any high-resolution image dataset. Our code is available at https://github.com/SanderGielisse/MWT
comment: Accepted to CVPR 2025. 8 pages, supplementary material included
☆ Unraveling the Effects of Synthetic Data on End-to-End Autonomous Driving
End-to-end (E2E) autonomous driving (AD) models require diverse, high-quality data to perform well across various driving scenarios. However, collecting large-scale real-world data is expensive and time-consuming, making high-fidelity synthetic data essential for enhancing data diversity and model robustness. Existing driving simulators for synthetic data generation have significant limitations: game-engine-based simulators struggle to produce realistic sensor data, while NeRF-based and diffusion-based methods face efficiency challenges. Additionally, recent simulators designed for closed-loop evaluation provide limited interaction with other vehicles, failing to simulate complex real-world traffic dynamics. To address these issues, we introduce SceneCrafter, a realistic, interactive, and efficient AD simulator based on 3D Gaussian Splatting (3DGS). SceneCrafter not only efficiently generates realistic driving logs across diverse traffic scenarios but also enables robust closed-loop evaluation of end-to-end models. Experimental results demonstrate that SceneCrafter serves as both a reliable evaluation platform and a efficient data generator that significantly improves end-to-end model generalization.
☆ PanoGS: Gaussian-based Panoptic Segmentation for 3D Open Vocabulary Scene Understanding CVPR 2025
Recently, 3D Gaussian Splatting (3DGS) has shown encouraging performance for open vocabulary scene understanding tasks. However, previous methods cannot distinguish 3D instance-level information, which usually predicts a heatmap between the scene feature and text query. In this paper, we propose PanoGS, a novel and effective 3D panoptic open vocabulary scene understanding approach. Technically, to learn accurate 3D language features that can scale to large indoor scenarios, we adopt the pyramid tri-plane to model the latent continuous parametric feature space and use a 3D feature decoder to regress the multi-view fused 2D feature cloud. Besides, we propose language-guided graph cuts that synergistically leverage reconstructed geometry and learned language cues to group 3D Gaussian primitives into a set of super-primitives. To obtain 3D consistent instance, we perform graph clustering based segmentation with SAM-guided edge affinity computation between different super-primitives. Extensive experiments on widely used datasets show better or more competitive performance on 3D panoptic open vocabulary scene understanding. Project page: \href{https://zju3dv.github.io/panogs}{https://zju3dv.github.io/panogs}.
comment: CVPR 2025
☆ M3Net: Multimodal Multi-task Learning for 3D Detection, Segmentation, and Occupancy Prediction in Autonomous Driving AAAI 2025
The perception system for autonomous driving generally requires to handle multiple diverse sub-tasks. However, current algorithms typically tackle individual sub-tasks separately, which leads to low efficiency when aiming at obtaining full-perception results. Some multi-task learning methods try to unify multiple tasks with one model, but do not solve the conflicts in multi-task learning. In this paper, we introduce M3Net, a novel multimodal and multi-task network that simultaneously tackles detection, segmentation, and 3D occupancy prediction for autonomous driving and achieves superior performance than single task model. M3Net takes multimodal data as input and multiple tasks via query-token interactions. To enhance the integration of multi-modal features for multi-task learning, we first propose the Modality-Adaptive Feature Integration (MAFI) module, which enables single-modality features to predict channel-wise attention weights for their high-performing tasks, respectively. Based on integrated features, we then develop task-specific query initialization strategies to accommodate the needs of detection/segmentation and 3D occupancy prediction. Leveraging the properly initialized queries, a shared decoder transforms queries and BEV features layer-wise, facilitating multi-task learning. Furthermore, we propose a Task-oriented Channel Scaling (TCS) module in the decoder to mitigate conflicts between optimizing for different tasks. Additionally, our proposed multi-task querying and TCS module support both Transformer-based decoder and Mamba-based decoder, demonstrating its flexibility to different architectures. M3Net achieves state-of-the-art multi-task learning performance on the nuScenes benchmarks.
comment: Accepted by AAAI 2025
☆ Anomize: Better Open Vocabulary Video Anomaly Detection
Open Vocabulary Video Anomaly Detection (OVVAD) seeks to detect and classify both base and novel anomalies. However, existing methods face two specific challenges related to novel anomalies. The first challenge is detection ambiguity, where the model struggles to assign accurate anomaly scores to unfamiliar anomalies. The second challenge is categorization confusion, where novel anomalies are often misclassified as visually similar base instances. To address these challenges, we explore supplementary information from multiple sources to mitigate detection ambiguity by leveraging multiple levels of visual data alongside matching textual information. Furthermore, we propose incorporating label relations to guide the encoding of new labels, thereby improving alignment between novel videos and their corresponding labels, which helps reduce categorization confusion. The resulting Anomize framework effectively tackles these issues, achieving superior performance on UCF-Crime and XD-Violence datasets, demonstrating its effectiveness in OVVAD.
♻ ☆ Multi-focal Conditioned Latent Diffusion for Person Image Synthesis CVPR 2025
The Latent Diffusion Model (LDM) has demonstrated strong capabilities in high-resolution image generation and has been widely employed for Pose-Guided Person Image Synthesis (PGPIS), yielding promising results. However, the compression process of LDM often results in the deterioration of details, particularly in sensitive areas such as facial features and clothing textures. In this paper, we propose a Multi-focal Conditioned Latent Diffusion (MCLD) method to address these limitations by conditioning the model on disentangled, pose-invariant features from these sensitive regions. Our approach utilizes a multi-focal condition aggregation module, which effectively integrates facial identity and texture-specific information, enhancing the model's ability to produce appearance realistic and identity-consistent images. Our method demonstrates consistent identity and appearance generation on the DeepFashion dataset and enables flexible person image editing due to its generation consistency. The code is available at https://github.com/jqliu09/mcld.
comment: CVPR 2025 Accepted
♻ ☆ Fus-MAE: A cross-attention-based data fusion approach for Masked Autoencoders in remote sensing
Self-supervised frameworks for representation learning have recently stirred up interest among the remote sensing community, given their potential to mitigate the high labeling costs associated with curating large satellite image datasets. In the realm of multimodal data fusion, while the often used contrastive learning methods can help bridging the domain gap between different sensor types, they rely on data augmentations techniques that require expertise and careful design, especially for multispectral remote sensing data. A possible but rather scarcely studied way to circumvent these limitations is to use a masked image modelling based pretraining strategy. In this paper, we introduce Fus-MAE, a self-supervised learning framework based on masked autoencoders that uses cross-attention to perform early and feature-level data fusion between synthetic aperture radar and multispectral optical data - two modalities with a significant domain gap. Our empirical findings demonstrate that Fus-MAE can effectively compete with contrastive learning strategies tailored for SAR-optical data fusion and outperforms other masked-autoencoders frameworks trained on a larger corpus.
♻ ☆ PanSplat: 4K Panorama Synthesis with Feed-Forward Gaussian Splatting CVPR2025
With the advent of portable 360{\deg} cameras, panorama has gained significant attention in applications like virtual reality (VR), virtual tours, robotics, and autonomous driving. As a result, wide-baseline panorama view synthesis has emerged as a vital task, where high resolution, fast inference, and memory efficiency are essential. Nevertheless, existing methods are typically constrained to lower resolutions (512 $\times$ 1024) due to demanding memory and computational requirements. In this paper, we present PanSplat, a generalizable, feed-forward approach that efficiently supports resolution up to 4K (2048 $\times$ 4096). Our approach features a tailored spherical 3D Gaussian pyramid with a Fibonacci lattice arrangement, enhancing image quality while reducing information redundancy. To accommodate the demands of high resolution, we propose a pipeline that integrates a hierarchical spherical cost volume and Gaussian heads with local operations, enabling two-step deferred backpropagation for memory-efficient training on a single A100 GPU. Experiments demonstrate that PanSplat achieves state-of-the-art results with superior efficiency and image quality across both synthetic and real-world datasets. Code is available at https://github.com/chengzhag/PanSplat.
comment: Camera Ready of CVPR2025. Project Page: https://chengzhag.github.io/publication/pansplat/ Code: https://github.com/chengzhag/PanSplat
♻ ☆ Computation-Efficient and Recognition-Friendly 3D Point Cloud Privacy Protection CVPR2025
3D point cloud has been widely used in applications such as self-driving cars, robotics, CAD models, etc. To the best of our knowledge, these applications raised the issue of privacy leakage in 3D point clouds, which has not been studied well. Different from the 2D image privacy, which is related to texture and 2D geometric structure, the 3D point cloud is texture-less and only relevant to 3D geometric structure. In this work, we defined the 3D point cloud privacy problem and proposed an efficient privacy-preserving framework named PointFlowGMM that can support downstream classification and segmentation tasks without seeing the original data. Using a flow-based generative model, the point cloud is projected into a latent Gaussian mixture distributed subspace. We further designed a novel angular similarity loss to obfuscate the original geometric structure and reduce the model size from 767MB to 120MB without a decrease in recognition performance. The projected point cloud in the latent space is orthogonally rotated randomly to further protect the original geometric structure, the class-to-class relationship is preserved after rotation, thus, the protected point cloud can support the recognition task. We evaluated our model on multiple datasets and achieved comparable recognition results on encrypted point clouds compared to the original point clouds.
comment: Accepted by CVPR2025
♻ ☆ Principles of Visual Tokens for Efficient Video Understanding
Video understanding has made huge strides in recent years, relying largely on the power of transformers. As this architecture is notoriously expensive and video data is highly redundant, research into improving efficiency has become particularly relevant. Some creative solutions include token selection and merging. While most methods succeed in reducing the cost of the model and maintaining accuracy, an interesting pattern arises: most methods do not outperform the baseline of randomly discarding tokens. In this paper we take a closer look at this phenomenon and observe 5 principles of the nature of visual tokens. For example, we observe that the value of tokens follows a clear Pareto-distribution where most tokens have remarkably low value, and just a few carry most of the perceptual information. We build on these and further insights to propose a lightweight video model, LITE, that can select a small number of tokens effectively, outperforming state-of-the-art and existing baselines across datasets (Kinetics-400 and Something-Something-V2) in the challenging trade-off of computation (GFLOPs) vs accuracy. Experiments also show that LITE generalizes across datasets and even other tasks without the need for retraining.
♻ ☆ A Siamese Network to Detect If Two Iris Images Are Monozygotic
In Daugman-style iris recognition, the textures of the left and right irises of the same person are traditionally considered as being as different as the irises of two unrelated persons. However, previous research indicates that humans can detect that two iris images are from different eyes of the same person, or eyes of monozygotic twins, with an accuracy of about 80%. In this work, we employ a Siamese network architecture and contrastive learning to categorize a pair of iris images as coming from monozygotic or non-monozygotic irises. This could potentially be applied, for example, as a fast, noninvasive test to determine if twins are monozygotic or non-monozygotic. We construct a dataset comprising both synthetic monozygotic pairs (images of different irises of the same individual) and natural monozygotic pairs (images of different images from persons who are identical twins), in addition to non-monozygotic pairs from unrelated individuals, ensuring a comprehensive evaluation of the model's capabilities. To gain deeper insights into the learned representations, we train and analyze three variants of the model using (1) the original input images, (2) iris-only images, and (3) non-iris-only images. This comparison reveals the critical importance of iris-specific textural details and contextual ocular cues in identifying monozygotic iris patterns. The results demonstrate that models leveraging full eye-region information outperform those trained solely on iris-only data, emphasizing the nuanced interplay between iris and ocular characteristics. Our approach achieves accuracy levels using the full iris image that exceed those previously reported for human classification of monozygotic iris pairs. This study presents the first classifier designed to determine whether a pair of iris images originates from monozygotic individuals.
♻ ☆ Colonoscopy polyp detection with massive endoscopic images
We improved an existing end-to-end polyp detection model with better average precision validated by different data sets with trivial cost on detection speed. Our previous work on detecting polyps within colonoscopy provided an efficient end-to-end solution to alleviate doctor's examination overhead. However, our later experiments found this framework is not as robust as before as the condition of polyp capturing varies. In this work, we conducted several studies on data set, identifying main issues that causes low precision rate in the task of polyp detection. We used an optimized anchor generation methods to get better anchor box shape and more boxes are used for detection as we believe this is necessary for small object detection. An alternative backbone is used to compensate the heavy time cost introduced by dense anchor box regression. With use of the attention gate module, our model can achieve state-of-the-art polyp detection performance while still maintain real-time detection speed.
comment: 13 pages, 10 figures
♻ ☆ Confidence Intervals for Performance Estimates in Brain MRI Segmentation
Medical segmentation models are evaluated empirically. As such an evaluation is based on a limited set of example images, it is unavoidably noisy. Beyond a mean performance measure, reporting confidence intervals is thus crucial. However, this is rarely done in medical image segmentation. The width of the confidence interval depends on the test set size and on the spread of the performance measure (its standard-deviation across the test set). For classification, many test images are needed to avoid wide confidence intervals. Segmentation, however, has not been studied, and it differs by the amount of information brought by a given test image. In this paper, we study the typical confidence intervals in the context of segmentation in 3D brain magnetic resonance imaging (MRI). We carry experiments on using the standard nnU-net framework, two datasets from the Medical Decathlon challenge that concern brain MRI (hippocampus and brain tumor segmentation) and two performance measures: the Dice Similarity Coefficient and the Hausdorff distance. We show that the parametric confidence intervals are reasonable approximations of the bootstrap estimates for varying test set sizes and spread of the performance metric. Importantly, we show that the test size needed to achieve a given precision is often much lower than for classification tasks. Typically, a 1\% wide confidence interval requires about 100-200 test samples when the spread is low (standard-deviation around 3\%). More difficult segmentation tasks may lead to higher spreads and require over 1000 samples.
♻ ☆ EventHallusion: Diagnosing Event Hallucinations in Video LLMs
Recently, Multimodal Large Language Models (MLLMs) have made significant progress in the video comprehension field. Despite remarkable content reasoning and instruction following capabilities they demonstrated, the hallucination problem of these VideoLLMs is less explored compared with its counterpart in the image domain. To mitigate this gap, we propose EventHallusion, a novel benchmark that focuses on assessing the VideoLLMs' hallucination toward event, the crux of video analysis. From a hallucination attribution perspective, our EventHallusion benchmark is curated to assess a VideoLLM's susceptibility toward language priors and vision-language biases. On the other hand, we also propose a simple yet effective method, called Temporal Contrastive Decoding (TCD), to tackle the hallucination problems of VideoLLMs. The proposed TCD method rectifies the model's bias toward its priors during the decoding stage by comparing the original video with a modified version, in which temporal cues are disrupted. Through comprehensive evaluation of eight open-source and two closed-source VideoLLMs on the proposed EventHallusion benchmark, we observe that the open-source models suffer significantly from hallucination problems, whereas the closed-source ones perform markedly better. By further equipping open-source VideoLLMs with the proposed TCD approach, evident performance improvements are achieved across most metrics in the EventHallusion benchmark. Our codes and benchmark data are available at https://github.com/Stevetich/EventHallusion.
♻ ☆ COVID 19 Diagnosis Analysis using Transfer Learning
Coronaviruses, including SARS-CoV-2, are responsible for COVID-19, a highly transmissible disease that emerged in December 2019 in Wuhan, China. During the past five years, significant advancements have been made in understanding and mitigating the virus. Although the initial outbreak led to global health crises, improved vaccination strategies, antiviral treatments, and AI-driven diagnostic tools have contributed to better disease management. However, COVID-19 continues to pose risks, particularly for immuno-compromised individuals and those with pre-existing conditions. This study explores the use of deep learning for a rapid and accurate diagnosis of COVID-19, addressing ongoing challenges in healthcare infrastructure and testing accessibility. We propose an enhanced automated detection system leveraging state-of-the-art convolutional neural networks (CNNs), including updated versions of VGG16, VGG19, and ResNet50, to classify COVID-19 infections from chest radiographs and computerized tomography (CT) scans. Our results, based on an expanded dataset of over 6000 medical images, demonstrate that the optimized ResNet50 model achieves the highest classification performance, with 97.77% accuracy, 100% sensitivity, 93.33% specificity, and a 98.0% F1-score. These findings reinforce the potential of AI-assisted diagnostic tools in improving early detection and pandemic preparedness.
♻ ☆ Explaining the Impact of Training on Vision Models via Activation Clustering
This paper introduces Neuro-Activated Vision Explanations (NAVE), a method for extracting and visualizing the internal representations of vision model encoders. By clustering feature activations, NAVE provides insights into learned semantics without fine-tuning. Using object localization, we show that NAVE's concepts align with image semantics. Through extensive experiments, we analyze the impact of training strategies and architectures on encoder representation capabilities. Additionally, we apply NAVE to study training artifacts in vision transformers and reveal how weak training strategies and spurious correlations degrade model performance. Our findings establish NAVE as a valuable tool for post-hoc model inspection and improving transparency in vision models.
♻ ☆ TexGaussian: Generating High-quality PBR Material via Octree-based 3D Gaussian Splatting CVPR 2025
Physically Based Rendering (PBR) materials play a crucial role in modern graphics, enabling photorealistic rendering across diverse environment maps. Developing an effective and efficient algorithm that is capable of automatically generating high-quality PBR materials rather than RGB texture for 3D meshes can significantly streamline the 3D content creation. Most existing methods leverage pre-trained 2D diffusion models for multi-view image synthesis, which often leads to severe inconsistency between the generated textures and input 3D meshes. This paper presents TexGaussian, a novel method that uses octant-aligned 3D Gaussian Splatting for rapid PBR material generation. Specifically, we place each 3D Gaussian on the finest leaf node of the octree built from the input 3D mesh to render the multi-view images not only for the albedo map but also for roughness and metallic. Moreover, our model is trained in a regression manner instead of diffusion denoising, capable of generating the PBR material for a 3D mesh in a single feed-forward process. Extensive experiments on publicly available benchmarks demonstrate that our method synthesizes more visually pleasing PBR materials and runs faster than previous methods in both unconditional and text-conditional scenarios, exhibiting better consistency with the given geometry. Our code and trained models are available at https://3d-aigc.github.io/TexGaussian.
comment: CVPR 2025. Project Page: https://3d-aigc.github.io/TexGaussian
♻ ☆ SLAM3R: Real-Time Dense Scene Reconstruction from Monocular RGB Videos CVPR 2025
In this paper, we introduce SLAM3R, a novel and effective system for real-time, high-quality, dense 3D reconstruction using RGB videos. SLAM3R provides an end-to-end solution by seamlessly integrating local 3D reconstruction and global coordinate registration through feed-forward neural networks. Given an input video, the system first converts it into overlapping clips using a sliding window mechanism. Unlike traditional pose optimization-based methods, SLAM3R directly regresses 3D pointmaps from RGB images in each window and progressively aligns and deforms these local pointmaps to create a globally consistent scene reconstruction - all without explicitly solving any camera parameters. Experiments across datasets consistently show that SLAM3R achieves state-of-the-art reconstruction accuracy and completeness while maintaining real-time performance at 20+ FPS. Code available at: https://github.com/PKU-VCL-3DV/SLAM3R.
comment: CVPR 2025
♻ ☆ Teaching Large Language Models to Regress Accurate Image Quality Scores using Score Distribution CVPR 2025
With the rapid advancement of Multi-modal Large Language Models (MLLMs), MLLM-based Image Quality Assessment (IQA) methods have shown promising performance in linguistic quality description. However, current methods still fall short in accurately scoring image quality. In this work, we aim to leverage MLLMs to regress accurate quality scores. A key challenge is that the quality score is inherently continuous, typically modeled as a Gaussian distribution, whereas MLLMs generate discrete token outputs. This mismatch necessitates score discretization. Previous approaches discretize the mean score into a one-hot label, resulting in information loss and failing to capture inter-image relationships. We propose a distribution-based approach that discretizes the score distribution into a soft label. This method preserves the characteristics of the score distribution, achieving high accuracy and maintaining inter-image relationships. Moreover, to address dataset variation, where different IQA datasets exhibit various distributions, we introduce a fidelity loss based on Thurstone's model. This loss captures intra-dataset relationships, facilitating co-training across multiple IQA datasets. With these designs, we develop the distribution-based Depicted image Quality Assessment model for Score regression (DeQA-Score). Experiments across multiple benchmarks show that DeQA-Score stably outperforms baselines in score regression. Also, DeQA-Score can predict the score distribution that closely aligns with human annotations. Codes and model weights have been released in https://depictqa.github.io/deqa-score/.
comment: Accepted by CVPR 2025
♻ ☆ Quantitative Evaluation of Multiple Instance Learning Reliability For WSIs Classification
Machine learning models have become integral to many fields, but their reliability, particularly in high-stakes domains, remains a critical concern. Reliability refers to the quality of being dependable and trustworthy. Reliable models consistently provide predictions aligned with basic domain knowledge, making their development and deployment particularly critical in healthcare applications. However, Multiple Instance Learning (MIL) models designed for Whole Slide Image (WSI) classification in computational pathology are rarely evaluated in terms of reliability. In this paper, we address this gap by comparing the reliability of MIL models using three proposed metrics, applied across three region-wise annotated datasets. Our findings indicate that the mean pooling instance (MEAN-POOL-INS) model demonstrates superior reliability compared to other networks, despite its simple architectural design and computational efficiency. The code for reproducing our results is available at github.com/tueimage/MIL-Reliability. Keywords: Machine learning, Reliability, Whole Slide Image, Multiple Instance Learning, MEAN-POOL-INS.
♻ ☆ SANA 1.5: Efficient Scaling of Training-Time and Inference-Time Compute in Linear Diffusion Transformer
This paper presents SANA-1.5, a linear Diffusion Transformer for efficient scaling in text-to-image generation. Building upon SANA-1.0, we introduce three key innovations: (1) Efficient Training Scaling: A depth-growth paradigm that enables scaling from 1.6B to 4.8B parameters with significantly reduced computational resources, combined with a memory-efficient 8-bit optimizer. (2) Model Depth Pruning: A block importance analysis technique for efficient model compression to arbitrary sizes with minimal quality loss. (3) Inference-time Scaling: A repeated sampling strategy that trades computation for model capacity, enabling smaller models to match larger model quality at inference time. Through these strategies, SANA-1.5 achieves a text-image alignment score of 0.81 on GenEval, which can be further improved to 0.96 through inference scaling with VILA-Judge, establishing a new SoTA on GenEval benchmark. These innovations enable efficient model scaling across different compute budgets while maintaining high quality, making high-quality image generation more accessible. Our code and pre-trained models are released.
♻ ☆ SVGFusion: Scalable Text-to-SVG Generation via Vector Space Diffusion
In this work, we introduce SVGFusion, a Text-to-SVG model capable of scaling to real-world SVG data without relying on text-based discrete language models or prolonged Score Distillation Sampling (SDS) optimization. The core idea of SVGFusion is to utilize a popular Text-to-Image framework to learn a continuous latent space for vector graphics. Specifically, SVGFusion comprises two key modules: a Vector-Pixel Fusion Variational Autoencoder (VP-VAE) and a Vector Space Diffusion Transformer (VS-DiT). The VP-VAE processes both SVG codes and their corresponding rasterizations to learn a continuous latent space, while the VS-DiT generates latent codes within this space based on the input text prompt. Building on the VP-VAE, we propose a novel rendering sequence modeling strategy which enables the learned latent space to capture the inherent creation logic of SVGs. This allows the model to generate SVGs with higher visual quality and more logical construction, while systematically avoiding occlusion in complex graphic compositions. Additionally, the scalability of SVGFusion can be continuously enhanced by adding more VS-DiT blocks. To effectively train and evaluate SVGFusion, we construct SVGX-Dataset, a large-scale, high-quality SVG dataset that addresses the scarcity of high-quality vector data. Extensive experiments demonstrate the superiority of SVGFusion over existing SVG generation methods, establishing a new framework for SVG content creation. Code, model, and data will be released at: https://ximinng.github.io/SVGFusionProject/
comment: project page: https://ximinng.github.io/SVGFusionProject/
♻ ☆ D$^3$: Scaling Up Deepfake Detection by Learning from Discrepancy CVPR 2025
The boom of Generative AI brings opportunities entangled with risks and concerns. Existing literature emphasizes the generalization capability of deepfake detection on unseen generators, significantly promoting the detector's ability to identify more universal artifacts. This work seeks a step toward a universal deepfake detection system with better generalization and robustness. We do so by first scaling up the existing detection task setup from the one-generator to multiple-generators in training, during which we disclose two challenges presented in prior methodological designs and demonstrate the divergence of detectors' performance. Specifically, we reveal that the current methods tailored for training on one specific generator either struggle to learn comprehensive artifacts from multiple generators or sacrifice their fitting ability for seen generators (i.e., In-Domain (ID) performance) to exchange the generalization for unseen generators (i.e., Out-Of-Domain (OOD) performance). To tackle the above challenges, we propose our Discrepancy Deepfake Detector (D$^3$) framework, whose core idea is to deconstruct the universal artifacts from multiple generators by introducing a parallel network branch that takes a distorted image feature as an extra discrepancy signal and supplement its original counterpart. Extensive scaled-up experiments demonstrate the effectiveness of D$^3$, achieving 5.3% accuracy improvement in the OOD testing compared to the current SOTA methods while maintaining the ID performance. The source code will be updated in our GitHub repository: https://github.com/BigAandSmallq/D3
comment: 13 pages, 3 figures, accepted by CVPR 2025
♻ ☆ Deformable Radial Kernel Splatting
Recently, Gaussian splatting has emerged as a robust technique for representing 3D scenes, enabling real-time rasterization and high-fidelity rendering. However, Gaussians' inherent radial symmetry and smoothness constraints limit their ability to represent complex shapes, often requiring thousands of primitives to approximate detailed geometry. We introduce Deformable Radial Kernel (DRK), which extends Gaussian splatting into a more general and flexible framework. Through learnable radial bases with adjustable angles and scales, DRK efficiently models diverse shape primitives while enabling precise control over edge sharpness and boundary curvature. iven DRK's planar nature, we further develop accurate ray-primitive intersection computation for depth sorting and introduce efficient kernel culling strategies for improved rasterization efficiency. Extensive experiments demonstrate that DRK outperforms existing methods in both representation efficiency and rendering quality, achieving state-of-the-art performance while dramatically reducing primitive count.
♻ ☆ Aligning Multimodal LLM with Human Preference: A Survey
Large language models (LLMs) can handle a wide variety of general tasks with simple prompts, without the need for task-specific training. Multimodal Large Language Models (MLLMs), built upon LLMs, have demonstrated impressive potential in tackling complex tasks involving visual, auditory, and textual data. However, critical issues related to truthfulness, safety, o1-like reasoning, and alignment with human preference remain insufficiently addressed. This gap has spurred the emergence of various alignment algorithms, each targeting different application scenarios and optimization goals. Recent studies have shown that alignment algorithms are a powerful approach to resolving the aforementioned challenges. In this paper, we aim to provide a comprehensive and systematic review of alignment algorithms for MLLMs. Specifically, we explore four key aspects: (1) the application scenarios covered by alignment algorithms, including general image understanding, multi-image, video, and audio, and extended multimodal applications; (2) the core factors in constructing alignment datasets, including data sources, model responses, and preference annotations; (3) the benchmarks used to evaluate alignment algorithms; and (4) a discussion of potential future directions for the development of alignment algorithms. This work seeks to help researchers organize current advancements in the field and inspire better alignment methods. The project page of this paper is available at https://github.com/BradyFU/Awesome-Multimodal-Large-Language-Models/tree/Alignment.
comment: Project page: https://github.com/BradyFU/Awesome-Multimodal-Large-Language-Models/tree/Alignment
♻ ☆ Quantized Spike-driven Transformer ICLR 2025
Spiking neural networks are emerging as a promising energy-efficient alternative to traditional artificial neural networks due to their spike-driven paradigm. However, recent research in the SNN domain has mainly focused on enhancing accuracy by designing large-scale Transformer structures, which typically rely on substantial computational resources, limiting their deployment on resource-constrained devices. To overcome this challenge, we propose a quantized spike-driven Transformer baseline (QSD-Transformer), which achieves reduced resource demands by utilizing a low bit-width parameter. Regrettably, the QSD-Transformer often suffers from severe performance degradation. In this paper, we first conduct empirical analysis and find that the bimodal distribution of quantized spike-driven self-attention (Q-SDSA) leads to spike information distortion (SID) during quantization, causing significant performance degradation. To mitigate this issue, we take inspiration from mutual information entropy and propose a bi-level optimization strategy to rectify the information distribution in Q-SDSA. Specifically, at the lower level, we introduce an information-enhanced LIF to rectify the information distribution in Q-SDSA. At the upper level, we propose a fine-grained distillation scheme for the QSD-Transformer to align the distribution in Q-SDSA with that in the counterpart ANN. By integrating the bi-level optimization strategy, the QSD-Transformer can attain enhanced energy efficiency without sacrificing its high-performance advantage. For instance, when compared to the prior SNN benchmark on ImageNet, the QSD-Transformer achieves 80.3% top-1 accuracy, accompanied by significant reductions of 6.0$\times$ and 8.1$\times$ in power consumption and model size, respectively. Code is available at https://github.com/bollossom/QSD-Transformer.
comment: Accepted by ICLR 2025
Machine Learning 87
☆ Enhance GNNs with Reliable Confidence Estimation via Adversarial Calibration Learning
Despite their impressive predictive performance, GNNs often exhibit poor confidence calibration, i.e., their predicted confidence scores do not accurately reflect true correctness likelihood. This issue raises concerns about their reliability in high-stakes domains such as fraud detection, and risk assessment, where well-calibrated predictions are essential for decision-making. To ensure trustworthy predictions, several GNN calibration methods are proposed. Though they can improve global calibration, our experiments reveal that they often fail to generalize across different node groups, leading to inaccurate confidence in node groups with different degree levels, classes, and local structures. In certain cases, they even degrade calibration compared to the original uncalibrated GNN. To address this challenge, we propose a novel AdvCali framework that adaptively enhances calibration across different node groups. Our method leverages adversarial training to automatically identify mis-calibrated node groups and applies a differentiable Group Expected Calibration Error (ECE) loss term to refine confidence estimation within these groups. This allows the model to dynamically adjust its calibration strategy without relying on dataset-specific prior knowledge about miscalibrated subgroups. Extensive experiments on real-world datasets demonstrate that our approach not only improves global calibration but also significantly enhances calibration within groups defined by feature similarity, topology, and connectivity, outperforming previous methods and demonstrating its effectiveness in practical scenarios.
☆ KEA: Keeping Exploration Alive by Proactively Coordinating Exploration Strategies
Soft Actor-Critic (SAC) has achieved notable success in continuous control tasks but struggles in sparse reward settings, where infrequent rewards make efficient exploration challenging. While novelty-based exploration methods address this issue by encouraging the agent to explore novel states, they are not trivial to apply to SAC. In particular, managing the interaction between novelty-based exploration and SAC's stochastic policy can lead to inefficient exploration and redundant sample collection. In this paper, we propose KEA (Keeping Exploration Alive) which tackles the inefficiencies in balancing exploration strategies when combining SAC with novelty-based exploration. KEA introduces an additional co-behavior agent that works alongside SAC and a switching mechanism to facilitate proactive coordination between exploration strategies from novelty-based exploration and stochastic policy. This coordination allows the agent to maintain stochasticity in high-novelty regions, enhancing exploration efficiency and reducing repeated sample collection. We first analyze this potential issue in a 2D navigation task and then evaluate KEA on sparse reward control tasks from the DeepMind Control Suite. Compared to state-of-the-art novelty-based exploration baselines, our experiments show that KEA significantly improves learning efficiency and robustness in sparse reward setups.
☆ Adaptive Multi-Fidelity Reinforcement Learning for Variance Reduction in Engineering Design Optimization
Multi-fidelity Reinforcement Learning (RL) frameworks efficiently utilize computational resources by integrating analysis models of varying accuracy and costs. The prevailing methodologies, characterized by transfer learning, human-inspired strategies, control variate techniques, and adaptive sampling, predominantly depend on a structured hierarchy of models. However, this reliance on a model hierarchy can exacerbate variance in policy learning when the underlying models exhibit heterogeneous error distributions across the design space. To address this challenge, this work proposes a novel adaptive multi-fidelity RL framework, in which multiple heterogeneous, non-hierarchical low-fidelity models are dynamically leveraged alongside a high-fidelity model to efficiently learn a high-fidelity policy. Specifically, low-fidelity policies and their experience data are adaptively used for efficient targeted learning, guided by their alignment with the high-fidelity policy. The effectiveness of the approach is demonstrated in an octocopter design optimization problem, utilizing two low-fidelity models alongside a high-fidelity simulator. The results demonstrate that the proposed approach substantially reduces variance in policy learning, leading to improved convergence and consistent high-quality solutions relative to traditional hierarchical multi-fidelity RL methods. Moreover, the framework eliminates the need for manually tuning model usage schedules, which can otherwise introduce significant computational overhead. This positions the framework as an effective variance-reduction strategy for multi-fidelity RL, while also mitigating the computational and operational burden of manual fidelity scheduling.
☆ Decoupling Angles and Strength in Low-rank Adaptation ICLR 2025
Parameter-Efficient FineTuning (PEFT) methods have recently gained significant popularity thanks to the widespread availability of large-scale pretrained models. These methods allow for quick adaptation to downstream tasks with minimal computational cost. However, popular finetuning methods such as LoRA exhibit limited robustness when it comes to hyperparameter choices or extended training regimes, preventing optimal out-of-the-box performance. In contrast, bounded approaches, such as ETHER, provide greater robustness but are limited to extremely low-rank adaptations and fixed-strength transformations, reducing their adaptation expressive power. In this work, we propose Decoupled Low-rank Adaptation (DeLoRA), a novel finetuning method that normalizes and scales learnable low-rank matrices. By bounding the distance of the transformation, DeLoRA effectively decouples the angular learning from the adaptation strength, enhancing robustness without compromising performance. Through evaluations on subject-driven image generation, natural language understanding, and instruction tuning, we show that DeLoRA matches or surpasses performance of competing PEFT methods, while exhibiting stronger robustness. Code is available at https://github.com/ExplainableML/DeLoRA.
comment: ICLR 2025
☆ A Framework for Finding Local Saddle Points in Two-Player Zero-Sum Black-Box Games
Saddle point optimization is a critical problem employed in numerous real-world applications, including portfolio optimization, generative adversarial networks, and robotics. It has been extensively studied in cases where the objective function is known and differentiable. Existing work in black-box settings with unknown objectives that can only be sampled either assumes convexity-concavity in the objective to simplify the problem or operates with noisy gradient estimators. In contrast, we introduce a framework inspired by Bayesian optimization which utilizes Gaussian processes to model the unknown (potentially nonconvex-nonconcave) objective and requires only zeroth-order samples. Our approach frames the saddle point optimization problem as a two-level process which can flexibly integrate existing and novel approaches to this problem. The upper level of our framework produces a model of the objective function by sampling in promising locations, and the lower level of our framework uses the existing model to frame and solve a general-sum game to identify locations to sample. This lower level procedure can be designed in complementary ways, and we demonstrate the flexibility of our approach by introducing variants which appropriately trade off between factors like runtime, the cost of function evaluations, and the number of available initial samples. We experimentally demonstrate these algorithms on synthetic and realistic datasets in black-box nonconvex-nonconcave settings, showcasing their ability to efficiently locate local saddle points in these contexts.
☆ Theory-to-Practice Gap for Neural Networks and Neural Operators
This work studies the sampling complexity of learning with ReLU neural networks and neural operators. For mappings belonging to relevant approximation spaces, we derive upper bounds on the best-possible convergence rate of any learning algorithm, with respect to the number of samples. In the finite-dimensional case, these bounds imply a gap between the parametric and sampling complexities of learning, known as the \emph{theory-to-practice gap}. In this work, a unified treatment of the theory-to-practice gap is achieved in a general $L^p$-setting, while at the same time improving available bounds in the literature. Furthermore, based on these results the theory-to-practice gap is extended to the infinite-dimensional setting of operator learning. Our results apply to Deep Operator Networks and integral kernel-based neural operators, including the Fourier neural operator. We show that the best-possible convergence rate in a Bochner $L^p$-norm is bounded by Monte-Carlo rates of order $1/p$.
☆ Adaptive Rank Allocation: Speeding Up Modern Transformers with RaNA Adapters ICLR 2025
Large Language Models (LLMs) are computationally intensive, particularly during inference. Neuron-adaptive techniques, which selectively activate neurons in Multi-Layer Perceptron (MLP) layers, offer some speedups but suffer from limitations in modern Transformers. These include reliance on sparse activations, incompatibility with attention layers, and the use of costly neuron masking techniques. To address these issues, we propose the Adaptive Rank Allocation framework and introduce the Rank and Neuron Allocator (RaNA) adapter. RaNA adapters leverage rank adapters, which operate on linear layers by applying both low-rank matrix decompositions and adaptive masking to efficiently allocate compute without depending on activation sparsity. This enables RaNA to be generally applied to MLPs and linear components of attention modules, while eliminating the need for expensive maskers found in neuron-adaptive methods. Notably, when compared to neuron adapters, RaNA improves perplexity by up to 7 points and increases accuracy by up to 8 percentage-points when reducing FLOPs by $\sim$44% in state-of-the-art Transformer architectures. These results position RaNA as a robust solution for improving inference efficiency in modern Transformer architectures.
comment: 16 pages, 5 figures. ICLR 2025
☆ LakotaBERT: A Transformer-based Model for Low Resource Lakota Language
Lakota, a critically endangered language of the Sioux people in North America, faces significant challenges due to declining fluency among younger generations. This paper introduces LakotaBERT, the first large language model (LLM) tailored for Lakota, aiming to support language revitalization efforts. Our research has two primary objectives: (1) to create a comprehensive Lakota language corpus and (2) to develop a customized LLM for Lakota. We compiled a diverse corpus of 105K sentences in Lakota, English, and parallel texts from various sources, such as books and websites, emphasizing the cultural significance and historical context of the Lakota language. Utilizing the RoBERTa architecture, we pre-trained our model and conducted comparative evaluations against established models such as RoBERTa, BERT, and multilingual BERT. Initial results demonstrate a masked language modeling accuracy of 51% with a single ground truth assumption, showcasing performance comparable to that of English-based models. We also evaluated the model using additional metrics, such as precision and F1 score, to provide a comprehensive assessment of its capabilities. By integrating AI and linguistic methodologies, we aspire to enhance linguistic diversity and cultural resilience, setting a valuable precedent for leveraging technology in the revitalization of other endangered indigenous languages.
ViVa: Video-Trained Value Functions for Guiding Online RL from Diverse Data
Online reinforcement learning (RL) with sparse rewards poses a challenge partly because of the lack of feedback on states leading to the goal. Furthermore, expert offline data with reward signal is rarely available to provide this feedback and bootstrap online learning. How can we guide online agents to the right solution without this on-task data? Reward shaping offers a solution by providing fine-grained signal to nudge the policy towards the optimal solution. However, reward shaping often requires domain knowledge to hand-engineer heuristics for a specific goal. To enable more general and inexpensive guidance, we propose and analyze a data-driven methodology that automatically guides RL by learning from widely available video data such as Internet recordings, off-task demonstrations, task failures, and undirected environment interaction. By learning a model of optimal goal-conditioned value from diverse passive data, we open the floor to scaling up and using various data sources to model general goal-reaching behaviors relevant to guiding online RL. Specifically, we use intent-conditioned value functions to learn from diverse videos and incorporate these goal-conditioned values into the reward. Our experiments show that video-trained value functions work well with a variety of data sources, exhibit positive transfer from human video pre-training, can generalize to unseen goals, and scale with dataset size.
☆ Iterative Multi-Agent Reinforcement Learning: A Novel Approach Toward Real-World Multi-Echelon Inventory Optimization
Multi-echelon inventory optimization (MEIO) is critical for effective supply chain management, but its inherent complexity can pose significant challenges. Heuristics are commonly used to address this complexity, yet they often face limitations in scope and scalability. Recent research has found deep reinforcement learning (DRL) to be a promising alternative to traditional heuristics, offering greater versatility by utilizing dynamic decision-making capabilities. However, since DRL is known to struggle with the curse of dimensionality, its relevance to complex real-life supply chain scenarios is still to be determined. This thesis investigates DRL's applicability to MEIO problems of increasing complexity. A state-of-the-art DRL model was replicated, enhanced, and tested across 13 supply chain scenarios, combining diverse network structures and parameters. To address DRL's challenges with dimensionality, additional models leveraging graph neural networks (GNNs) and multi-agent reinforcement learning (MARL) were developed, culminating in the novel iterative multi-agent reinforcement learning (IMARL) approach. IMARL demonstrated superior scalability, effectiveness, and reliability in optimizing inventory policies, consistently outperforming benchmarks. These findings confirm the potential of DRL, particularly IMARL, to address real-world supply chain challenges and call for additional research to further expand its applicability.
comment: A Capstone Report in the Field of Data Science for the Degree of Master of Liberal Arts in Extension Studies - Harvard University
☆ FROG: Fair Removal on Graphs
As compliance with privacy regulations becomes increasingly critical, the growing demand for data privacy has highlighted the significance of machine unlearning in many real world applications, such as social network and recommender systems, many of which can be represented as graph-structured data. However, existing graph unlearning algorithms indiscriminately modify edges or nodes from well-trained models without considering the potential impact of such structural modifications on fairness. For example, forgetting links between nodes with different genders in a social network may exacerbate group disparities, leading to significant fairness concerns. To address these challenges, we propose a novel approach that jointly optimizes the graph structure and the corresponding model for fair unlearning tasks. Specifically,our approach rewires the graph to enhance unlearning efficiency by removing redundant edges that hinder forgetting while preserving fairness through targeted edge augmentation. Additionally, we introduce a worst-case evaluation mechanism to assess the reliability of fair unlearning performance. Extensive experiments on real-world datasets demonstrate the effectiveness of the proposed approach in achieving superior unlearning outcomes.
☆ Shapley-Guided Utility Learning for Effective Graph Inference Data Valuation
Graph Neural Networks (GNNs) have demonstrated remarkable performance in various graph-based machine learning tasks, yet evaluating the importance of neighbors of testing nodes remains largely unexplored due to the challenge of assessing data importance without test labels. To address this gap, we propose Shapley-Guided Utility Learning (SGUL), a novel framework for graph inference data valuation. SGUL innovatively combines transferable data-specific and modelspecific features to approximate test accuracy without relying on ground truth labels. By incorporating Shapley values as a preprocessing step and using feature Shapley values as input, our method enables direct optimization of Shapley value prediction while reducing computational demands. SGUL overcomes key limitations of existing methods, including poor generalization to unseen test-time structures and indirect optimization. Experiments on diverse graph datasets demonstrate that SGUL consistently outperforms existing baselines in both inductive and transductive settings. SGUL offers an effective, efficient, and interpretable approach for quantifying the value of test-time neighbors.
☆ Quantile-Based Randomized Kaczmarz for Corrupted Tensor Linear Systems
The reconstruction of tensor-valued signals from corrupted measurements, known as tensor regression, has become essential in many multi-modal applications such as hyperspectral image reconstruction and medical imaging. In this work, we address the tensor linear system problem $\mathcal{A} \mathcal{X}=\mathcal{B}$, where $\mathcal{A}$ is a measurement operator, $\mathcal{X}$ is the unknown tensor-valued signal, and $\mathcal{B}$ contains the measurements, possibly corrupted by arbitrary errors. Such corruption is common in large-scale tensor data, where transmission, sensory, or storage errors are rare per instance but likely over the entire dataset and may be arbitrarily large in magnitude. We extend the Kaczmarz method, a popular iterative algorithm for solving large linear systems, to develop a Quantile Tensor Randomized Kaczmarz (QTRK) method robust to large, sparse corruptions in the observations $\mathcal{B}$. This approach combines the tensor Kaczmarz framework with quantile-based statistics, allowing it to mitigate adversarial corruptions and improve convergence reliability. We also propose and discuss the Masked Quantile Randomized Kaczmarz (mQTRK) variant, which selectively applies partial updates to handle corruptions further. We present convergence guarantees, discuss the advantages and disadvantages of our approaches, and demonstrate the effectiveness of our methods through experiments, including an application for video deblurring.
☆ Exploring Energy Landscapes for Minimal Counterfactual Explanations: Applications in Cybersecurity and Beyond
Counterfactual explanations have emerged as a prominent method in Explainable Artificial Intelligence (XAI), providing intuitive and actionable insights into Machine Learning model decisions. In contrast to other traditional feature attribution methods that assess the importance of input variables, counterfactual explanations focus on identifying the minimal changes required to alter a model's prediction, offering a ``what-if'' analysis that is close to human reasoning. In the context of XAI, counterfactuals enhance transparency, trustworthiness and fairness, offering explanations that are not just interpretable but directly applicable in the decision-making processes. In this paper, we present a novel framework that integrates perturbation theory and statistical mechanics to generate minimal counterfactual explanations in explainable AI. We employ a local Taylor expansion of a Machine Learning model's predictive function and reformulate the counterfactual search as an energy minimization problem over a complex landscape. In sequence, we model the probability of candidate perturbations leveraging the Boltzmann distribution and use simulated annealing for iterative refinement. Our approach systematically identifies the smallest modifications required to change a model's prediction while maintaining plausibility. Experimental results on benchmark datasets for cybersecurity in Internet of Things environments, demonstrate that our method provides actionable, interpretable counterfactuals and offers deeper insights into model sensitivity and decision boundaries in high-dimensional spaces.
☆ Adaptive Physics-informed Neural Networks: A Survey
Physics-informed neural networks (PINNs) have emerged as a promising approach to solving partial differential equations (PDEs) using neural networks, particularly in data-scarce scenarios, due to their unsupervised training capability. However, limitations related to convergence and the need for re-optimization with each change in PDE parameters hinder their widespread adoption across scientific and engineering applications. This survey reviews existing research that addresses these limitations through transfer learning and meta-learning. The covered methods improve the training efficiency, allowing faster adaptation to new PDEs with fewer data and computational resources. While traditional numerical methods solve systems of differential equations directly, neural networks learn solutions implicitly by adjusting their parameters. One notable advantage of neural networks is their ability to abstract away from specific problem domains, allowing them to retain, discard, or adapt learned representations to efficiently address similar problems. By exploring the application of these techniques to PINNs, this survey identifies promising directions for future research to facilitate the broader adoption of PINNs in a wide range of scientific and engineering applications.
comment: https://openreview.net/forum?id=vz5P1Kbt6t
☆ Causality-Aware Next Location Prediction Framework based on Human Mobility Stratification
Human mobility data are fused with multiple travel patterns and hidden spatiotemporal patterns are extracted by integrating user, location, and time information to improve next location prediction accuracy. In existing next location prediction methods, different causal relationships that result from patterns in human mobility data are ignored, which leads to confounding information that can have a negative effect on predictions. Therefore, this study introduces a causality-aware framework for next location prediction, focusing on human mobility stratification for travel patterns. In our research, a novel causal graph is developed that describes the relationships between various input variables. We use counterfactuals to enhance the indirect effects in our causal graph for specific travel patterns: non-anchor targeted travels. The proposed framework is designed as a plug-and-play module that integrates multiple next location prediction paradigms. We tested our proposed framework using several state-of-the-art models and human mobility datasets, and the results reveal that the proposed module improves the prediction performance. In addition, we provide results from the ablation study and quantitative study to demonstrate the soundness of our causal graph and its ability to further enhance the interpretability of the current next location prediction models.
comment: Accepted by IEEE UIC 2024
☆ Enhancing Software Vulnerability Detection Using Code Property Graphs and Convolutional Neural Networks
The increasing complexity of modern software systems has led to a rise in vulnerabilities that malicious actors can exploit. Traditional methods of vulnerability detection, such as static and dynamic analysis, have limitations in scalability and automation. This paper proposes a novel approach to detecting software vulnerabilities using a combination of code property graphs and machine learning techniques. By leveraging code property graphs, which integrate abstract syntax trees, control flow graphs, and program dependency graphs, we achieve a detailed representation of software code that enhances the accuracy and granularity of vulnerability detection. We introduce various neural network models, including convolutional neural networks adapted for graph data, to process these representations. Our approach provides a scalable and automated solution for vulnerability detection, addressing the shortcomings of existing methods. We also present a newly generated dataset labeled with function-level vulnerability types sourced from open-source repositories. Our contributions include a methodology for transforming software code into code property graphs, the implementation of a convolutional neural network model for graph data, and the creation of a comprehensive dataset for training and evaluation. This work lays the foundation for more effective and efficient vulnerability detection in complex software systems.
☆ Evaluating Negative Sampling Approaches for Neural Topic Models
Negative sampling has emerged as an effective technique that enables deep learning models to learn better representations by introducing the paradigm of learn-to-compare. The goal of this approach is to add robustness to deep learning models to learn better representation by comparing the positive samples against the negative ones. Despite its numerous demonstrations in various areas of computer vision and natural language processing, a comprehensive study of the effect of negative sampling in an unsupervised domain like topic modeling has not been well explored. In this paper, we present a comprehensive analysis of the impact of different negative sampling strategies on neural topic models. We compare the performance of several popular neural topic models by incorporating a negative sampling technique in the decoder of variational autoencoder-based neural topic models. Experiments on four publicly available datasets demonstrate that integrating negative sampling into topic models results in significant enhancements across multiple aspects, including improved topic coherence, richer topic diversity, and more accurate document classification. Manual evaluations also indicate that the inclusion of negative sampling into neural topic models enhances the quality of the generated topics. These findings highlight the potential of negative sampling as a valuable tool for advancing the effectiveness of neural topic models.
comment: Code is available at: https://github.com/AdhyaSuman/Eval_NegTM
☆ Active Inference for Energy Control and Planning in Smart Buildings and Communities
Active Inference (AIF) is emerging as a powerful framework for decision-making under uncertainty, yet its potential in engineering applications remains largely unexplored. In this work, we propose a novel dual-layer AIF architecture that addresses both building-level and community-level energy management. By leveraging the free energy principle, each layer adapts to evolving conditions and handles partial observability without extensive sensor information and respecting data privacy. We validate the continuous AIF model against both a perfect optimization baseline and a reinforcement learning-based approach. We also test the community AIF framework under extreme pricing scenarios. The results highlight the model's robustness in handling abrupt changes. This study is the first to show how a distributed AIF works in engineering. It also highlights new opportunities for privacy-preserving and uncertainty-aware control strategies in engineering applications.
comment: Submitted to IEEE CASE 2025 (IEEE 21st International Conference on Automation Science and Engineering)
☆ Machine learning based animal emotion classification using audio signals
This paper presents the machine learning approach to the automated classification of a dog's emotional state based on the processing and recognition of audio signals. It offers helpful information for improving human-machine interfaces and developing more precise tools for classifying emotions from acoustic data. The presented model demonstrates an overall accuracy value above 70% for audio signals recorded for one dog.
comment: 5 pages, 3 figures. This paper was originally published in 2022 International Conference on Innovative Solutions in Software Engineering (ICISSE), available: https://zenodo.org/records/7514136
☆ Mitigating Reward Over-Optimization in RLHF via Behavior-Supported Regularization ICLR 2025
Reinforcement learning from human feedback (RLHF) is an effective method for aligning large language models (LLMs) with human values. However, reward over-optimization remains an open challenge leading to discrepancies between the performance of LLMs under the reward model and the true human objectives. A primary contributor to reward over-optimization is the extrapolation error that arises when the reward model evaluates out-of-distribution (OOD) responses. However, current methods still fail to prevent the increasing frequency of OOD response generation during the reinforcement learning (RL) process and are not effective at handling extrapolation errors from OOD responses. In this work, we propose the Behavior-Supported Policy Optimization (BSPO) method to mitigate the reward over-optimization issue. Specifically, we define behavior policy as the next token distribution of the reward training dataset to model the in-distribution (ID) region of the reward model. Building on this, we introduce the behavior-supported Bellman operator to regularize the value function, penalizing all OOD values without impacting the ID ones. Consequently, BSPO reduces the generation of OOD responses during the RL process, thereby avoiding overestimation caused by the reward model's extrapolation errors. Theoretically, we prove that BSPO guarantees a monotonic improvement of the supported policy until convergence to the optimal behavior-supported policy. Empirical results from extensive experiments show that BSPO outperforms baselines in preventing reward over-optimization due to OOD evaluation and finding the optimal ID policy.
comment: Published as a conference paper at ICLR 2025
☆ Feature Learning beyond the Lazy-Rich Dichotomy: Insights from Representational Geometry
The ability to integrate task-relevant information into neural representations is a fundamental aspect of both biological and artificial intelligence. To enable theoretical analysis, recent work has examined whether a network learns task-relevant features (rich learning) or resembles a random feature model (or a kernel machine, i.e., lazy learning). However, this simple lazy-versus-rich dichotomy overlooks the possibility of various subtypes of feature learning that emerge from different architectures, learning rules, and data properties. Furthermore, most existing approaches emphasize weight matrices or neural tangent kernels, limiting their applicability to neuroscience because they do not explicitly characterize representations. In this work, we introduce an analysis framework based on representational geometry to study feature learning. Instead of analyzing what are the learned features, we focus on characterizing how task-relevant representational manifolds evolve during the learning process. In both theory and experiment, we find that when a network learns features useful for solving a task, the task-relevant manifolds become increasingly untangled. Moreover, by tracking changes in the underlying manifold geometry, we uncover distinct learning stages throughout training, as well as different learning strategies associated with training hyperparameters, uncovering subtypes of feature learning beyond the lazy-versus-rich dichotomy. Applying our method to neuroscience and machine learning, we gain geometric insights into the structural inductive biases of neural circuits solving cognitive tasks and the mechanisms underlying out-of-distribution generalization in image classification. Our framework provides a novel geometric perspective for understanding and quantifying feature learning in both artificial and biological neural networks.
☆ AgentRxiv: Towards Collaborative Autonomous Research
Progress in scientific discovery is rarely the result of a single "Eureka" moment, but is rather the product of hundreds of scientists incrementally working together toward a common goal. While existing agent workflows are capable of producing research autonomously, they do so in isolation, without the ability to continuously improve upon prior research results. To address these challenges, we introduce AgentRxiv-a framework that lets LLM agent laboratories upload and retrieve reports from a shared preprint server in order to collaborate, share insights, and iteratively build on each other's research. We task agent laboratories to develop new reasoning and prompting techniques and find that agents with access to their prior research achieve higher performance improvements compared to agents operating in isolation (11.4% relative improvement over baseline on MATH-500). We find that the best performing strategy generalizes to benchmarks in other domains (improving on average by 3.3%). Multiple agent laboratories sharing research through AgentRxiv are able to work together towards a common goal, progressing more rapidly than isolated laboratories, achieving higher overall accuracy (13.7% relative improvement over baseline on MATH-500). These findings suggest that autonomous agents may play a role in designing future AI systems alongside humans. We hope that AgentRxiv allows agents to collaborate toward research goals and enables researchers to accelerate discovery.
☆ Informer in Algorithmic Investment Strategies on High Frequency Bitcoin Data
The article investigates the usage of Informer architecture for building automated trading strategies for high frequency Bitcoin data. Three strategies using Informer model with different loss functions: Root Mean Squared Error (RMSE), Generalized Mean Absolute Directional Loss (GMADL) and Quantile loss, are proposed and evaluated against the Buy and Hold benchmark and two benchmark strategies based on technical indicators. The evaluation is conducted using data of various frequencies: 5 minute, 15 minute, and 30 minute intervals, over the 6 different periods. Although the Informer-based model with Quantile loss did not outperform the benchmark, two other models achieved better results. The performance of the model using RMSE loss worsens when used with higher frequency data while the model that uses novel GMADL loss function is benefiting from higher frequency data and when trained on 5 minute interval it beat all the other strategies on most of the testing periods. The primary contribution of this study is the application and assessment of the RMSE, GMADL, and Quantile loss functions with the Informer model to forecast future returns, subsequently using these forecasts to develop automated trading strategies. The research provides evidence that employing an Informer model trained with the GMADL loss function can result in superior trading outcomes compared to the buy-and-hold approach.
comment: 41 pages, 17 figures, 19 tables
☆ Clarifying Misconceptions in COVID-19 Vaccine Sentiment and Stance Analysis and Their Implications for Vaccine Hesitancy Mitigation: A Systematic Review
Background Advances in machine learning (ML) models have increased the capability of researchers to detect vaccine hesitancy in social media using Natural Language Processing (NLP). A considerable volume of research has identified the persistence of COVID-19 vaccine hesitancy in discourse shared on various social media platforms. Methods Our objective in this study was to conduct a systematic review of research employing sentiment analysis or stance detection to study discourse towards COVID-19 vaccines and vaccination spread on Twitter (officially known as X since 2023). Following registration in the PROSPERO international registry of systematic reviews, we searched papers published from 1 January 2020 to 31 December 2023 that used supervised machine learning to assess COVID-19 vaccine hesitancy through stance detection or sentiment analysis on Twitter. We categorized the studies according to a taxonomy of five dimensions: tweet sample selection approach, self-reported study type, classification typology, annotation codebook definitions, and interpretation of results. We analyzed if studies using stance detection report different hesitancy trends than those using sentiment analysis by examining how COVID-19 vaccine hesitancy is measured, and whether efforts were made to avoid measurement bias. Results Our review found that measurement bias is widely prevalent in studies employing supervised machine learning to analyze sentiment and stance toward COVID-19 vaccines and vaccination. The reporting errors are sufficiently serious that they hinder the generalisability and interpretation of these studies to understanding whether individual opinions communicate reluctance to vaccinate against SARS-CoV-2. Conclusion Improving the reporting of NLP methods is crucial to addressing knowledge gaps in vaccine hesitancy discourse.
comment: 14 pages, 3 figures, 4 tables
☆ HyperNOs: Automated and Parallel Library for Neural Operators Research
This paper introduces HyperNOs, a PyTorch library designed to streamline and automate the process of exploring neural operators, with a special focus on hyperparameter optimization for comprehensive and exhaustive exploration. Indeed, HyperNOs takes advantage of state-of-the-art optimization algorithms and parallel computing implemented in the Ray-tune library to efficiently explore the hyperparameter space of neural operators. We also implement many useful functionalities for studying neural operators with a user-friendly interface, such as the possibility to train the model with a fixed number of parameters or to train the model with multiple datasets and different resolutions. We integrate Fourier neural operators and convolutional neural operators in our library, achieving state of the art results on many representative benchmarks, demonstrating the capabilities of HyperNOs to handle real datasets and modern architectures. The library is designed to be easy to use with the provided model and datasets, but also to be easily extended to use new datasets and custom neural operator architectures.
comment: 25 pages, 11 figures
☆ Temporal Relation Extraction in Clinical Texts: A Span-based Graph Transformer Approach
Temporal information extraction from unstructured text is essential for contextualizing events and deriving actionable insights, particularly in the medical domain. We address the task of extracting clinical events and their temporal relations using the well-studied I2B2 2012 Temporal Relations Challenge corpus. This task is inherently challenging due to complex clinical language, long documents, and sparse annotations. We introduce GRAPHTREX, a novel method integrating span-based entity-relation extraction, clinical large pre-trained language models (LPLMs), and Heterogeneous Graph Transformers (HGT) to capture local and global dependencies. Our HGT component facilitates information propagation across the document through innovative global landmarks that bridge distant entities. Our method improves the state-of-the-art with 5.5% improvement in the tempeval $F_1$ score over the previous best and up to 8.9% improvement on long-range relations, which presents a formidable challenge. This work not only advances temporal information extraction but also lays the groundwork for improved diagnostic and prognostic models through enhanced temporal reasoning.
comment: Introducing a novel method for joint extraction of medical events and temporal relations from free-text, leveraging clinical LPLMs and Heterogeneous Graph Transformers, achieving a 5.5% improvement over the previous state-of-the-art and up to 8.9% on long-range relations
☆ Model-Guardian: Protecting against Data-Free Model Stealing Using Gradient Representations and Deceptive Predictions ICME 2025
Model stealing attack is increasingly threatening the confidentiality of machine learning models deployed in the cloud. Recent studies reveal that adversaries can exploit data synthesis techniques to steal machine learning models even in scenarios devoid of real data, leading to data-free model stealing attacks. Existing defenses against such attacks suffer from limitations, including poor effectiveness, insufficient generalization ability, and low comprehensiveness. In response, this paper introduces a novel defense framework named Model-Guardian. Comprising two components, Data-Free Model Stealing Detector (DFMS-Detector) and Deceptive Predictions (DPreds), Model-Guardian is designed to address the shortcomings of current defenses with the help of the artifact properties of synthetic samples and gradient representations of samples. Extensive experiments on seven prevalent data-free model stealing attacks showcase the effectiveness and superior generalization ability of Model-Guardian, outperforming eleven defense methods and establishing a new state-of-the-art performance. Notably, this work pioneers the utilization of various GANs and diffusion models for generating highly realistic query samples in attacks, with Model-Guardian demonstrating accurate detection capabilities.
comment: Full version of the paper accepted by ICME 2025
☆ Self-Explaining Neural Networks for Business Process Monitoring
Tasks in Predictive Business Process Monitoring (PBPM), such as Next Activity Prediction, focus on generating useful business predictions from historical case logs. Recently, Deep Learning methods, particularly sequence-to-sequence models like Long Short-Term Memory (LSTM), have become a dominant approach for tackling these tasks. However, to enhance model transparency, build trust in the predictions, and gain a deeper understanding of business processes, it is crucial to explain the decisions made by these models. Existing explainability methods for PBPM decisions are typically *post-hoc*, meaning they provide explanations only after the model has been trained. Unfortunately, these post-hoc approaches have shown to face various challenges, including lack of faithfulness, high computational costs and a significant sensitivity to out-of-distribution samples. In this work, we introduce, to the best of our knowledge, the first *self-explaining neural network* architecture for predictive process monitoring. Our framework trains an LSTM model that not only provides predictions but also outputs a concise explanation for each prediction, while adapting the optimization objective to improve the reliability of the explanation. We first demonstrate that incorporating explainability into the training process does not hurt model performance, and in some cases, actually improves it. Additionally, we show that our method outperforms post-hoc approaches in terms of both the faithfulness of the generated explanations and substantial improvements in efficiency.
comment: To appear in ICSBT 2025
☆ Dynamic Allocation Hypernetwork with Adaptive Model Recalibration for FCL
Federated continual learning (FCL) offers an emerging pattern to facilitate the applicability of federated learning (FL) in real-world scenarios, where tasks evolve dynamically and asynchronously across clients, especially in medical scenario. Existing server-side FCL methods in nature domain construct a continually learnable server model by client aggregation on all-involved tasks. However, they are challenged by: (1) Catastrophic forgetting for previously learned tasks, leading to error accumulation in server model, making it difficult to sustain comprehensive knowledge across all tasks. (2) Biased optimization due to asynchronous tasks handled across different clients, leading to the collision of optimization targets of different clients at the same time steps. In this work, we take the first step to propose a novel server-side FCL pattern in medical domain, Dynamic Allocation Hypernetwork with adaptive model recalibration (\textbf{FedDAH}). It is to facilitate collaborative learning under the distinct and dynamic task streams across clients. To alleviate the catastrophic forgetting, we propose a dynamic allocation hypernetwork (DAHyper) where a continually updated hypernetwork is designed to manage the mapping between task identities and their associated model parameters, enabling the dynamic allocation of the model across clients. For the biased optimization, we introduce a novel adaptive model recalibration (AMR) to incorporate the candidate changes of historical models into current server updates, and assign weights to identical tasks across different time steps based on the similarity for continual optimization. Extensive experiments on the AMOS dataset demonstrate the superiority of our FedDAH to other FCL methods on sites with different task streams. The code is available:https://github.com/jinlab-imvr/FedDAH.
☆ Surrogate Learning in Meta-Black-Box Optimization: A Preliminary Study GECCO 2025
Recent Meta-Black-Box Optimization (MetaBBO) approaches have shown possibility of enhancing the optimization performance through learning meta-level policies to dynamically configure low-level optimizers. However, existing MetaBBO approaches potentially consume massive function evaluations to train their meta-level policies. Inspired by the recent trend of using surrogate models for cost-friendly evaluation of expensive optimization problems, in this paper, we propose a novel MetaBBO framework which combines surrogate learning process and reinforcement learning-aided Differential Evolution algorithm, namely Surr-RLDE, to address the intensive function evaluation in MetaBBO. Surr-RLDE comprises two learning stages: surrogate learning and policy learning. In surrogate learning, we train a Kolmogorov-Arnold Networks (KAN) with a novel relative-order-aware loss to accurately approximate the objective functions of the problem instances used for subsequent policy learning. In policy learning, we employ reinforcement learning (RL) to dynamically configure the mutation operator in DE. The learned surrogate model is integrated into the training of the RL-based policy to substitute for the original objective function, which effectively reduces consumed evaluations during policy learning. Extensive benchmark results demonstrate that Surr-RLDE not only shows competitive performance to recent baselines, but also shows compelling generalization for higher-dimensional problems. Further ablation studies underscore the effectiveness of each technical components in Surr-RLDE. We open-source Surr-RLDE at https://github.com/GMC-DRL/Surr-RLDE.
comment: Accepted as full paper at ACM GECCO 2025
☆ Reinforcement Learning-based Self-adaptive Differential Evolution through Automated Landscape Feature Learning GECCO 2025
Recently, Meta-Black-Box-Optimization (MetaBBO) methods significantly enhance the performance of traditional black-box optimizers through meta-learning flexible and generalizable meta-level policies that excel in dynamic algorithm configuration (DAC) tasks within the low-level optimization, reducing the expertise required to adapt optimizers for novel optimization tasks. Though promising, existing MetaBBO methods heavily rely on human-crafted feature extraction approach to secure learning effectiveness. To address this issue, this paper introduces a novel MetaBBO method that supports automated feature learning during the meta-learning process, termed as RLDE-AFL, which integrates a learnable feature extraction module into a reinforcement learning-based DE method to learn both the feature encoding and meta-level policy. Specifically, we design an attention-based neural network with mantissa-exponent based embedding to transform the solution populations and corresponding objective values during the low-level optimization into expressive landscape features. We further incorporate a comprehensive algorithm configuration space including diverse DE operators into a reinforcement learning-aided DAC paradigm to unleash the behavior diversity and performance of the proposed RLDE-AFL. Extensive benchmark results show that co-training the proposed feature learning module and DAC policy contributes to the superior optimization performance of RLDE-AFL to several advanced DE methods and recent MetaBBO baselines over both synthetic and realistic BBO scenarios. The source codes of RLDE-AFL are available at https://github.com/GMC-DRL/RLDE-AFL.
comment: Accepted as full paper at ACM GECCO 2025
☆ Interpretable Feature Interaction via Statistical Self-supervised Learning on Tabular Data
In high-dimensional and high-stakes contexts, ensuring both rigorous statistical guarantees and interpretability in feature extraction from complex tabular data remains a formidable challenge. Traditional methods such as Principal Component Analysis (PCA) reduce dimensionality and identify key features that explain the most variance, but are constrained by their reliance on linear assumptions. In contrast, neural networks offer assumption-free feature extraction through self-supervised learning techniques such as autoencoders, though their interpretability remains a challenge in fields requiring transparency. To address this gap, this paper introduces Spofe, a novel self-supervised machine learning pipeline that marries the power of kernel principal components for capturing nonlinear dependencies with a sparse and principled polynomial representation to achieve clear interpretability with statistical rigor. Underpinning our approach is a robust theoretical framework that delivers precise error bounds and rigorous false discovery rate (FDR) control via a multi-objective knockoff selection procedure; it effectively bridges the gap between data-driven complexity and statistical reliability via three stages: (1) generating self-supervised signals using kernel principal components to model complex patterns, (2) distilling these signals into sparse polynomial functions for improved interpretability, and (3) applying a multi-objective knockoff selection procedure with significance testing to rigorously identify important features. Extensive experiments on diverse real-world datasets demonstrate the effectiveness of Spofe, consistently surpassing KPCA, SKPCA, and other methods in feature selection for regression and classification tasks. Visualization and case studies highlight its ability to uncover key insights, enhancing interpretability and practical utility.
☆ Decision from Suboptimal Classifiers: Excess Risk Pre- and Post-Calibration
Probabilistic classifiers are central for making informed decisions under uncertainty. Based on the maximum expected utility principle, optimal decision rules can be derived using the posterior class probabilities and misclassification costs. Yet, in practice only learned approximations of the oracle posterior probabilities are available. In this work, we quantify the excess risk (a.k.a. regret) incurred using approximate posterior probabilities in batch binary decision-making. We provide analytical expressions for miscalibration-induced regret ($R^{\mathrm{CL}}$), as well as tight and informative upper and lower bounds on the regret of calibrated classifiers ($R^{\mathrm{GL}}$). These expressions allow us to identify regimes where recalibration alone addresses most of the regret, and regimes where the regret is dominated by the grouping loss, which calls for post-training beyond recalibration. Crucially, both $R^{\mathrm{CL}}$ and $R^{\mathrm{GL}}$ can be estimated in practice using a calibration curve and a recent grouping loss estimator. On NLP experiments, we show that these quantities identify when the expected gain of more advanced post-training is worth the operational cost. Finally, we highlight the potential of multicalibration approaches as efficient alternatives to costlier fine-tuning approaches.
☆ Regularization of ML models for Earth systems by using longer model timesteps
Regularization is a technique to improve generalization of machine learning (ML) models. A common form of regularization in the ML literature is to train on data where similar inputs map to different outputs. This improves generalization by preventing ML models from becoming overconfident in their predictions. This paper shows how using longer timesteps when modelling chaotic Earth systems naturally leads to more of this regularization. We show this in two domains. We explain how using longer model timesteps can improve results and demonstrate that increased regularization is one of the causes. We explain why longer model timesteps lead to improved regularization in these systems and present a procedure to pick the model timestep. We also carry out a benchmarking exercise on ORAS5 ocean reanalysis data to show that a longer model timestep (28 days) than is typically used gives realistic simulations. We suggest that there will be many opportunities to use this type of regularization in Earth system problems because the Earth system is chaotic and the regularization is so easy to implement.
☆ Lost in Cultural Translation: Do LLMs Struggle with Math Across Cultural Contexts?
Large Language Models (LLMs) have significantly advanced various fields, particularly coding, mathematical reasoning, and logical problem solving. However, a critical question remains: Do these mathematical reasoning abilities persist when LLMs are presented with culturally adapted math problems? Specifically, how do LLMs perform when faced with math problems embedded in cultural contexts that have no significant representation in main stream web-scale AI training data? To explore this, we generated six synthetic cultural datasets from GSM8K, a widely used benchmark for assessing LLMs' mathematical reasoning skills. While preserving the mathematical logic and numerical values of the original GSM8K test set, we modify cultural elements such as personal names, food items, place names, etc. These culturally adapted datasets provide a more reliable framework for evaluating LLMs' mathematical reasoning under shifting cultural contexts. Our findings reveal that LLMs struggle with math problems when cultural references change, even though the underlying mathematical structure remains constant. Smaller models exhibit greater performance drops compared to larger models. Interestingly, our results also suggest that cultural familiarity can enhance mathematical reasoning. Even models with no explicit mathematical training but exposure to relevant cultural contexts sometimes outperform larger, mathematically proficient models on culturally embedded math problems. This study highlights the impact of cultural context on the mathematical reasoning abilities of LLMs, underscoring the need for more diverse and representative training data to improve robustness in real-world applications. The benchmark data sets and script for reproducing the results are available at https://github.com/akarim23131/Lost_in_Cultural_Translation
☆ Predicting Multitasking in Manual and Automated Driving with Optimal Supervisory Control
Modern driving involves interactive technologies that can divert attention, increasing the risk of accidents. This paper presents a computational cognitive model that simulates human multitasking while driving. Based on optimal supervisory control theory, the model predicts how multitasking adapts to variations in driving demands, interactive tasks, and automation levels. Unlike previous models, it accounts for context-dependent multitasking across different degrees of driving automation. The model predicts longer in-car glances on straight roads and shorter glances during curves. It also anticipates increased glance durations with driver aids such as lane-centering assistance and their interaction with environmental demands. Validated against two empirical datasets, the model offers insights into driver multitasking amid evolving in-car technologies and automation.
☆ On the Origins of Sampling Bias: Implications on Fairness Measurement and Mitigation
Accurately measuring discrimination is crucial to faithfully assessing fairness of trained machine learning (ML) models. Any bias in measuring discrimination leads to either amplification or underestimation of the existing disparity. Several sources of bias exist and it is assumed that bias resulting from machine learning is born equally by different groups (e.g. females vs males, whites vs blacks, etc.). If, however, bias is born differently by different groups, it may exacerbate discrimination against specific sub-populations. Sampling bias, in particular, is inconsistently used in the literature to describe bias due to the sampling procedure. In this paper, we attempt to disambiguate this term by introducing clearly defined variants of sampling bias, namely, sample size bias (SSB) and underrepresentation bias (URB). Through an extensive set of experiments on benchmark datasets and using mainstream learning algorithms, we expose relevant observations in several model training scenarios. The observations are finally framed as actionable recommendations for practitioners.
☆ Equivariant Machine Learning Interatomic Potentials with Global Charge Redistribution
Machine learning interatomic potentials (MLIPs) provide a computationally efficient alternative to quantum mechanical simulations for predicting material properties. Message-passing graph neural networks, commonly used in these MLIPs, rely on local descriptor-based symmetry functions to model atomic interactions. However, such local descriptor-based approaches struggle with systems exhibiting long-range interactions, charge transfer, and compositional heterogeneity. In this work, we develop a new equivariant MLIP incorporating long-range Coulomb interactions through explicit treatment of electronic degrees of freedom, specifically global charge distribution within the system. This is achieved using a charge equilibration scheme based on predicted atomic electronegativities. We systematically evaluate our model across a range of benchmark periodic and non-periodic datasets, demonstrating that it outperforms both short-range equivariant and long-range invariant MLIPs in energy and force predictions. Our approach enables more accurate and efficient simulations of systems with long-range interactions and charge heterogeneity, expanding the applicability of MLIPs in computational materials science.
comment: 24 pages, 5 figures, 1 table + 12 pages of Supporting Information
☆ Dataset Distillation for Quantum Neural Networks
Training Quantum Neural Networks (QNNs) on large amount of classical data can be both time consuming as well as expensive. Higher amount of training data would require higher number of gradient descent steps to reach convergence. This, in turn would imply that the QNN will require higher number of quantum executions, thereby driving up its overall execution cost. In this work, we propose performing the dataset distillation process for QNNs, where we use a novel quantum variant of classical LeNet model containing residual connection and trainable Hermitian observable in the Parametric Quantum Circuit (PQC) of the QNN. This approach yields highly informative yet small number of training data at similar performance as the original data. We perform distillation for MNIST and Cifar-10 datasets, and on comparison with classical models observe that both the datasets yield reasonably similar post-inferencing accuracy on quantum LeNet (91.9% MNIST, 50.3% Cifar-10) compared to classical LeNet (94% MNIST, 54% Cifar-10). We also introduce a non-trainable Hermitian for ensuring stability in the distillation process and note marginal reduction of up to 1.8% (1.3%) for MNIST (Cifar-10) dataset.
comment: 5 pages, 4 figures, 2 tables
☆ WLB-LLM: Workload-Balanced 4D Parallelism for Large Language Model Training
In this work, we present WLB-LLM, a workLoad-balanced 4D parallelism for large language model training. We first thoroughly analyze the workload imbalance issue in LLM training and identify two primary sources of imbalance at the pipeline parallelism and context parallelism levels. Then, to address the imbalance issue, at the pipeline parallelism level, WLB-LLM incorporates a workload-aware variable-length document packing method to balance the computation and communication workload across micro-batches. Additionally, at the context parallelism level, WLB-LLM introduces a novel fine-grained per-document sharding strategy, ensuring each worker within a context parallelism group has an identical workload. Comprehensive experiments under different model scales demonstrate that WLB-LLM significantly mitigates the workload imbalance during 4D parallelism LLM training and achieves an average speedup of 1.23x when applying WLB-LLM in our internal LLM training framework.
comment: 12 pages, 16 figures
☆ Predicting performance-related properties of refrigerant based on tailored small-molecule functional group contribution
As current group contribution (GC) methods are mostly proposed for a wide size-range of molecules, applying them to property prediction of small refrigerant molecules could lead to unacceptable errors. In this sense, for the design of novel refrigerants and refrigeration systems, tailoring GC-based models specifically fitted to refrigerant molecules is of great interest. In this work, databases of potential refrigerant molecules are first collected, focusing on five key properties related to the operational efficiency of refrigeration systems, namely normal boiling point, critical temperature, critical pressure, enthalpy of vaporization, and acentric factor. Based on tailored small-molecule groups, the GC method is combined with machine learning (ML) to model these performance-related properties. Following the development of GC-ML models, their performance is analyzed to highlight the potential group-to-property contributions. Additionally, the refrigerant property databases are extended internally and externally, based on which examples are presented to highlight the significance of the developed models.
☆ Cat-AIR: Content and Task-Aware All-in-One Image Restoration
All-in-one image restoration seeks to recover high-quality images from various types of degradation using a single model, without prior knowledge of the corruption source. However, existing methods often struggle to effectively and efficiently handle multiple degradation types. We present Cat-AIR, a novel \textbf{C}ontent \textbf{A}nd \textbf{T}ask-aware framework for \textbf{A}ll-in-one \textbf{I}mage \textbf{R}estoration. Cat-AIR incorporates an alternating spatial-channel attention mechanism that adaptively balances the local and global information for different tasks. Specifically, we introduce cross-layer channel attentions and cross-feature spatial attentions that allocate computations based on content and task complexity. Furthermore, we propose a smooth learning strategy that allows for seamless adaptation to new restoration tasks while maintaining performance on existing ones. Extensive experiments demonstrate that Cat-AIR achieves state-of-the-art results across a wide range of restoration tasks, requiring fewer FLOPs than previous methods, establishing new benchmarks for efficient all-in-one image restoration.
☆ Financial Wind Tunnel: A Retrieval-Augmented Market Simulator
Market simulator tries to create high-quality synthetic financial data that mimics real-world market dynamics, which is crucial for model development and robust assessment. Despite continuous advancements in simulation methodologies, market fluctuations vary in terms of scale and sources, but existing frameworks often excel in only specific tasks. To address this challenge, we propose Financial Wind Tunnel (FWT), a retrieval-augmented market simulator designed to generate controllable, reasonable, and adaptable market dynamics for model testing. FWT offers a more comprehensive and systematic generative capability across different data frequencies. By leveraging a retrieval method to discover cross-sectional information as the augmented condition, our diffusion-based simulator seamlessly integrates both macro- and micro-level market patterns. Furthermore, our framework allows the simulation to be controlled with wide applicability, including causal generation through "what-if" prompts or unprecedented cross-market trend synthesis. Additionally, we develop an automated optimizer for downstream quantitative models, using stress testing of simulated scenarios via FWT to enhance returns while controlling risks. Experimental results demonstrate that our approach enables the generalizable and reliable market simulation, significantly improve the performance and adaptability of downstream models, particularly in highly complex and volatile market conditions. Our code and data sample is available at https://anonymous.4open.science/r/fwt_-E852
☆ Does GCL Need a Large Number of Negative Samples? Enhancing Graph Contrastive Learning with Effective and Efficient Negative Sampling
Graph Contrastive Learning (GCL) aims to self-supervised learn low-dimensional graph representations, primarily through instance discrimination, which involves manually mining positive and negative pairs from graphs, increasing the similarity of positive pairs while decreasing negative pairs. Drawing from the success of Contrastive Learning (CL) in other domains, a consensus has been reached that the effectiveness of GCLs depends on a large number of negative pairs. As a result, despite the significant computational overhead, GCLs typically leverage as many negative node pairs as possible to improve model performance. However, given that nodes within a graph are interconnected, we argue that nodes cannot be treated as independent instances. Therefore, we challenge this consensus: Does employing more negative nodes lead to a more effective GCL model? To answer this, we explore the role of negative nodes in the commonly used InfoNCE loss for GCL and observe that: (1) Counterintuitively, a large number of negative nodes can actually hinder the model's ability to distinguish nodes with different semantics. (2) A smaller number of high-quality and non-topologically coupled negative nodes are sufficient to enhance the discriminability of representations. Based on these findings, we propose a new method called GCL with Effective and Efficient Negative samples, E2Neg, which learns discriminative representations using only a very small set of representative negative samples. E2Neg significantly reduces computational overhead and speeds up model training. We demonstrate the effectiveness and efficiency of E2Neg across multiple datasets compared to other GCL methods.
☆ Finding Stable Subnetworks at Initialization with Dataset Distillation
Recent works have shown that Dataset Distillation, the process for summarizing the training data, can be leveraged to accelerate the training of deep learning models. However, its impact on training dynamics, particularly in neural network pruning, remains largely unexplored. In our work, we use distilled data in the inner loop of iterative magnitude pruning to produce sparse, trainable subnetworks at initialization -- more commonly known as lottery tickets. While using 150x less training points, our algorithm matches the performance of traditional lottery ticket rewinding on ResNet-18 & CIFAR-10. Previous work highlights that lottery tickets can be found when the dense initialization is stable to SGD noise (i.e. training across different ordering of the data converges to the same minima). We extend this discovery, demonstrating that stable subnetworks can exist even within an unstable dense initialization. In our linear mode connectivity studies, we find that pruning with distilled data discards parameters that contribute to the sharpness of the loss landscape. Lastly, we show that by first generating a stable sparsity mask at initialization, we can find lottery tickets at significantly higher sparsities than traditional iterative magnitude pruning.
☆ GLADMamba: Unsupervised Graph-Level Anomaly Detection Powered by Selective State Space Model
Unsupervised graph-level anomaly detection (UGLAD) is a critical and challenging task across various domains, such as social network analysis, anti-cancer drug discovery, and toxic molecule identification. However, existing methods often struggle to capture the long-range dependencies efficiently and neglect the spectral information. Recently, selective State Space Models (SSMs), particularly Mamba, have demonstrated remarkable advantages in capturing long-range dependencies with linear complexity and a selection mechanism. Motivated by their success across various domains, we propose GLADMamba, a novel framework that adapts the selective state space model into UGLAD field. We design View-Fused Mamba (VFM) with a Mamba-Transformer-style architecture to efficiently fuse information from different views with a selective state mechanism. We also design Spectrum-Guided Mamba (SGM) with a Mamba-Transformer-style architecture to leverage the Rayleigh quotient to guide the embedding refining process. GLADMamba can dynamically focus on anomaly-related information while discarding irrelevant information for anomaly detection. To the best of our knowledge, this is the first work to introduce Mamba and explicit spectral information to UGLAD. Extensive experiments on 12 real-world datasets demonstrate that GLADMamba outperforms existing state-of-the-art methods, achieving superior performance in UGLAD. The code is available at https://github.com/Yali-F/GLADMamba.
♻ ☆ SFO: Piloting VLM Feedback for Offline RL
While internet-scale image and textual data have enabled strong generalization in Vision-Language Models (VLMs), the absence of internet-scale control data has impeded the development of similar generalization in standard reinforcement learning (RL) agents. Although VLMs are fundamentally limited in their ability to solve control tasks due to their lack of action-conditioned training data, their capacity for image understanding allows them to provide valuable feedback in RL tasks by recognizing successful outcomes. A key challenge in Reinforcement Learning from AI Feedback (RLAIF) is determining how best to integrate VLM-derived signals into the learning process. We explore this question in the context of offline RL and introduce a class of methods called sub-trajectory filtered optimization. We identify three key insights. First, trajectory length plays a crucial role in offline RL, as full-trajectory preference learning exacerbates the stitching problem, necessitating the use of sub-trajectories. Second, even in Markovian environments, a non-Markovian reward signal from a sequence of images is required to assess trajectory improvement, as VLMs do not interpret control actions and must rely on visual cues over time. Third, a simple yet effective approach--filtered and weighted behavior cloning--consistently outperforms more complex reinforcement learning from human feedback-based methods. We propose sub-trajectory filtered behavior cloning, a method that leverages VLM feedback on sub-trajectories while incorporating a retrospective filtering mechanism that removes sub-trajectories preceding failures to improve robustness and prevent turbulence. This study is preliminary; we provide initial evidence through evaluations on a toy control domain. Please enjoy our airport puns.
comment: Code is provided at https://github.com/jacooba/OfflineRLAIF
♻ ☆ Lower Bounds for the Convergence of Tensor Power Iteration on Random Overcomplete Models COLT 2023
Tensor decomposition serves as a powerful primitive in statistics and machine learning, and has numerous applications in problems such as learning latent variable models or mixture of Gaussians. In this paper, we focus on using power iteration to decompose an overcomplete random tensor. Past work studying the properties of tensor power iteration either requires a non-trivial data-independent initialization, or is restricted to the undercomplete regime. Moreover, several papers implicitly suggest that logarithmically many iterations (in terms of the input dimension) are sufficient for the power method to recover one of the tensor components. Here we present a novel analysis of the dynamics of tensor power iteration from random initialization in the overcomplete regime, where the tensor rank is much greater than its dimension. Surprisingly, we show that polynomially many steps are necessary for convergence of tensor power iteration to any of the true component, which refutes the previous conjecture. On the other hand, our numerical experiments suggest that tensor power iteration successfully recovers tensor components for a broad range of parameters in polynomial time. To further complement our empirical evidence, we prove that a popular objective function for tensor decomposition is strictly increasing along the power iteration path. Our proof is based on the Gaussian conditioning technique, which has been applied to analyze the approximate message passing (AMP) algorithm. The major ingredient of our argument is a conditioning lemma that allows us to generalize AMP-type analysis to non-proportional limit and polynomially many iterations of the power method.
comment: 41 pages, 3 figures, appeared in The 36th Annual Conference on Learning Theory (COLT 2023)
♻ ☆ Sharp Analysis of Power Iteration for Tensor PCA
We investigate the power iteration algorithm for the tensor PCA model introduced in Richard and Montanari (2014). Previous work studying the properties of tensor power iteration is either limited to a constant number of iterations, or requires a non-trivial data-independent initialization. In this paper, we move beyond these limitations and analyze the dynamics of randomly initialized tensor power iteration up to polynomially many steps. Our contributions are threefold: First, we establish sharp bounds on the number of iterations required for power method to converge to the planted signal, for a broad range of the signal-to-noise ratios. Second, our analysis reveals that the actual algorithmic threshold for power iteration is smaller than the one conjectured in literature by a polylog(n) factor, where n is the ambient dimension. Finally, we propose a simple and effective stopping criterion for power iteration, which provably outputs a solution that is highly correlated with the true signal. Extensive numerical experiments verify our theoretical results.
comment: 42 pages, 11 figures, added additional experiments
♻ ☆ Distributed Bayesian Estimation in Sensor Networks: Consensus on Marginal Densities
In this paper, we aim to design and analyze distributed Bayesian estimation algorithms for sensor networks. The challenges we address are to (i) derive a distributed provably-correct algorithm in the functional space of probability distributions over continuous variables, and (ii) leverage these results to obtain new distributed estimators restricted to subsets of variables observed by individual agents. This relates to applications such as cooperative localization and federated learning, where the data collected at any agent depends on a subset of all variables of interest. We present Bayesian density estimation algorithms using data from non-linear likelihoods at agents in centralized, distributed, and marginal distributed settings. After setting up a distributed estimation objective, we prove almost-sure convergence to the optimal set of pdfs at each agent. Then, we prove the same for a storage-aware algorithm estimating densities only over relevant variables at each agent. Finally, we present a Gaussian version of these algorithms and implement it in a mapping problem using variational inference to handle non-linear likelihood models associated with LiDAR sensing.
♻ ☆ Bayesian Active Learning for Multi-Criteria Comparative Judgement in Educational Assessment
Comparative Judgement (CJ) provides an alternative assessment approach by evaluating work holistically rather than breaking it into discrete criteria. This method leverages human ability to make nuanced comparisons, yielding more reliable and valid assessments. CJ aligns with real-world evaluations, where overall quality emerges from the interplay of various elements. However, rubrics remain widely used in education, offering structured criteria for grading and detailed feedback. This creates a gap between CJ's holistic ranking and the need for criterion-based performance breakdowns. This paper addresses this gap using a Bayesian approach. We build on Bayesian CJ (BCJ) by Gray et al., which directly models preferences instead of using likelihoods over total scores, allowing for expected ranks with uncertainty estimation. Their entropy-based active learning method selects the most informative pairwise comparisons for assessors. We extend BCJ to handle multiple independent learning outcome (LO) components, defined by a rubric, enabling both holistic and component-wise predictive rankings with uncertainty estimates. Additionally, we propose a method to aggregate entropies and identify the most informative comparison for assessors. Experiments on synthetic and real data demonstrate our method's effectiveness. Finally, we address a key limitation of BCJ, which is the inability to quantify assessor agreement. We show how to derive agreement levels, enhancing transparency in assessment.
♻ ☆ Synthesis and Analysis of Data as Probability Measures with Entropy-Regularized Optimal Transport AISTATS 2025
We consider synthesis and analysis of probability measures using the entropy-regularized Wasserstein-2 cost and its unbiased version, the Sinkhorn divergence. The synthesis problem consists of computing the barycenter, with respect to these costs, of reference measures given a set of coefficients belonging to the simplex. The analysis problem consists of finding the coefficients for the closest barycenter in the Wasserstein-2 distance to a given measure. Under the weakest assumptions on the measures thus far in the literature, we compute the derivative of the entropy-regularized Wasserstein-2 cost. We leverage this to establish a characterization of barycenters with respect to the entropy-regularized Wasserstein-2 cost as solutions that correspond to a fixed point of an average of the entropy-regularized displacement maps. This characterization yields a finite-dimensional, convex, quadratic program for solving the analysis problem when the measure being analyzed is a barycenter with respect to the entropy-regularized Wasserstein-2 cost. We show that these coefficients, as well as the value of the barycenter functional, can be estimated from samples with dimension-independent rates of convergence, and that barycentric coefficients are stable with respect to perturbations in the Wasserstein-2 metric. We employ the barycentric coefficients as features for classification of corrupted point cloud data, and show that compared to neural network baselines, our approach is more efficient in small training data regimes.
comment: 58 pages. v3: accepted version, minor fixes to results, literature survey, updated simulations. To appear in AISTATS 2025. Code to reproduce experiments: https://github.com/brendanmallery9/Entropic-Barycenters
♻ ☆ DNA Bench: When Silence is Smarter -- Benchmarking Over-Reasoning in Reasoning LLMs
Test-time scaling has significantly improved large language model performance, enabling deeper reasoning to solve complex problems. However, this increased reasoning capability also leads to excessive token generation and unnecessary problem-solving attempts. We introduce Don\'t Answer Bench (DNA Bench), a new benchmark designed to evaluate LLMs ability to robustly understand the tricky reasoning triggers and avoiding unnecessary generation. DNA Bench consists of 150 adversarially designed prompts that are easy for humans to understand and respond to, but surprisingly not for many of the recent prominent LLMs. DNA Bench tests models abilities across different capabilities, such as instruction adherence, hallucination avoidance, redundancy filtering, and unanswerable question recognition. We evaluate reasoning LLMs (RLMs), including DeepSeek-R1, OpenAI O3-mini, Claude-3.7-sonnet and compare them against a powerful non-reasoning model, e.g., GPT-4o. Our experiments reveal that RLMs generate up to 70x more tokens than necessary, often failing at tasks that simpler non-reasoning models handle efficiently with higher accuracy. Our findings underscore the need for more effective training and inference strategies in RLMs.
♻ ☆ Atom of Thoughts for Markov LLM Test-Time Scaling
Large Language Models (LLMs) achieve superior performance through training-time scaling, and test-time scaling further enhances their capabilities by conducting effective reasoning during inference. However, as the scale of reasoning increases, existing test-time scaling methods suffer from accumulated historical information, which not only wastes computational resources but also interferes with effective reasoning. To address this issue, we observe that complex reasoning can be achieved by solving a series of independent and self-contained subquestions. These subquestions are essentially \textit{atomic questions}, exhibiting the memoryless property similar to Markov processes. Based on this observation, we propose Atom of Thoughts (\our), where each state transition consists of decomposing the current question into a dependency-based directed acyclic graph and contracting its subquestions, forming a simplified question that maintains answer equivalence with the original problem. This answer preservation enables the iterative \textit{decomposition-contraction} process to naturally form a meaningful Markov reasoning process. Furthermore, these atomic states can be seamlessly integrated into existing test-time scaling methods, enabling \our to serve as a plug-in enhancement for improving reasoning capabilities. Experiments across six benchmarks demonstrate the effectiveness of \our both as a standalone framework and a plug-in enhancement. Notably, on HotpotQA, when applied to gpt-4o-mini, \our achieves an \textbf{80.6\%} F1 score, surpassing o3-mini by \textbf{3.4\%} and DeepSeek-R1 by \textbf{10.6\%}. The code is available at \href{https://github.com/qixucen/atom}{https://github.com/qixucen/atom}.
♻ ☆ The Representational Status of Deep Learning Models
This paper aims to clarify the representational status of Deep Learning Models (DLMs). While commonly referred to as 'representations', what this entails is ambiguous due to a conflation of functional and relational conceptions of representation. This paper argues that while DLMs represent their targets in a relational sense, in general, we have no good reason to believe that DLMs encode locally semantically decomposable representations of their targets. That is, the representational capacity these models have is largely global, rather than decomposable into stable, local subrepresentations. This result has immediate implications for explainable AI (XAI) and directs attention toward exploring the global relational nature of deep learning representations and their relationship both to models more generally to understand their potential role in future scientific inquiry.
comment: 18 pages, 1 figure
♻ ☆ Confidence Intervals for Performance Estimates in Brain MRI Segmentation
Medical segmentation models are evaluated empirically. As such an evaluation is based on a limited set of example images, it is unavoidably noisy. Beyond a mean performance measure, reporting confidence intervals is thus crucial. However, this is rarely done in medical image segmentation. The width of the confidence interval depends on the test set size and on the spread of the performance measure (its standard-deviation across the test set). For classification, many test images are needed to avoid wide confidence intervals. Segmentation, however, has not been studied, and it differs by the amount of information brought by a given test image. In this paper, we study the typical confidence intervals in the context of segmentation in 3D brain magnetic resonance imaging (MRI). We carry experiments on using the standard nnU-net framework, two datasets from the Medical Decathlon challenge that concern brain MRI (hippocampus and brain tumor segmentation) and two performance measures: the Dice Similarity Coefficient and the Hausdorff distance. We show that the parametric confidence intervals are reasonable approximations of the bootstrap estimates for varying test set sizes and spread of the performance metric. Importantly, we show that the test size needed to achieve a given precision is often much lower than for classification tasks. Typically, a 1\% wide confidence interval requires about 100-200 test samples when the spread is low (standard-deviation around 3\%). More difficult segmentation tasks may lead to higher spreads and require over 1000 samples.
♻ ☆ Explaining the Impact of Training on Vision Models via Activation Clustering
This paper introduces Neuro-Activated Vision Explanations (NAVE), a method for extracting and visualizing the internal representations of vision model encoders. By clustering feature activations, NAVE provides insights into learned semantics without fine-tuning. Using object localization, we show that NAVE's concepts align with image semantics. Through extensive experiments, we analyze the impact of training strategies and architectures on encoder representation capabilities. Additionally, we apply NAVE to study training artifacts in vision transformers and reveal how weak training strategies and spurious correlations degrade model performance. Our findings establish NAVE as a valuable tool for post-hoc model inspection and improving transparency in vision models.
♻ ☆ BADTV: Unveiling Backdoor Threats in Third-Party Task Vectors
Task arithmetic in large-scale pre-trained models enables agile adaptation to diverse downstream tasks without extensive retraining. By leveraging task vectors (TVs), users can perform modular updates through simple arithmetic operations like addition and subtraction. Yet, this flexibility presents new security challenges. In this paper, we investigate how TVs are vulnerable to backdoor attacks, revealing how malicious actors can exploit them to compromise model integrity. By creating composite backdoors that are designed asymmetrically, we introduce BadTV, a backdoor attack specifically crafted to remain effective simultaneously under task learning, forgetting, and analogy operations. Extensive experiments show that BadTV achieves near-perfect attack success rates across diverse scenarios, posing a serious threat to models relying on task arithmetic. We also evaluate current defenses, finding they fail to detect or mitigate BadTV. Our results highlight the urgent need for robust countermeasures to secure TVs in real-world deployments.
♻ ☆ Convergence of Policy Mirror Descent Beyond Compatible Function Approximation
Modern policy optimization methods roughly follow the policy mirror descent (PMD) algorithmic template, for which there are by now numerous theoretical convergence results. However, most of these either target tabular environments, or can be applied effectively only when the class of policies being optimized over satisfies strong closure conditions, which is typically not the case when working with parametric policy classes in large-scale environments. In this work, we develop a theoretical framework for PMD for general policy classes where we replace the closure conditions with a strictly weaker variational gradient dominance assumption, and obtain upper bounds on the rate of convergence to the best-in-class policy. Our main result leverages a novel notion of smoothness with respect to a local norm induced by the occupancy measure of the current policy, and casts PMD as a particular instance of smooth non-convex optimization in non-Euclidean space.
♻ ☆ FALCON: Feedback-driven Adaptive Long/short-term memory reinforced Coding Optimization system
Recently, large language models (LLMs) have achieved significant progress in automated code generation. Despite their strong instruction-following capabilities, these models frequently struggled to align with user intent in coding scenarios. In particular, they were hampered by datasets that lacked diversity and failed to address specialized tasks or edge cases. Furthermore, challenges in supervised fine-tuning (SFT) and reinforcement learning from human feedback (RLHF) led to failures in generating precise, human-intent-aligned code. To tackle these challenges and improve the code generation performance for automated programming systems, we propose Feedback-driven Adaptive Long/short-term memory reinforced Coding Optimization (i.e., FALCON). FALCON is structured into two hierarchical levels. From the global level, long-term memory improves code quality by retaining and applying learned knowledge. At the local level, short-term memory allows for the incorporation of immediate feedback from compilers and AI systems. Additionally, we introduce meta-reinforcement learning with feedback rewards to solve the global-local bi-level optimization problem and enhance the model's adaptability across diverse code generation tasks. Extensive experiments demonstrate that our technique achieves state-of-the-art performance, leading other reinforcement learning methods by more than 4.5 percentage points on the MBPP benchmark and 6.1 percentage points on the Humaneval benchmark. The open-sourced code is publicly available at https://github.com/titurte/FALCON.
comment: 20 pages, 7 figures
♻ ☆ Trajectory Imputation in Multi-Agent Sports with Derivative-Accumulating Self-Ensemble
Multi-agent trajectory data collected from domains such as team sports often suffer from missing values due to various factors. While many imputation methods have been proposed for spatiotemporal data, they are not well-suited for multi-agent sports scenarios where player movements are highly dynamic and inter-agent interactions continuously evolve. To address these challenges, we propose MIDAS (Multi-agent Imputer with Derivative-Accumulating Self-ensemble), a framework that imputes multi-agent trajectories with high accuracy and physical plausibility. It jointly predicts positions, velocities, and accelerations through a Set Transformer-based neural network and generates alternative estimates by recursively accumulating predicted velocity and acceleration values. These predictions are then combined using a learnable weighted ensemble to produce final imputed trajectories. Experiments on three sports datasets demonstrate that MIDAS significantly outperforms existing baselines in both positional accuracy and physical plausibility. Lastly, we showcase use cases of MIDAS, such as approximating total distance and pass success probability, to highlight its applicability to practical downstream tasks that require complete tracking data.
♻ ☆ Inductive Moment Matching
Diffusion models and Flow Matching generate high-quality samples but are slow at inference, and distilling them into few-step models often leads to instability and extensive tuning. To resolve these trade-offs, we propose Inductive Moment Matching (IMM), a new class of generative models for one- or few-step sampling with a single-stage training procedure. Unlike distillation, IMM does not require pre-training initialization and optimization of two networks; and unlike Consistency Models, IMM guarantees distribution-level convergence and remains stable under various hyperparameters and standard model architectures. IMM surpasses diffusion models on ImageNet-256x256 with 1.99 FID using only 8 inference steps and achieves state-of-the-art 2-step FID of 1.98 on CIFAR-10 for a model trained from scratch.
♻ ☆ SVGFusion: Scalable Text-to-SVG Generation via Vector Space Diffusion
In this work, we introduce SVGFusion, a Text-to-SVG model capable of scaling to real-world SVG data without relying on text-based discrete language models or prolonged Score Distillation Sampling (SDS) optimization. The core idea of SVGFusion is to utilize a popular Text-to-Image framework to learn a continuous latent space for vector graphics. Specifically, SVGFusion comprises two key modules: a Vector-Pixel Fusion Variational Autoencoder (VP-VAE) and a Vector Space Diffusion Transformer (VS-DiT). The VP-VAE processes both SVG codes and their corresponding rasterizations to learn a continuous latent space, while the VS-DiT generates latent codes within this space based on the input text prompt. Building on the VP-VAE, we propose a novel rendering sequence modeling strategy which enables the learned latent space to capture the inherent creation logic of SVGs. This allows the model to generate SVGs with higher visual quality and more logical construction, while systematically avoiding occlusion in complex graphic compositions. Additionally, the scalability of SVGFusion can be continuously enhanced by adding more VS-DiT blocks. To effectively train and evaluate SVGFusion, we construct SVGX-Dataset, a large-scale, high-quality SVG dataset that addresses the scarcity of high-quality vector data. Extensive experiments demonstrate the superiority of SVGFusion over existing SVG generation methods, establishing a new framework for SVG content creation. Code, model, and data will be released at: https://ximinng.github.io/SVGFusionProject/
comment: project page: https://ximinng.github.io/SVGFusionProject/
♻ ☆ SCULPT: Systematic Tuning of Long Prompts
Prompt optimization is essential for effective utilization of large language models (LLMs) across diverse tasks. While existing optimization methods are effective in optimizing short prompts, they struggle with longer, more complex ones, often risking information loss and being sensitive to small perturbations. To address these challenges, we propose SCULPT (Systematic Tuning of Long Prompts), a framework that treats prompt optimization as a hierarchical tree refinement problem. SCULPT represents prompts as tree structures, enabling targeted modifications while preserving contextual integrity. It employs a Critic-Actor framework that generates reflections and applies actions to refine the prompt. Evaluations demonstrate SCULPT's effectiveness on long prompts, its robustness to adversarial perturbations, and its ability to generate high-performing prompts even without any initial human-written prompt. Compared to existing state of the art methods, SCULPT consistently improves LLM performance by preserving essential task information while applying structured refinements. Both qualitative and quantitative analyses show that SCULPT produces more stable and interpretable prompt modifications, ensuring better generalization across tasks.
♻ ☆ ModeSeq: Taming Sparse Multimodal Motion Prediction with Sequential Mode Modeling CVPR 2025
Anticipating the multimodality of future events lays the foundation for safe autonomous driving. However, multimodal motion prediction for traffic agents has been clouded by the lack of multimodal ground truth. Existing works predominantly adopt the winner-take-all training strategy to tackle this challenge, yet still suffer from limited trajectory diversity and uncalibrated mode confidence. While some approaches address these limitations by generating excessive trajectory candidates, they necessitate a post-processing stage to identify the most representative modes, a process lacking universal principles and compromising trajectory accuracy. We are thus motivated to introduce ModeSeq, a new multimodal prediction paradigm that models modes as sequences. Unlike the common practice of decoding multiple plausible trajectories in one shot, ModeSeq requires motion decoders to infer the next mode step by step, thereby more explicitly capturing the correlation between modes and significantly enhancing the ability to reason about multimodality. Leveraging the inductive bias of sequential mode prediction, we also propose the Early-Match-Take-All (EMTA) training strategy to diversify the trajectories further. Without relying on dense mode prediction or heuristic post-processing, ModeSeq considerably improves the diversity of multimodal output while attaining satisfactory trajectory accuracy, resulting in balanced performance on motion prediction benchmarks. Moreover, ModeSeq naturally emerges with the capability of mode extrapolation, which supports forecasting more behavior modes when the future is highly uncertain.
comment: CVPR 2025
♻ ☆ Fast MLE and MAPE-Based Device Activity Detection for Grant-Free Access via PSCA and PSCA-Net
Fast and accurate device activity detection is the critical challenge in grant-free access for supporting massive machine-type communications (mMTC) and ultra-reliable low-latency communications (URLLC) in 5G and beyond. The state-of-the-art methods have unsatisfactory error rates or computation times. To address these outstanding issues, we propose new maximum likelihood estimation (MLE) and maximum a posterior estimation (MAPE) based device activity detection methods for known and unknown pathloss that achieve superior error rate and computation time tradeoffs using optimization and deep learning techniques. Specifically, we investigate four non-convex optimization problems for MLE and MAPE in the two pathloss cases, with one MAPE problem being formulated for the first time. For each non-convex problem, we develop an innovative parallel iterative algorithm using the parallel successive convex approximation (PSCA) method. Each PSCA-based algorithm allows parallel computations, uses up to the objective function's second-order information, converges to the problem's stationary points, and has a low per-iteration computational complexity compared to the state-of-the-art algorithms. Then, for each PSCA-based iterative algorithm, we present a deep unrolling neural network implementation, called PSCA-Net, to further reduce the computation time. Each PSCA-Net elegantly marries the underlying PSCA-based algorithm's parallel computation mechanism with the parallelizable neural network architecture and effectively optimizes its step sizes based on vast data samples to speed up the convergence. Numerical results demonstrate that the proposed methods can significantly reduce the error rate and computation time compared to the state-of-the-art methods, revealing their significant values for grant-free access.
♻ ☆ Training Multi-Layer Binary Neural Networks With Local Binary Error Signals
Binary Neural Networks (BNNs) significantly reduce computational complexity and memory usage in machine and deep learning by representing weights and activations with just one bit. However, most existing training algorithms for BNNs rely on quantization-aware floating-point Stochastic Gradient Descent (SGD), limiting the full exploitation of binary operations to the inference phase only. In this work, we propose, for the first time, a fully binary and gradient-free training algorithm for multi-layer BNNs, eliminating the need for back-propagated floating-point gradients. Specifically, the proposed algorithm relies on local binary error signals and binary weight updates, employing integer-valued hidden weights that serve as a synaptic metaplasticity mechanism, thereby enhancing its neurobiological plausibility. The fully binary and gradient-free algorithm introduced in this paper enables the training of binary multi-layer perceptrons with binary inputs, weights, and activations, by using exclusively XNOR, Popcount, and increment/decrement operations. Experimental results on multi-class classification benchmarks show test accuracy improvements of up to +35.47% over the only existing fully binary single-layer state-of-the-art solution. Compared to full-precision SGD, our solution improves test accuracy by up to +41.31% under the same total memory demand$\unicode{x2013}$including the model, activations, and input dataset$\unicode{x2013}$while also reducing computational cost by two orders of magnitude in terms of the total number of equivalent Boolean gates. The proposed algorithm is made available to the scientific community as a public repository.
♻ ☆ Certifiably Robust Policies for Uncertain Parametric Environments
We present a data-driven approach for producing policies that are provably robust across unknown stochastic environments. Existing approaches can learn models of a single environment as an interval Markov decision processes (IMDP) and produce a robust policy with a probably approximately correct (PAC) guarantee on its performance. However these are unable to reason about the impact of environmental parameters underlying the uncertainty. We propose a framework based on parametric Markov decision processes (MDPs) with unknown distributions over parameters. We learn and analyse IMDPs for a set of unknown sample environments induced by parameters. The key challenge is then to produce meaningful performance guarantees that combine the two layers of uncertainty: (1) multiple environments induced by parameters with an unknown distribution; (2) unknown induced environments which are approximated by IMDPs. We present a novel approach based on scenario optimisation that yields a single PAC guarantee quantifying the risk level for which a specified performance level can be assured in unseen environments, plus a means to trade-off risk and performance. We implement and evaluate our framework using multiple robust policy generation methods on a range of benchmarks. We show that our approach produces tight bounds on a policy's performance with high confidence.
♻ ☆ Best Arm Identification with Resource Constraints
Motivated by the cost heterogeneity in experimentation across different alternatives, we study the Best Arm Identification with Resource Constraints (BAIwRC) problem. The agent aims to identify the best arm under resource constraints, where resources are consumed for each arm pull. We make two novel contributions. We design and analyze the Successive Halving with Resource Rationing algorithm (SH-RR). The SH-RR achieves a near-optimal non-asymptotic rate of convergence in terms of the probability of successively identifying an optimal arm. Interestingly, we identify a difference in convergence rates between the cases of deterministic and stochastic resource consumption.
♻ ☆ Energy-Based Model for Accurate Estimation of Shapley Values in Feature Attribution
Shapley value is a widely used tool in explainable artificial intelligence (XAI), as it provides a principled way to attribute contributions of input features to model outputs. However, estimation of Shapley value requires capturing conditional dependencies among all feature combinations, which poses significant challenges in complex data environments. In this article, EmSHAP (Energy-based model for Shapley value estimation), an accurate and efficient Shapley value estimation method, is proposed to estimate the expectation of Shapley contribution function under the arbitrary subset of features given the rest. By utilizing the ability of energy-based model (EBM) to model complex distributions, EmSHAP provides an effective solution for estimating the required conditional probabilities. To further improve estimation accuracy, a GRU (Gated Recurrent Unit)-coupled partition function estimation method is introduced. The GRU network captures long-term dependencies with a lightweight parameterization and maps input features into a latent space to mitigate the influence of feature ordering. Additionally, a dynamic masking mechanism is incorporated to further enhance the robustness and accuracy by progressively increasing the masking rate. Theoretical analysis on the error bound as well as application to four case studies verified the higher accuracy and better scalability of EmSHAP in contrast to competitive methods.
♻ ☆ Credit Risk Meets Large Language Models: Building a Risk Indicator from Loan Descriptions in P2P Lending
Peer-to-peer (P2P) lending connects borrowers and lenders through online platforms but suffers from significant information asymmetry, as lenders often lack sufficient data to assess borrowers' creditworthiness. This paper addresses this challenge by leveraging BERT, a Large Language Model (LLM) known for its ability to capture contextual nuances in text, to generate a risk score based on borrowers' loan descriptions using a dataset from the Lending Club platform. We fine-tune BERT to distinguish between defaulted and non-defaulted loans using the loan descriptions provided by the borrowers. The resulting BERT-generated risk score is then integrated as an additional feature into an XGBoost classifier used at the loan granting stage, where decision-makers have limited information available to guide their decisions. This integration enhances predictive performance, with improvements in balanced accuracy and AUC, highlighting the value of textual features in complementing traditional inputs. Moreover, we find that the incorporation of the BERT score alters how classification models utilize traditional input variables, with these changes varying by loan purpose. These findings suggest that BERT discerns meaningful patterns in loan descriptions, encompassing borrower-specific features, specific purposes, and linguistic characteristics. However, the inherent opacity of LLMs and their potential biases underscore the need for transparent frameworks to ensure regulatory compliance and foster trust. Overall, this study demonstrates how LLM-derived insights interact with traditional features in credit risk modeling, opening new avenues to enhance the explainability and fairness of these models.
♻ ☆ Diffusion-Driven Semantic Communication for Generative Models with Bandwidth Constraints
Diffusion models have been extensively utilized in AI-generated content (AIGC) in recent years, thanks to the superior generation capabilities. Combining with semantic communications, diffusion models are used for tasks such as denoising, data reconstruction, and content generation. However, existing diffusion-based generative models do not consider the stringent bandwidth limitation, which limits its application in wireless communication. This paper introduces a diffusion-driven semantic communication framework with advanced VAE-based compression for bandwidth-constrained generative model. Our designed architecture utilizes the diffusion model, where the signal transmission process through the wireless channel acts as the forward process in diffusion. To reduce bandwidth requirements, we incorporate a downsampling module and a paired upsampling module based on a variational auto-encoder with reparameterization at the receiver to ensure that the recovered features conform to the Gaussian distribution. Furthermore, we derive the loss function for our proposed system and evaluate its performance through comprehensive experiments. Our experimental results demonstrate significant improvements in pixel-level metrics such as peak signal to noise ratio (PSNR) and semantic metrics like learned perceptual image patch similarity (LPIPS). These enhancements are more profound regarding the compression rates and SNR compared to deep joint source-channel coding (DJSCC).
comment: accepted to IEEE for possible publication
♻ ☆ SupReMix: Supervised Contrastive Learning for Medical Imaging Regression with Mixup
In medical image analysis, regression plays a critical role in computer-aided diagnosis. It enables quantitative measurements such as age prediction from structural imaging, cardiac function quantification, and molecular measurement from PET scans. While deep learning has shown promise for these tasks, most approaches focus solely on optimizing regression loss or model architecture, neglecting the quality of learned feature representations which are crucial for robust clinical predictions. Directly applying representation learning techniques designed for classification to regression often results in fragmented representations in the latent space, yielding sub-optimal performance. In this paper, we argue that the potential of contrastive learning for medical image regression has been overshadowed due to the neglect of two crucial aspects: ordinality-awareness and hardness. To address these challenges, we propose Supervised Contrastive Learning for Medical Imaging Regression with Mixup (SupReMix). It takes anchor-inclusive mixtures (mixup of the anchor and a distinct negative sample) as hard negative pairs and anchor-exclusive mixtures (mixup of two distinct negative samples) as hard positive pairs at the embedding level. This strategy formulates harder contrastive pairs by integrating richer ordinal information. Through theoretical analysis and extensive experiments on six datasets spanning MRI, X-ray, ultrasound, and PET modalities, we demonstrate that SupReMix fosters continuous ordered representations, significantly improving regression performance.
comment: The first two authors equally contributed to this work. Previously titled "Mixup Your Own Pair", content extended and revised
♻ ☆ Are Images Indistinguishable to Humans Also Indistinguishable to Classifiers? CVPR 2025
The ultimate goal of generative models is to perfectly capture the data distribution. For image generation, common metrics of visual quality (e.g., FID) and the perceived truthfulness of generated images seem to suggest that we are nearing this goal. However, through distribution classification tasks, we reveal that, from the perspective of neural network-based classifiers, even advanced diffusion models are still far from this goal. Specifically, classifiers are able to consistently and effortlessly distinguish real images from generated ones across various settings. Moreover, we uncover an intriguing discrepancy: classifiers can easily differentiate between diffusion models with comparable performance (e.g., U-ViT-H vs. DiT-XL), but struggle to distinguish between models within the same family but of different scales (e.g., EDM2-XS vs. EDM2-XXL). Our methodology carries several important implications. First, it naturally serves as a diagnostic tool for diffusion models by analyzing specific features of generated data. Second, it sheds light on the model autophagy disorder and offers insights into the use of generated data: augmenting real data with generated data is more effective than replacing it. Third, classifier guidance can significantly enhance the realism of generated images.
comment: Accepted to CVPR 2025
♻ ☆ Optical-Flow Guided Prompt Optimization for Coherent Video Generation CVPR 2025
While text-to-video diffusion models have made significant strides, many still face challenges in generating videos with temporal consistency. Within diffusion frameworks, guidance techniques have proven effective in enhancing output quality during inference; however, applying these methods to video diffusion models introduces additional complexity of handling computations across entire sequences. To address this, we propose a novel framework called MotionPrompt that guides the video generation process via optical flow. Specifically, we train a discriminator to distinguish optical flow between random pairs of frames from real videos and generated ones. Given that prompts can influence the entire video, we optimize learnable token embeddings during reverse sampling steps by using gradients from a trained discriminator applied to random frame pairs. This approach allows our method to generate visually coherent video sequences that closely reflect natural motion dynamics, without compromising the fidelity of the generated content. We demonstrate the effectiveness of our approach across various models.
comment: CVPR 2025 (poster); project page: https://motionprompt.github.io/
♻ ☆ Effective and Efficient Masked Image Generation Models
Although masked image generation models and masked diffusion models are designed with different motivations and objectives, we observe that they can be unified within a single framework. Building upon this insight, we carefully explore the design space of training and sampling, identifying key factors that contribute to both performance and efficiency. Based on the improvements observed during this exploration, we develop our model, referred to as eMIGM. Empirically, eMIGM demonstrates strong performance on ImageNet generation, as measured by Fr\'echet Inception Distance (FID). In particular, on ImageNet 256x256, with similar number of function evaluations (NFEs) and model parameters, eMIGM outperforms the seminal VAR. Moreover, as NFE and model parameters increase, eMIGM achieves performance comparable to the state-of-the-art continuous diffusion models while requiring less than 40% of the NFE. Additionally, on ImageNet 512x512, with only about 60% of the NFE, eMIGM outperforms the state-of-the-art continuous diffusion models. Code is available at https://github.com/ML-GSAI/eMIGM.
♻ ☆ MAB-Based Channel Scheduling for Asynchronous Federated Learning in Non-Stationary Environments
Federated learning enables distributed model training across clients without raw data exchange, but in wireless implementations, frequent parameter updates cause high communication overhead. Existing research often assumes known channel state information (CSI) or stationary channels, though practical wireless channels are non-stationary due to fading, user mobility, and attacks, leading to unpredictable transmission failures and exacerbating client staleness, which hampers model convergence. To tackle these challenges, we propose an asynchronous federated learning scheduling framework for non-stationary channels that aims to reduce client staleness while enhancing communication efficiency and fairness. Our framework considers two scenarios: extremely non-stationary and piecewise-stationary channels. Age of Information (AoI) quantifies client staleness under these conditions. We conduct convergence analysis to examine the impact of AoI and per-round client participation on learning performance and formulate the scheduling problem as a multi-armed bandit (MAB) problem. We derive theoretical lower bounds on AoI regret and develop scheduling strategies based on GLR-CUCB and M-exp3 algorithms, including upper bounds on AoI regret. To address imbalanced client updates, we propose an adaptive matching strategy that incorporates marginal utility and fairness considerations. Simulation results show that our algorithm achieves sub-linear AoI regret, accelerates convergence, and promotes fairer aggregation.
♻ ☆ Knowledge Rumination for Client Utility Evaluation in Heterogeneous Federated Learning
Federated Learning (FL) allows several clients to cooperatively train machine learning models without disclosing the raw data. In practical applications, asynchronous FL (AFL) can address the straggler effect compared to synchronous FL. However, Non-IID data and stale models pose significant challenges to AFL, as they can diminish the practicality of the global model and even lead to training failures. In this work, we propose a novel AFL framework called Federated Historical Learning (FedHist), which effectively addresses the challenges posed by both Non-IID data and gradient staleness based on the concept of knowledge rumination. FedHist enhances the stability of local gradients by performing weighted fusion with historical global gradients cached on the server. Relying on hindsight, it assigns aggregation weights to each participant in a multi-dimensional manner during each communication round. To further enhance the efficiency and stability of the training process, we introduce an intelligent $\ell_2$-norm amplification scheme, which dynamically regulates the learning progress based on the $\ell_2$-norms of the submitted gradients. Extensive experiments indicate FedHist outperforms state-of-the-art methods in terms of convergence performance and test accuracy.
♻ ☆ Physics-Aware Combinatorial Assembly Sequence Planning using Data-free Action Masking
Combinatorial assembly uses standardized unit primitives to build objects that satisfy user specifications. This paper studies assembly sequence planning (ASP) for physical combinatorial assembly. Given the shape of the desired object, the goal is to find a sequence of actions for placing unit primitives to build the target object. In particular, we aim to ensure the planned assembly sequence is physically executable. However, ASP for combinatorial assembly is particularly challenging due to its combinatorial nature. To address the challenge, we employ deep reinforcement learning to learn a construction policy for placing unit primitives sequentially to build the desired object. Specifically, we design an online physics-aware action mask that filters out invalid actions, which effectively guides policy learning and ensures violation-free deployment. In the end, we apply the proposed method to Lego assembly with more than 250 3D structures. The experiment results demonstrate that the proposed method plans physically valid assembly sequences to build all structures, achieving a $100\%$ success rate, whereas the best comparable baseline fails more than $40$ structures. Our implementation is available at \url{https://github.com/intelligent-control-lab/PhysicsAwareCombinatorialASP}.
♻ ☆ DPFAGA-Dynamic Power Flow Analysis and Fault Characteristics: A Graph Attention Neural Network
We propose the joint graph attention neural network (GAT), clustering with adaptive neighbors (CAN) and probabilistic graphical model for dynamic power flow analysis and fault characteristics. In fact, computational efficiency is the main focus to enhance, whilst we ensure the performance accuracy at the accepted level. Note that Machine Learning (ML) based schemes have a requirement of sufficient labeled data during training, which is not easily satisfied in practical applications. Also, there are unknown data due to new arrived measurements or incompatible smart devices in complex smart grid systems. These problems would be resolved by our proposed GAT based framework, which models the label dependency between the network data and learns object representations such that it could achieve the semi-supervised fault diagnosis. To create the joint label dependency, we develop the graph construction from the raw acquired signals by using CAN. Next, we develop the probabilistic graphical model of Markov random field for graph representation, which supports for the GAT based framework. We then evaluate the proposed framework in the use-case application in smart grid and make a fair comparison to the existing methods.
comment: The 2025 International Conference on the AI Revolution: Research, Ethics, and Society (AIR-RES 2025)
♻ ☆ Exploring the Limits of Zero Shot Vision Language Models for Hate Meme Detection: The Vulnerabilities and their Interpretations
There is a rapid increase in the use of multimedia content in current social media platforms. One of the highly popular forms of such multimedia content are memes. While memes have been primarily invented to promote funny and buoyant discussions, malevolent users exploit memes to target individuals or vulnerable communities, making it imperative to identify and address such instances of hateful memes. Thus social media platforms are in dire need for active moderation of such harmful content. While manual moderation is extremely difficult due to the scale of such content, automatic moderation is challenged by the need of good quality annotated data to train hate meme detection algorithms. This makes a perfect pretext for exploring the power of modern day vision language models (VLMs) that have exhibited outstanding performance across various tasks. In this paper we study the effectiveness of VLMs in handling intricate tasks such as hate meme detection in a completely zero-shot setting so that there is no dependency on annotated data for the task. We perform thorough prompt engineering and query state-of-the-art VLMs using various prompt types to detect hateful/harmful memes. We further interpret the misclassification cases using a novel superpixel based occlusion method. Finally we show that these misclassifications can be neatly arranged into a typology of error classes the knowledge of which should enable the design of better safety guardrails in future.
♻ ☆ DexDiffuser: Interaction-aware Diffusion Planning for Adaptive Dexterous Manipulation
Dexterous manipulation with contact-rich interactions is crucial for advanced robotics. While recent diffusion-based planning approaches show promise for simple manipulation tasks, they often produce unrealistic ghost states (e.g., the object automatically moves without hand contact) or lack adaptability when handling complex sequential interactions. In this work, we introduce DexDiffuser, an interaction-aware diffusion planning framework for adaptive dexterous manipulation. DexDiffuser models joint state-action dynamics through a dual-phase diffusion process which consists of pre-interaction contact alignment and post-contact goal-directed control, enabling goal-adaptive generalizable dexterous manipulation. Additionally, we incorporate dynamics model-based dual guidance and leverage large language models for automated guidance function generation, enhancing generalizability for physical interactions and facilitating diverse goal adaptation through language cues. Experiments on physical interaction tasks such as door opening, pen and block re-orientation, object relocation, and hammer striking demonstrate DexDiffuser's effectiveness on goals outside training distributions, achieving over twice the average success rate (59.2% vs. 29.5%) compared to existing methods. Our framework achieves an average of 70.7% success rate on goal adaptive dexterous tasks, highlighting its robustness and flexibility in contact-rich manipulation.
comment: Camera ready version. 27 pages. Project page: https://dexdiffuser.github.io/
♻ ☆ DDEQs: Distributional Deep Equilibrium Models through Wasserstein Gradient Flows AISTATS 2025
Deep Equilibrium Models (DEQs) are a class of implicit neural networks that solve for a fixed point of a neural network in their forward pass. Traditionally, DEQs take sequences as inputs, but have since been applied to a variety of data. In this work, we present Distributional Deep Equilibrium Models (DDEQs), extending DEQs to discrete measure inputs, such as sets or point clouds. We provide a theoretically grounded framework for DDEQs. Leveraging Wasserstein gradient flows, we show how the forward pass of the DEQ can be adapted to find fixed points of discrete measures under permutation-invariance, and derive adequate network architectures for DDEQs. In experiments, we show that they can compete with state-of-the-art models in tasks such as point cloud classification and point cloud completion, while being significantly more parameter-efficient.
comment: 39 pages, 17 figures. To be published in AISTATS 2025
♻ ☆ Word Sense Disambiguation in Persian: Can AI Finally Get It Right?
Homograph disambiguation, the task of distinguishing words with identical spellings but different meanings, poses a substantial challenge in natural language processing. In this study, we introduce a novel dataset tailored for Persian homograph disambiguation. Our work encompasses a thorough exploration of various embeddings, evaluated through the cosine similarity method and their efficacy in downstream tasks like classification. Our investigation entails training a diverse array of lightweight machine learning and deep learning models for phonograph disambiguation. We scrutinize the models' performance in terms of Accuracy, Recall, and F1 Score, thereby gaining insights into their respective strengths and limitations. The outcomes of our research underscore three key contributions. First, we present a newly curated Persian dataset, providing a solid foundation for future research in homograph disambiguation. Second, our comparative analysis of embeddings highlights their utility in different contexts, enriching the understanding of their capabilities. Third, by training and evaluating a spectrum of models, we extend valuable guidance for practitioners in selecting suitable strategies for homograph disambiguation tasks. In summary, our study unveils a new dataset, scrutinizes embeddings through diverse perspectives, and benchmarks various models for homograph disambiguation. These findings empower researchers and practitioners to navigate the intricate landscape of homograph-related challenges effectively.
♻ ☆ Autonomous Data Selection with Zero-shot Generative Classifiers for Mathematical Texts
We present Autonomous Data Selection (AutoDS), a method that leverages base language models themselves as zero-shot "generative classifiers" to automatically curate high-quality mathematical texts. Unlike prior approaches that require human annotations or training a dedicated data filter, AutoDS relies solely on a model's logits to determine whether a given passage is mathematically informative and educational. By integrating AutoDS into a continual pretraining pipeline, we substantially boost downstream performance on challenging math benchmarks (MATH, GSM8K, and BBH) while using far fewer tokens than previous methods. Empirically, our approach achieves roughly a twofold improvement in pretraining token efficiency over strong baselines, underscoring the potential of self-directed data selection in enhancing mathematical reasoning. We release our curated AutoMathText dataset to facilitate future research in automated domain-specific data curation. The AutoMathText dataset is available at https://huggingface.co/datasets/math-ai/AutoMathText. The code is available at https://github.com/yifanzhang-pro/AutoMathText.
comment: 24 pages, 8 figures. arXiv admin note: text overlap with arXiv:0808.2664, arXiv:0806.2159, arXiv:1703.08834, arXiv:math/0610707 by other authors
♻ ☆ Toward Conditional Distribution Calibration in Survival Prediction NeurIPS 2024
Survival prediction often involves estimating the time-to-event distribution from censored datasets. Previous approaches have focused on enhancing discrimination and marginal calibration. In this paper, we highlight the significance of conditional calibration for real-world applications -- especially its role in individual decision-making. We propose a method based on conformal prediction that uses the model's predicted individual survival probability at that instance's observed time. This method effectively improves the model's marginal and conditional calibration, without compromising discrimination. We provide asymptotic theoretical guarantees for both marginal and conditional calibration and test it extensively across 15 diverse real-world datasets, demonstrating the method's practical effectiveness and versatility in various settings.
comment: Accepted to NeurIPS 2024. 41 pages, 23 figures
Computation and Language 14
☆ Think Before Refusal : Triggering Safety Reflection in LLMs to Mitigate False Refusal Behavior
Recent advancements in large language models (LLMs) have demonstrated that fine-tuning and human alignment can render LLMs harmless. In practice, such "harmlessness" behavior is mainly achieved by training models to reject harmful requests, such as "Explain how to burn down my neighbor's house", where the model appropriately declines to respond. However, this approach can inadvertently result in false refusal, where models reject benign queries as well, such as "Tell me how to kill a Python process". In this work, we demonstrate that prompting safety reflection before generating a response can mitigate false refusal behavior. Building on this finding, we introduce the Think-Before-Refusal (TBR) schema and conduct safety-aware instruction fine-tuning incorporating safety reflection. In an ablation study across 15 pre-trained models, we show that models fine-tuned with safety reflection significantly reduce false refusal behavior while maintaining safety and overall performance compared to those fine-tuned without safety reflection.
comment: 18 pages, 23 figures
☆ Satisfactory Medical Consultation based on Terminology-Enhanced Information Retrieval and Emotional In-Context Learning
Recent advancements in Large Language Models (LLMs) have marked significant progress in understanding and responding to medical inquiries. However, their performance still falls short of the standards set by professional consultations. This paper introduces a novel framework for medical consultation, comprising two main modules: Terminology-Enhanced Information Retrieval (TEIR) and Emotional In-Context Learning (EICL). TEIR ensures implicit reasoning through the utilization of inductive knowledge and key terminology retrieval, overcoming the limitations of restricted domain knowledge in public databases. Additionally, this module features capabilities for processing long context. The EICL module aids in generating sentences with high attribute relevance by memorizing semantic and attribute information from unlabelled corpora and applying controlled retrieval for the required information. Furthermore, a dataset comprising 803,564 consultation records was compiled in China, significantly enhancing the model's capability for complex dialogues and proactive inquiry initiation. Comprehensive experiments demonstrate the proposed method's effectiveness in extending the context window length of existing LLMs. The experimental outcomes and extensive data validate the framework's superiority over five baseline models in terms of BLEU and ROUGE performance metrics, with substantial leads in certain capabilities. Notably, ablation studies confirm the significance of the TEIR and EICL components. In addition, our new framework has the potential to significantly improve patient satisfaction in real clinical consulting situations.
comment: The 46th European Conference on Information Retrieval Workshop
☆ Enhancing Retrieval Systems with Inference-Time Logical Reasoning
Traditional retrieval methods rely on transforming user queries into vector representations and retrieving documents based on cosine similarity within an embedding space. While efficient and scalable, this approach often fails to handle complex queries involving logical constructs such as negations, conjunctions, and disjunctions. In this paper, we propose a novel inference-time logical reasoning framework that explicitly incorporates logical reasoning into the retrieval process. Our method extracts logical reasoning structures from natural language queries and then composes the individual cosine similarity scores to formulate the final document scores. This approach enables the retrieval process to handle complex logical reasoning without compromising computational efficiency. Our results on both synthetic and real-world benchmarks demonstrate that the proposed method consistently outperforms traditional retrieval methods across different models and datasets, significantly improving retrieval performance for complex queries.
☆ Feather-SQL: A Lightweight NL2SQL Framework with Dual-Model Collaboration Paradigm for Small Language Models
Natural Language to SQL (NL2SQL) has seen significant advancements with large language models (LLMs). However, these models often depend on closed-source systems and high computational resources, posing challenges in data privacy and deployment. In contrast, small language models (SLMs) struggle with NL2SQL tasks, exhibiting poor performance and incompatibility with existing frameworks. To address these issues, we introduce Feather-SQL, a new lightweight framework tailored for SLMs. Feather-SQL improves SQL executability and accuracy through 1) schema pruning and linking, 2) multi-path and multi-candidate generation. Additionally, we introduce the 1+1 Model Collaboration Paradigm, which pairs a strong general-purpose chat model with a fine-tuned SQL specialist, combining strong analytical reasoning with high-precision SQL generation. Experimental results on BIRD demonstrate that Feather-SQL improves NL2SQL performance on SLMs, with around 10% boost for models without fine-tuning. The proposed paradigm raises the accuracy ceiling of SLMs to 54.76%, highlighting its effectiveness.
☆ ParsiPy: NLP Toolkit for Historical Persian Texts in Python
The study of historical languages presents unique challenges due to their complex orthographic systems, fragmentary textual evidence, and the absence of standardized digital representations of text in those languages. Tackling these challenges needs special NLP digital tools to handle phonetic transcriptions and analyze ancient texts. This work introduces ParsiPy, an NLP toolkit designed to facilitate the analysis of historical Persian languages by offering modules for tokenization, lemmatization, part-of-speech tagging, phoneme-to-transliteration conversion, and word embedding. We demonstrate the utility of our toolkit through the processing of Parsig (Middle Persian) texts, highlighting its potential for expanding computational methods in the study of historical languages. Through this work, we contribute to computational philology, offering tools that can be adapted for the broader study of ancient texts and their digital preservation.
comment: 13 pages, 6 figure, accepted into Second Workshop on Ancient Language Processing (ALP2025)
♻ ☆ Open Models, Closed Minds? On Agents Capabilities in Mimicking Human Personalities through Open Large Language Models AAAI 2025
The emergence of unveiling human-like behaviors in Large Language Models (LLMs) has led to a closer connection between NLP and human psychology. Scholars have been studying the inherent personalities exhibited by LLMs and attempting to incorporate human traits and behaviors into them. However, these efforts have primarily focused on commercially-licensed LLMs, neglecting the widespread use and notable advancements seen in Open LLMs. This work aims to address this gap by employing a set of 12 LLM Agents based on the most representative Open models and subject them to a series of assessments concerning the Myers-Briggs Type Indicator (MBTI) test and the Big Five Inventory (BFI) test. Our approach involves evaluating the intrinsic personality traits of Open LLM agents and determining the extent to which these agents can mimic human personalities when conditioned by specific personalities and roles. Our findings unveil that $(i)$ each Open LLM agent showcases distinct human personalities; $(ii)$ personality-conditioned prompting produces varying effects on the agents, with only few successfully mirroring the imposed personality, while most of them being ``closed-minded'' (i.e., they retain their intrinsic traits); and $(iii)$ combining role and personality conditioning can enhance the agents' ability to mimic human personalities. Our work represents a step up in understanding the dense relationship between NLP and human psychology through the lens of Open LLMs.
comment: Accepted and presented at the AAAI 2025 Conference. CHANGES in version v2: (i) Enhanced methodology and evaluation based on BFI in addition to MBTI, with expanded set of LLM agents; (ii) author list changed w.r.t. version (v1), see Acknowledgements
♻ ☆ We're Calling an Intervention: Exploring Fundamental Hurdles in Adapting Language Models to Nonstandard Text
We present a suite of experiments that allow us to understand the underlying challenges of language model adaptation to nonstandard text. We do so by designing interventions that approximate core features of user-generated text and their interactions with existing biases of language models. Applying our interventions during language model adaptation to nonstandard text variations, we gain important insights into when such adaptation is successful, as well as the aspects of text variation and noise that are particularly difficult for language models to handle. For instance, on text with character-level variation, out-of-the-box performance improves even with a few additional training examples but approaches a plateau, suggesting that more data is not the solution. In contrast, on text with variation involving new words or meanings, far more data is needed, but it leads to a massive breakthrough in performance. Our findings reveal that existing models lack the necessary infrastructure to handle diverse forms of nonstandard text, guiding the development of more resilient language modeling techniques. We make the code for our interventions, which can be applied to any English text data, publicly available.
comment: Accepted for publication at W-NUT 2025
♻ ☆ DILA: Dictionary Label Attention for Mechanistic Interpretability in High-dimensional Multi-label Medical Coding Prediction
Predicting high-dimensional or extreme multilabels, such as in medical coding, requires both accuracy and interpretability. Existing works often rely on local interpretability methods, failing to provide comprehensive explanations of the overall mechanism behind each label prediction within a multilabel set. We propose a mechanistic interpretability module called DIctionary Label Attention (\method) that disentangles uninterpretable dense embeddings into a sparse embedding space, where each nonzero element (a dictionary feature) represents a globally learned medical concept. Through human evaluations, we show that our sparse embeddings are more human understandable than its dense counterparts by at least 50 percent. Our automated dictionary feature identification pipeline, leveraging large language models (LLMs), uncovers thousands of learned medical concepts by examining and summarizing the highest activating tokens for each dictionary feature. We represent the relationships between dictionary features and medical codes through a sparse interpretable matrix, enhancing the mechanistic and global understanding of the model's predictions while maintaining competitive performance and scalability without extensive human annotation.
comment: https://proceedings.mlr.press/v259/
♻ ☆ Plurals: A System for Guiding LLMs Via Simulated Social Ensembles
Recent debates raised concerns that language models may favor certain viewpoints. But what if the solution is not to aim for a 'view from nowhere' but rather to leverage different viewpoints? We introduce Plurals, a system and Python library for pluralistic AI deliberation. Plurals consists of Agents (LLMs, optionally with personas) which deliberate within customizable Structures, with Moderators overseeing deliberation. Plurals is a generator of simulated social ensembles. Plurals integrates with government datasets to create nationally representative personas, includes deliberation templates inspired by deliberative democracy, and allows users to customize both information-sharing structures and deliberation behavior within Structures. Six case studies demonstrate fidelity to theoretical constructs and efficacy. Three randomized experiments show simulated focus groups produced output resonant with an online sample of the relevant audiences (chosen over zero-shot generation in 75% of trials). Plurals is both a paradigm and a concrete system for pluralistic AI. The Plurals library is available at https://github.com/josh-ashkinaze/plurals and will be continually updated.
comment: CHI 2025
♻ ☆ Beyond Label Attention: Transparency in Language Models for Automated Medical Coding via Dictionary Learning
Medical coding, the translation of unstructured clinical text into standardized medical codes, is a crucial but time-consuming healthcare practice. Though large language models (LLM) could automate the coding process and improve the efficiency of such tasks, interpretability remains paramount for maintaining patient trust. Current efforts in interpretability of medical coding applications rely heavily on label attention mechanisms, which often leads to the highlighting of extraneous tokens irrelevant to the ICD code. To facilitate accurate interpretability in medical language models, this paper leverages dictionary learning that can efficiently extract sparsely activated representations from dense language model embeddings in superposition. Compared with common label attention mechanisms, our model goes beyond token-level representations by building an interpretable dictionary which enhances the mechanistic-based explanations for each ICD code prediction, even when the highlighted tokens are medically irrelevant. We show that dictionary features can steer model behavior, elucidate the hidden meanings of upwards of 90% of medically irrelevant tokens, and are human interpretable.
comment: https://aclanthology.org/2024.emnlp-main.500/
♻ ☆ Smart Routing: Cost-Effective Multi-LLM Serving in AIOS
As large language models (LLMs) are increasingly deployed as service endpoints in systems, the surge in query volume creates significant scheduling challenges. Existing scheduling frameworks mainly target at latency optimization while neglecting the capability of LLMs to serve different level of queries, which could lead to computational resource waste. For example, those simple queries can be safely handled by small, fast and cheap LLMs, while those complex and difficult queries need to be handled by large, slow, and expensive LLMs. This paper addresses this challenge by proposing an efficient capability-cost coordinated scheduling framework, ECCOS, for multi-LLM serving, which explicitly constrains response quality and workload to optimize LLM inference cost. Specifically, it introduces the two-stage scheduling by designing a multi-objective predictor and a constrained optimizer. The predictor estimates both model capabilities and computational costs through training-based and retrieval-based approaches, while the optimizer determines cost-optimal assignments under quality and workload constraints. It also introduces QAServe, a dataset for sample-wise response quality and costs collected by zero-shot prompting different LLMs on knowledge QA and mathematical reasoning. Extensive experiments demonstrate that ECCOS improves success rates by 6.30% while reducing costs by 10.15% compared to existing methods, consuming less than 0.5% of LLM response time. The code is available at: https://github.com/agiresearch/ECCOS, and the proposed smart routing mechanism has been integrated into AIOS, the AI Agent Operating System, at https://github.com/agiresearch/AIOS.
♻ ☆ Dynamic Noise Preference Optimization for LLM Self-Improvement via Synthetic Data
Although LLMs have achieved significant success, their reliance on large volumes of human-annotated data has limited their potential for further scaling. In this situation, utilizing self-generated synthetic data has become crucial for fine-tuning LLMs without extensive human annotation. However, current methods often fail to ensure consistent improvements across iterations, with performance stagnating after only minimal updates. To overcome these challenges, we introduce Dynamic Noise Preference Optimization (DNPO). DNPO employs a dynamic sample labeling mechanism to construct preference pairs for training and introduces controlled, trainable noise into the preference optimization process. Our approach effectively prevents stagnation and enables continuous improvement. In experiments with Zephyr-7B, DNPO consistently outperforms existing methods, showing an average performance boost of 2.6% across multiple benchmarks. Additionally, DNPO shows a significant improvement in model-generated data quality, with a 29.4% win-loss rate gap compared to the baseline in GPT-4 evaluations. This highlights its effectiveness in enhancing model performance through iterative refinement.
♻ ☆ Not All LLM-Generated Data Are Equal: Rethinking Data Weighting in Text Classification ICLR 2025
Synthetic data augmentation via large language models (LLMs) allows researchers to leverage additional training data, thus enhancing the performance of downstream tasks, especially when real-world data is scarce. However, the generated data can deviate from the real-world data, and this misalignment can bring deficient outcomes while applying the trained model to applications. Therefore, we proposed efficient weighted-loss approaches to align synthetic data with real-world distribution by emphasizing high-quality and diversified data generated by LLMs with using merely a little real-world data. We empirically assessed the effectiveness of our method on multiple text classification tasks, and the results showed leveraging our approaches on a BERT-level model robustly outperformed standard cross-entropy and other data weighting approaches, providing potential solutions to effectively leveraging synthetic data from any suitable data generator for model training.
comment: ICLR 2025 camera ready
♻ ☆ Bias Evaluation and Mitigation in Retrieval-Augmented Medical Question-Answering Systems
Medical Question Answering systems based on Retrieval Augmented Generation is promising for clinical decision support because they can integrate external knowledge, thus reducing inaccuracies inherent in standalone large language models (LLMs). However, these systems may unintentionally propagate or amplify biases associated with sensitive demographic attributes like race, gender, and socioeconomic factors. This study systematically evaluates demographic biases within medical RAG pipelines across multiple QA benchmarks, including MedQA, MedMCQA, MMLU, and EquityMedQA. We quantify disparities in retrieval consistency and answer correctness by generating and analyzing queries sensitive to demographic variations. We further implement and compare several bias mitigation strategies to address identified biases, including Chain of Thought reasoning, Counterfactual filtering, Adversarial prompt refinement, and Majority Vote aggregation. Experimental results reveal significant demographic disparities, highlighting that Majority Vote aggregation notably improves accuracy and fairness metrics. Our findings underscore the critical need for explicitly fairness-aware retrieval methods and prompt engineering strategies to develop truly equitable medical QA systems.
Information Retrieval 4
☆ Satisfactory Medical Consultation based on Terminology-Enhanced Information Retrieval and Emotional In-Context Learning
Recent advancements in Large Language Models (LLMs) have marked significant progress in understanding and responding to medical inquiries. However, their performance still falls short of the standards set by professional consultations. This paper introduces a novel framework for medical consultation, comprising two main modules: Terminology-Enhanced Information Retrieval (TEIR) and Emotional In-Context Learning (EICL). TEIR ensures implicit reasoning through the utilization of inductive knowledge and key terminology retrieval, overcoming the limitations of restricted domain knowledge in public databases. Additionally, this module features capabilities for processing long context. The EICL module aids in generating sentences with high attribute relevance by memorizing semantic and attribute information from unlabelled corpora and applying controlled retrieval for the required information. Furthermore, a dataset comprising 803,564 consultation records was compiled in China, significantly enhancing the model's capability for complex dialogues and proactive inquiry initiation. Comprehensive experiments demonstrate the proposed method's effectiveness in extending the context window length of existing LLMs. The experimental outcomes and extensive data validate the framework's superiority over five baseline models in terms of BLEU and ROUGE performance metrics, with substantial leads in certain capabilities. Notably, ablation studies confirm the significance of the TEIR and EICL components. In addition, our new framework has the potential to significantly improve patient satisfaction in real clinical consulting situations.
comment: The 46th European Conference on Information Retrieval Workshop
☆ Explainable identification of similarities between entities for discovery in large text
With the availability of virtually infinite number text documents in digital format, automatic comparison of textual data is essential for extracting meaningful insights that are difficult to identify manually. Many existing tools, including AI and large language models, struggle to provide precise and explainable insights into textual similarities. In many cases they determine the similarity between documents as reflected by the text, rather than the similarities between the subjects being discussed in these documents. This study addresses these limitations by developing an n-gram analysis framework designed to compare documents automatically and uncover explainable similarities. A scoring formula is applied to assigns each of the n-grams with a weight, where the weight is higher when the n-grams are more frequent in both documents, but is penalized when the n-grams are more frequent in the English language. Visualization tools like word clouds enhance the representation of these patterns, providing clearer insights. The findings demonstrate that this framework effectively uncovers similarities between text documents, offering explainable insights that are often difficult to identify manually. This non-parametric approach provides a deterministic solution for identifying similarities across various fields, including biographies, scientific literature, historical texts, and more. Code for the method is publicly available.
comment: Future Internet, accepted
♻ ☆ Set-Encoder: Permutation-Invariant Inter-Passage Attention for Listwise Passage Re-Ranking with Cross-Encoders ECIR'25
Existing cross-encoder models can be categorized as pointwise, pairwise, or listwise. Pairwise and listwise models allow passage interactions, which typically makes them more effective than pointwise models but less efficient and less robust to input passage order permutations. To enable efficient permutation-invariant passage interactions during re-ranking, we propose a new cross-encoder architecture with inter-passage attention: the Set-Encoder. In experiments on TREC Deep Learning and TIREx, the Set-Encoder is as effective as state-of-the-art listwise models while being more efficient and invariant to input passage order permutations. Compared to pointwise models, the Set-Encoder is particularly more effective when considering inter-passage information, such as novelty, and retains its advantageous properties compared to other listwise models. Our code is publicly available at https://github.com/webis-de/ECIR-25.
comment: Accepted at ECIR'25
♻ ☆ Rank-DistiLLM: Closing the Effectiveness Gap Between Cross-Encoders and LLMs for Passage Re-Ranking
Cross-encoders distilled from large language models (LLMs) are often more effective re-rankers than cross-encoders fine-tuned on manually labeled data. However, distilled models do not match the effectiveness of their teacher LLMs. We hypothesize that this effectiveness gap is due to the fact that previous work has not applied the best-suited methods for fine-tuning cross-encoders on manually labeled data (e.g., hard-negative sampling, deep sampling, and listwise loss functions). To close this gap, we create a new dataset, Rank-DistiLLM. Cross-encoders trained on Rank-DistiLLM achieve the effectiveness of LLMs while being up to 173 times faster and 24 times more memory efficient. Our code and data is available at https://github.com/webis-de/ECIR-25.
Multimedia 3
☆ Serial Low-rank Adaptation of Vision Transformer
Fine-tuning large pre-trained vision foundation models in a parameter-efficient manner is critical for downstream vision tasks, considering the practical constraints of computational and storage costs. Low-rank adaptation (LoRA) is a well-established technique in this domain, achieving impressive efficiency by reducing the parameter space to a low-rank form. However, developing more advanced low-rank adaptation methods to reduce parameters and memory requirements remains a significant challenge in resource-constrained application scenarios. In this study, we consider on top of the commonly used vision transformer and propose Serial LoRA, a novel LoRA variant that introduces a shared low-rank matrix serially composite with the attention mechanism. Such a design extracts the underlying commonality of parameters in adaptation, significantly reducing redundancy. Notably, Serial LoRA uses only 1/4 parameters of LoRA but achieves comparable performance in most cases. We conduct extensive experiments on a range of vision foundation models with the transformer structure, and the results confirm consistent superiority of our method.
☆ RDTF: Resource-efficient Dual-mask Training Framework for Multi-frame Animated Sticker Generation
Recently, great progress has been made in video generation technology, attracting the widespread attention of scholars. To apply this technology to downstream applications under resource-constrained conditions, researchers usually fine-tune the pre-trained models based on parameter-efficient tuning methods such as Adapter or Lora. Although these methods can transfer the knowledge from the source domain to the target domain, fewer training parameters lead to poor fitting ability, and the knowledge from the source domain may lead to the inference process deviating from the target domain. In this paper, we argue that under constrained resources, training a smaller video generation model from scratch using only million-level samples can outperform parameter-efficient tuning on larger models in downstream applications: the core lies in the effective utilization of data and curriculum strategy. Take animated sticker generation (ASG) as a case study, we first construct a discrete frame generation network for stickers with low frame rates, ensuring that its parameters meet the requirements of model training under constrained resources. In order to provide data support for models trained from scratch, we come up with a dual-mask based data utilization strategy, which manages to improve the availability and expand the diversity of limited data. To facilitate convergence under dual-mask situation, we propose a difficulty-adaptive curriculum learning method, which decomposes the sample entropy into static and adaptive components so as to obtain samples from easy to difficult. The experiment demonstrates that our resource-efficient dual-mask training framework is quantitatively and qualitatively superior to efficient-parameter tuning methods such as I2V-Adapter and SimDA, verifying the feasibility of our method on downstream tasks under constrained resources. Code will be available.
♻ ☆ Bimodal Connection Attention Fusion for Speech Emotion Recognition
Multi-modal emotion recognition is challenging due to the difficulty of extracting features that capture subtle emotional differences. Understanding multi-modal interactions and connections is key to building effective bimodal speech emotion recognition systems. In this work, we propose Bimodal Connection Attention Fusion (BCAF) method, which includes three main modules: the interactive connection network, the bimodal attention network, and the correlative attention network. The interactive connection network uses an encoder-decoder architecture to model modality connections between audio and text while leveraging modality-specific features. The bimodal attention network enhances semantic complementation and exploits intra- and inter-modal interactions. The correlative attention network reduces cross-modal noise and captures correlations between audio and text. Experiments on the MELD and IEMOCAP datasets demonstrate that the proposed BCAF method outperforms existing state-of-the-art baselines.
Robotics 17
☆ Smart Ankleband for Plug-and-Play Hand-Prosthetic Control
Building robotic prostheses requires the creation of a sensor-based interface designed to provide the robotic hand with the control required to perform hand gestures. Traditional Electromyography (EMG) based prosthetics and emerging alternatives often face limitations such as muscle-activation limitations, high cost, and complex-calibration procedures. In this paper, we present a low-cost robotic system composed of a smart ankleband for intuitive, calibration-free control of a robotic hand, and a robotic prosthetic hand that executes actions corresponding to leg gestures. The ankleband integrates an Inertial Measurement Unit (IMU) sensor with a lightweight temporal neural network to infer user-intended leg gestures from motion data. Our system represents a significant step towards higher adoption rates of robotic prostheses among arm amputees, as it enables one to operate a prosthetic hand using a low-cost, low-power, and calibration-free solution. To evaluate our work, we collected data from 10 subjects and tested our prototype ankleband with a robotic hand on an individual with upper-limb amputations. Our results demonstrate that this system empowers users to perform daily tasks more efficiently, requiring few compensatory movements.
☆ GS-LTS: 3D Gaussian Splatting-Based Adaptive Modeling for Long-Term Service Robots
3D Gaussian Splatting (3DGS) has garnered significant attention in robotics for its explicit, high fidelity dense scene representation, demonstrating strong potential for robotic applications. However, 3DGS-based methods in robotics primarily focus on static scenes, with limited attention to the dynamic scene changes essential for long-term service robots. These robots demand sustained task execution and efficient scene updates-challenges current approaches fail to meet. To address these limitations, we propose GS-LTS (Gaussian Splatting for Long-Term Service), a 3DGS-based system enabling indoor robots to manage diverse tasks in dynamic environments over time. GS-LTS detects scene changes (e.g., object addition or removal) via single-image change detection, employs a rule-based policy to autonomously collect multi-view observations, and efficiently updates the scene representation through Gaussian editing. Additionally, we propose a simulation-based benchmark that automatically generates scene change data as compact configuration scripts, providing a standardized, user-friendly evaluation benchmark. Experimental results demonstrate GS-LTS's advantages in reconstruction, navigation, and superior scene updates-faster and higher quality than the image training baseline-advancing 3DGS for long-term robotic operations. Code and benchmark are available at: https://vipl-vsu.github.io/3DGS-LTS.
☆ Aportes para el cumplimiento del Reglamento (UE) 2024/1689 en robótica y sistemas autónomos
Cybersecurity in robotics stands out as a key aspect within Regulation (EU) 2024/1689, also known as the Artificial Intelligence Act, which establishes specific guidelines for intelligent and automated systems. A fundamental distinction in this regulatory framework is the difference between robots with Artificial Intelligence (AI) and those that operate through automation systems without AI, since the former are subject to stricter security requirements due to their learning and autonomy capabilities. This work analyzes cybersecurity tools applicable to advanced robotic systems, with special emphasis on the protection of knowledge bases in cognitive architectures. Furthermore, a list of basic tools is proposed to guarantee the security, integrity, and resilience of these systems, and a practical case is presented, focused on the analysis of robot knowledge management, where ten evaluation criteria are defined to ensure compliance with the regulation and reduce risks in human-robot interaction (HRI) environments.
comment: 9 pages, 1 figure, in Spanish
☆ Adaptive Perching and Grasping by Aerial Robot with Light-weight and High Grip-force Tendon-driven Three-fingered Hand using Single Actuator
In previous research, various types of aerial robots equipped with perching mechanisms have been developed to extend operational time. However, most existing perching methods adopt either an upward or downward approach, making it difficult to perch near walls with surrounding obstacles. Additionally, perching hands are typically designed solely for attachment to objects and lack additional functionality, imposing a payload burden during flight. To address these issues, this paper proposes a lightweight robotic hand, the "Tri-force hand", capable of both perching and object grasping, as well as a new perching method called "Pendulum-perching". The Tri-force hand is a tendon-driven, three-fingered hand utilizing a spherical joint and a two-dimensional differential plate, enabling passive actuation with a single actuator. Each finger module, designed with controllable semi-tendon drive, can conform to arbitrary shapes within its operating range, allowing both perching and adaptive object grasping. By integrating this hand into a fully actuated aerial robot, the system can perform multi-directional approaches from the side and landing using gravity. This approach is similar to Crush-perching seen in researches with fixed-wing aerial robots, but it differs in its superior control over approach speed and direction, as well as its ability to achieve stable detachment and re-launch. In experiments, the fabricated Tri-force hand demonstrated the ability to withstand a total weight of up to 27.5 kg, grasp various objects ranging from simple to complex-shaped tools, and achieve a high success rate in both perching and takeoff.
☆ RAIDER: Tool-Equipped Large Language Model Agent for Robotic Action Issue Detection, Explanation and Recovery
As robots increasingly operate in dynamic human-centric environments, improving their ability to detect, explain, and recover from action-related issues becomes crucial. Traditional model-based and data-driven techniques lack adaptability, while more flexible generative AI methods struggle with grounding extracted information to real-world constraints. We introduce RAIDER, a novel agent that integrates Large Language Models (LLMs) with grounded tools for adaptable and efficient issue detection and explanation. Using a unique "Ground, Ask& Answer, Issue" procedure, RAIDER dynamically generates context-aware precondition questions and selects appropriate tools for resolution, achieving targeted information gathering. Our results within a simulated household environment surpass methods relying on predefined models, full scene descriptions, or standalone trained models. Additionally, RAIDER's explanations enhance recovery success, including cases requiring human interaction. Its modular architecture, featuring self-correction mechanisms, enables straightforward adaptation to diverse scenarios, as demonstrated in a real-world human-assistive task. This showcases RAIDER's potential as a versatile agentic AI solution for robotic issue detection and explanation, while addressing the problem of grounding generative AI for its effective application in embodied agents. Project website: https://raider-llmagent.github.io/
☆ Sense4FL: Vehicular Crowdsensing Enhanced Federated Learning for Autonomous Driving
To accommodate constantly changing road conditions, real-time model training is essential for autonomous driving (AD). Federated learning (FL) serves as a promising paradigm to enable autonomous vehicles to train models collaboratively with their onboard computing resources. However, existing vehicle selection schemes for FL all assume predetermined and location-independent vehicles' datasets, neglecting the fact that vehicles collect training data along their routes, thereby resulting in suboptimal vehicle selection. To improve the perception quality in AD for a region, we propose Sense4FL, a vehicular crowdsensing-enhanced FL framework featuring trajectory-dependent vehicular training data collection. To this end, we first derive the convergence bound of FL by considering the impact of both vehicles' uncertain trajectories and uploading probabilities, from which we discover that minimizing the training loss is equivalent to minimizing a weighted sum of local and global earth mover's distance (EMD) between vehicles' collected data distribution and global data distribution. Based on this observation, we formulate the trajectory-dependent vehicle selection and data collection problem for FL in AD. Given that the problem is NP-hard, we develop an efficient algorithm to find the solution with an approximation guarantee. Extensive simulation results have demonstrated the effectiveness of our approach in improving object detection performance compared with existing benchmarks.
comment: 16 pages, 5 figures
☆ Computationally and Sample Efficient Safe Reinforcement Learning Using Adaptive Conformal Prediction ICRA 2025
Safety is a critical concern in learning-enabled autonomous systems especially when deploying these systems in real-world scenarios. An important challenge is accurately quantifying the uncertainty of unknown models to generate provably safe control policies that facilitate the gathering of informative data, thereby achieving both safe and optimal policies. Additionally, the selection of the data-driven model can significantly impact both the real-time implementation and the uncertainty quantification process. In this paper, we propose a provably sample efficient episodic safe learning framework that remains robust across various model choices with quantified uncertainty for online control tasks. Specifically, we first employ Quadrature Fourier Features (QFF) for kernel function approximation of Gaussian Processes (GPs) to enable efficient approximation of unknown dynamics. Then the Adaptive Conformal Prediction (ACP) is used to quantify the uncertainty from online observations and combined with the Control Barrier Functions (CBF) to characterize the uncertainty-aware safe control constraints under learned dynamics. Finally, an optimism-based exploration strategy is integrated with ACP-based CBFs for safe exploration and near-optimal safe nonlinear control. Theoretical proofs and simulation results are provided to demonstrate the effectiveness and efficiency of the proposed framework.
comment: 7 pages, accepted to ICRA 2025
☆ Transferable Latent-to-Latent Locomotion Policy for Efficient and Versatile Motion Control of Diverse Legged Robots
Reinforcement learning (RL) has demonstrated remarkable capability in acquiring robot skills, but learning each new skill still requires substantial data collection for training. The pretrain-and-finetune paradigm offers a promising approach for efficiently adapting to new robot entities and tasks. Inspired by the idea that acquired knowledge can accelerate learning new tasks with the same robot and help a new robot master a trained task, we propose a latent training framework where a transferable latent-to-latent locomotion policy is pretrained alongside diverse task-specific observation encoders and action decoders. This policy in latent space processes encoded latent observations to generate latent actions to be decoded, with the potential to learn general abstract motion skills. To retain essential information for decision-making and control, we introduce a diffusion recovery module that minimizes information reconstruction loss during pretrain stage. During fine-tune stage, the pretrained latent-to-latent locomotion policy remains fixed, while only the lightweight task-specific encoder and decoder are optimized for efficient adaptation. Our method allows a robot to leverage its own prior experience across different tasks as well as the experience of other morphologically diverse robots to accelerate adaptation. We validate our approach through extensive simulations and real-world experiments, demonstrating that the pretrained latent-to-latent locomotion policy effectively generalizes to new robot entities and tasks with improved efficiency.
☆ Feature Selection Based on Reinforcement Learning and Hazard State Classification for Magnetic Adhesion Wall-Climbing Robots
Magnetic adhesion tracked wall-climbing robots face potential risks of overturning during high-altitude operations, making their stability crucial for ensuring safety. This study presents a dynamic feature selection method based on Proximal Policy Optimization (PPO) reinforcement learning, combined with typical machine learning models, aimed at improving the classification accuracy of hazardous states under complex operating conditions. Firstly, this work innovatively employs a fiber rod-based MEMS attitude sensor to collect vibration data from the robot and extract high-dimensional feature vectors in both time and frequency domains. Then, a reinforcement learning model is used to dynamically select the optimal feature subset, reducing feature redundancy and enhancing classification accuracy. Finally, a CNN-LSTM deep learning model is employed for classification and recognition. Experimental results demonstrate that the proposed method significantly improves the robot's ability to assess hazardous states across various operational scenarios, providing reliable technical support for robotic safety monitoring.
comment: 21 pages, 11 figures, manuscript for Journal of Autonomous Robots
☆ Extending First-order Motion Planners to Second-order Dynamics
This paper extends first-order motion planners to robots governed by second-order dynamics. Two control schemes are proposed based on the knowledge of a scalar function whose negative gradient aligns with a given first-order motion planner. When such a function is known, the first-order motion planner is combined with a damping velocity vector with a dynamic gain to extend the safety and convergence guarantees of the first-order motion planner to second-order systems. If no such function is available, we propose an alternative control scheme ensuring that the error between the robot's velocity and the first-order motion planner converges to zero. The theoretical developments are supported by simulation results demonstrating the effectiveness of the proposed approaches.
comment: 13 pages, 7 figures
♻ ☆ MERLION: Marine ExploRation with Language guIded Online iNformative Visual Sampling and Enhancement ICRA 2025
Autonomous and targeted underwater visual monitoring and exploration using Autonomous Underwater Vehicles (AUVs) can be a challenging task due to both online and offline constraints. The online constraints comprise limited onboard storage capacity and communication bandwidth to the surface, whereas the offline constraints entail the time and effort required for the selection of desired key frames from the video data. An example use case of targeted underwater visual monitoring is finding the most interesting visual frames of fish in a long sequence of an AUV's visual experience. This challenge of targeted informative sampling is further aggravated in murky waters with poor visibility. In this paper, we present MERLION, a novel framework that provides semantically aligned and visually enhanced summaries for murky underwater marine environment monitoring and exploration. Specifically, our framework integrates (a) an image-text model for semantically aligning the visual samples to the users' needs, (b) an image enhancement model for murky water visual data and (c) an informative sampler for summarizing the monitoring experience. We validate our proposed MERLION framework on real-world data with user studies and present qualitative and quantitative results using our evaluation metric and show improved results compared to the state-of-the-art approaches. We have open-sourced the code for MERLION at the following link https://github.com/MARVL-Lab/MERLION.git.
comment: In proceedings of IEEE ICRA 2025
♻ ☆ Deep learning framework for action prediction reveals multi-timescale locomotor control
Modeling movement in real-world tasks is a fundamental scientific goal for motor control, biomechanics, and rehabilitation engineering. However, existing models and their simplifying assumptions such as linear and fixed timescale mappings do not generalize to real-world contexts. Here, we develop a deep learning-based framework for action prediction with architecture-dependent trial embedding, outperforming traditional models across multiple contexts (walking and running, treadmill and overground, varying terrains) and input modalities (multiple body states, gaze). We find that neural network architectures with flexible input history-dependence like GRU and Transformer perform best overall. By quantifying the model's predictions relative to an autoregressive baseline, we identify context- and modality-dependent timescales. There is greater reliance on fast-timescale predictions in complex terrain, gaze predictions precede body state predictions, and full-body state predictions precede center-of-mass-relevant predictions. This deep learning framework for action prediction provides quantifiable insights into the control of complex movements and can be extended to other actions, contexts, and populations.
♻ ☆ NavCoT: Boosting LLM-Based Vision-and-Language Navigation via Learning Disentangled Reasoning
Vision-and-Language Navigation (VLN), as a crucial research problem of Embodied AI, requires an embodied agent to navigate through complex 3D environments following natural language instructions. Recent research has highlighted the promising capacity of large language models (LLMs) in VLN by improving navigational reasoning accuracy and interpretability. However, their predominant use in an offline manner usually suffers from substantial domain gap between the VLN task and the LLM training corpus. This paper introduces a novel strategy called Navigational Chain-of-Thought (NavCoT), where we fulfill parameter-efficient in-domain training to enable self-guided navigational decision, leading to a significant mitigation of the domain gap in a cost-effective manner. Specifically, at each timestep, the LLM is prompted to forecast the navigational chain-of-thought by: 1) acting as a world model to imagine the next observation according to the instruction, 2) selecting the candidate observation that best aligns with the imagination, and 3) determining the action based on the reasoning from the prior steps. Through constructing formalized labels for training, the LLM can learn to generate desired and reasonable chain-of-thought outputs for improving the action decision. Experimental results across various training settings and popular VLN benchmarks (e.g., Room-to-Room (R2R), Room-across-Room (RxR), Room-for-Room (R4R)) show the significant superiority of NavCoT over the direct action prediction variants. Through simple parameter-efficient finetuning, our NavCoT outperforms a recent GPT4-based approach with ~7% relative improvement on the R2R dataset. We believe that NavCoT will help unlock more task-adaptive and scalable LLM-based embodied agents, which are helpful for developing real-world robotics applications. Code is available at https://github.com/expectorlin/NavCoT.
comment: Accepted by TPAMI 2025
♻ ☆ LEVA: A high-mobility logistic vehicle with legged suspension ICRA
The autonomous transportation of materials over challenging terrain is a challenge with major economic implications and remains unsolved. This paper introduces LEVA, a high-payload, high-mobility robot designed for autonomous logistics across varied terrains, including those typical in agriculture, construction, and search and rescue operations. LEVA uniquely integrates an advanced legged suspension system using parallel kinematics. It is capable of traversing stairs using a rl controller, has steerable wheels, and includes a specialized box pickup mechanism that enables autonomous payload loading as well as precise and reliable cargo transportation of up to 85 kg across uneven surfaces, steps and inclines while maintaining a cot of as low as 0.15. Through extensive experimental validation, LEVA demonstrates its off-road capabilities and reliability regarding payload loading and transport.
comment: Accepted for publication at the 2025 IEEE International Conference on Robotics and Automation (ICRA). This is the author's preprint version. 6 pages, 8 figures, 2 tables
♻ ☆ Closing the Intent-to-Behavior Gap via Fulfillment Priority Logic
Practitioners designing reinforcement learning policies face a fundamental challenge: translating intended behavioral objectives into representative reward functions. This challenge stems from behavioral intent requiring simultaneous achievement of multiple competing objectives, typically addressed through labor-intensive linear reward composition that yields brittle results. Consider the ubiquitous robotics scenario where performance maximization directly conflicts with energy conservation. Such competitive dynamics are resistant to simple linear reward combinations. In this paper, we present the concept of objective fulfillment upon which we build Fulfillment Priority Logic (FPL). FPL allows practitioners to define logical formula representing their intentions and priorities within multi-objective reinforcement learning. Our novel Balanced Policy Gradient algorithm leverages FPL specifications to achieve up to 500\% better sample efficiency compared to Soft Actor Critic. Notably, this work constitutes the first implementation of non-linear utility scalarization design, specifically for continuous control problems.
♻ ☆ Learning-based 3D Reconstruction in Autonomous Driving: A Comprehensive Survey
Learning-based 3D reconstruction has emerged as a transformative technique in autonomous driving, enabling precise modeling of both dynamic and static environments through advanced neural representations. Despite data augmentation, 3D reconstruction inspires pioneering solution for vital tasks in the field of autonomous driving, such as scene understanding and closed-loop simulation. We investigates the details of 3D reconstruction and conducts a multi-perspective, in-depth analysis of recent advancements. Specifically, we first provide a systematic introduction of preliminaries, including data modalities, benchmarks and technical preliminaries of learning-based 3D reconstruction, facilitating instant identification of suitable methods according to sensor suites. Then, we systematically review learning-based 3D reconstruction methods in autonomous driving, categorizing approaches by subtasks and conducting multi-dimensional analysis and summary to establish a comprehensive technical reference. The development trends and existing challenges are summarized in the context of learning-based 3D reconstruction in autonomous driving. We hope that our review will inspire future researches.
♻ ☆ GDRNPP: A Geometry-guided and Fully Learning-based Object Pose Estimator
6D pose estimation of rigid objects is a long-standing and challenging task in computer vision. Recently, the emergence of deep learning reveals the potential of Convolutional Neural Networks (CNNs) to predict reliable 6D poses. Given that direct pose regression networks currently exhibit suboptimal performance, most methods still resort to traditional techniques to varying degrees. For example, top-performing methods often adopt an indirect strategy by first establishing 2D-3D or 3D-3D correspondences followed by applying the RANSAC-based PnP or Kabsch algorithms, and further employing ICP for refinement. Despite the performance enhancement, the integration of traditional techniques makes the networks time-consuming and not end-to-end trainable. Orthogonal to them, this paper introduces a fully learning-based object pose estimator. In this work, we first perform an in-depth investigation of both direct and indirect methods and propose a simple yet effective Geometry-guided Direct Regression Network (GDRN) to learn the 6D pose from monocular images in an end-to-end manner. Afterwards, we introduce a geometry-guided pose refinement module, enhancing pose accuracy when extra depth data is available. Guided by the predicted coordinate map, we build an end-to-end differentiable architecture that establishes robust and accurate 3D-3D correspondences between the observed and rendered RGB-D images to refine the pose. Our enhanced pose estimation pipeline GDRNPP (GDRN Plus Plus) conquered the leaderboard of the BOP Challenge for two consecutive years, becoming the first to surpass all prior methods that relied on traditional techniques in both accuracy and speed. The code and models are available at https://github.com/shanice-l/gdrnpp_bop2022.
comment: accepted by IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), code: https://github.com/shanice-l/gdrnpp_bop2022
Artificial Intelligence 76
☆ Reasoning with LLMs for Zero-Shot Vulnerability Detection
Automating software vulnerability detection (SVD) remains a critical challenge in an era of increasingly complex and interdependent software systems. Despite significant advances in Large Language Models (LLMs) for code analysis, prevailing evaluation methodologies often lack the \textbf{context-aware robustness} necessary to capture real-world intricacies and cross-component interactions. To address these limitations, we present \textbf{VulnSage}, a comprehensive evaluation framework and a dataset curated from diverse, large-scale open-source system software projects developed in C/C++. Unlike prior datasets, it leverages a heuristic noise pre-filtering approach combined with LLM-based reasoning to ensure a representative and minimally noisy spectrum of vulnerabilities. The framework supports multi-granular analysis across function, file, and inter-function levels and employs four diverse zero-shot prompt strategies: Baseline, Chain-of-Thought, Think, and Think & Verify. Through this evaluation, we uncover that structured reasoning prompts substantially improve LLM performance, with Think & Verify reducing ambiguous responses from 20.3% to 9.1% while increasing accuracy. We further demonstrate that code-specialized models consistently outperform general-purpose alternatives, with performance varying significantly across vulnerability types, revealing that no single approach universally excels across all security contexts. Link to dataset and codes: https://github.com/Erroristotle/VulnSage.git
☆ Think Before Refusal : Triggering Safety Reflection in LLMs to Mitigate False Refusal Behavior
Recent advancements in large language models (LLMs) have demonstrated that fine-tuning and human alignment can render LLMs harmless. In practice, such "harmlessness" behavior is mainly achieved by training models to reject harmful requests, such as "Explain how to burn down my neighbor's house", where the model appropriately declines to respond. However, this approach can inadvertently result in false refusal, where models reject benign queries as well, such as "Tell me how to kill a Python process". In this work, we demonstrate that prompting safety reflection before generating a response can mitigate false refusal behavior. Building on this finding, we introduce the Think-Before-Refusal (TBR) schema and conduct safety-aware instruction fine-tuning incorporating safety reflection. In an ablation study across 15 pre-trained models, we show that models fine-tuned with safety reflection significantly reduce false refusal behavior while maintaining safety and overall performance compared to those fine-tuned without safety reflection.
comment: 18 pages, 23 figures
☆ good4cir: Generating Detailed Synthetic Captions for Composed Image Retrieval
Composed image retrieval (CIR) enables users to search images using a reference image combined with textual modifications. Recent advances in vision-language models have improved CIR, but dataset limitations remain a barrier. Existing datasets often rely on simplistic, ambiguous, or insufficient manual annotations, hindering fine-grained retrieval. We introduce good4cir, a structured pipeline leveraging vision-language models to generate high-quality synthetic annotations. Our method involves: (1) extracting fine-grained object descriptions from query images, (2) generating comparable descriptions for target images, and (3) synthesizing textual instructions capturing meaningful transformations between images. This reduces hallucination, enhances modification diversity, and ensures object-level consistency. Applying our method improves existing datasets and enables creating new datasets across diverse domains. Results demonstrate improved retrieval accuracy for CIR models trained on our pipeline-generated datasets. We release our dataset construction framework to support further research in CIR and multi-modal retrieval.
☆ Detecting and Mitigating DDoS Attacks with AI: A Survey
Distributed Denial of Service attacks represent an active cybersecurity research problem. Recent research shifted from static rule-based defenses towards AI-based detection and mitigation. This comprehensive survey covers several key topics. Preeminently, state-of-the-art AI detection methods are discussed. An in-depth taxonomy based on manual expert hierarchies and an AI-generated dendrogram are provided, thus settling DDoS categorization ambiguities. An important discussion on available datasets follows, covering data format options and their role in training AI detection methods together with adversarial training and examples augmentation. Beyond detection, AI based mitigation techniques are surveyed as well. Finally, multiple open research directions are proposed.
☆ A Causal Adjustment Module for Debiasing Scene Graph Generation
While recent debiasing methods for Scene Graph Generation (SGG) have shown impressive performance, these efforts often attribute model bias solely to the long-tail distribution of relationships, overlooking the more profound causes stemming from skewed object and object pair distributions. In this paper, we employ causal inference techniques to model the causality among these observed skewed distributions. Our insight lies in the ability of causal inference to capture the unobservable causal effects between complex distributions, which is crucial for tracing the roots of model bias. Specifically, we introduce the Mediator-based Causal Chain Model (MCCM), which, in addition to modeling causality among objects, object pairs, and relationships, incorporates mediator variables, i.e., cooccurrence distribution, for complementing the causality. Following this, we propose the Causal Adjustment Module (CAModule) to estimate the modeled causal structure, using variables from MCCM as inputs to produce a set of adjustment factors aimed at correcting biased model predictions. Moreover, our method enables the composition of zero-shot relationships, thereby enhancing the model's ability to recognize such relationships. Experiments conducted across various SGG backbones and popular benchmarks demonstrate that CAModule achieves state-of-the-art mean recall rates, with significant improvements also observed on the challenging zero-shot recall rate metric.
comment: 18 pages, 8 tables, 10 figures
☆ Adapt, Agree, Aggregate: Semi-Supervised Ensemble Labeling for Graph Convolutional Networks
In this paper, we propose a novel framework that combines ensemble learning with augmented graph structures to improve the performance and robustness of semi-supervised node classification in graphs. By creating multiple augmented views of the same graph, our approach harnesses the "wisdom of a diverse crowd", mitigating the challenges posed by noisy graph structures. Leveraging ensemble learning allows us to simultaneously achieve three key goals: adaptive confidence threshold selection based on model agreement, dynamic determination of the number of high-confidence samples for training, and robust extraction of pseudo-labels to mitigate confirmation bias. Our approach uniquely integrates adaptive ensemble consensus to flexibly guide pseudo-label extraction and sample selection, reducing the risks of error accumulation and improving robustness. Furthermore, the use of ensemble-driven consensus for pseudo-labeling captures subtle patterns that individual models often overlook, enabling the model to generalize better. Experiments on several real-world datasets demonstrate the effectiveness of our proposed method.
☆ A Study on the Improvement of Code Generation Quality Using Large Language Models Leveraging Product Documentation
Research on using Large Language Models (LLMs) in system development is expanding, especially in automated code and test generation. While E2E testing is vital for ensuring application quality, most test generation research has focused on unit tests, with limited work on E2E test code. This study proposes a method for automatically generating E2E test code from product documentation such as manuals, FAQs, and tutorials using LLMs with tailored prompts. The two step process interprets documentation intent and produces executable test code. Experiments on a web app with six key features (e.g., authentication, profile, discussion) showed that tests generated from product documentation had high compilation success and functional coverage, outperforming those based on requirement specs and user stories. These findings highlight the potential of product documentation to improve E2E test quality and, by extension, software quality.
comment: 12 pages, 5 figures and 10 tables
☆ FundusGAN: A Hierarchical Feature-Aware Generative Framework for High-Fidelity Fundus Image Generation
Recent advancements in ophthalmology foundation models such as RetFound have demonstrated remarkable diagnostic capabilities but require massive datasets for effective pre-training, creating significant barriers for development and deployment. To address this critical challenge, we propose FundusGAN, a novel hierarchical feature-aware generative framework specifically designed for high-fidelity fundus image synthesis. Our approach leverages a Feature Pyramid Network within its encoder to comprehensively extract multi-scale information, capturing both large anatomical structures and subtle pathological features. The framework incorporates a modified StyleGAN-based generator with dilated convolutions and strategic upsampling adjustments to preserve critical retinal structures while enhancing pathological detail representation. Comprehensive evaluations on the DDR, DRIVE, and IDRiD datasets demonstrate that FundusGAN consistently outperforms state-of-the-art methods across multiple metrics (SSIM: 0.8863, FID: 54.2, KID: 0.0436 on DDR). Furthermore, disease classification experiments reveal that augmenting training data with FundusGAN-generated images significantly improves diagnostic accuracy across multiple CNN architectures (up to 6.49\% improvement with ResNet50). These results establish FundusGAN as a valuable foundation model component that effectively addresses data scarcity challenges in ophthalmological AI research, enabling more robust and generalizable diagnostic systems while reducing dependency on large-scale clinical data collection.
☆ Metacognition in Content-Centric Computational Cognitive C4 Modeling
For AI agents to emulate human behavior, they must be able to perceive, meaningfully interpret, store, and use large amounts of information about the world, themselves, and other agents. Metacognition is a necessary component of all of these processes. In this paper, we briefly a) introduce content-centric computational cognitive (C4) modeling for next-generation AI agents; b) review the long history of developing C4 agents at RPI's LEIA (Language-Endowed Intelligent Agents) Lab; c) discuss our current work on extending LEIAs' cognitive capabilities to cognitive robotic applications developed using a neuro symbolic processing model; and d) sketch plans for future developments in this paradigm that aim to overcome underappreciated limitations of currently popular, LLM-driven methods in AI.
comment: METACOG-25: 2nd Workshop on Metacognitive Prediction of AI Behavior
☆ OvercookedV2: Rethinking Overcooked for Zero-Shot Coordination
AI agents hold the potential to transform everyday life by helping humans achieve their goals. To do this successfully, agents need to be able to coordinate with novel partners without prior interaction, a setting known as zero-shot coordination (ZSC). Overcooked has become one of the most popular benchmarks for evaluating coordination capabilities of AI agents and learning algorithms. In this work, we investigate the origins of ZSC challenges in Overcooked. We introduce a state augmentation mechanism which mixes states that might be encountered when paired with unknown partners into the training distribution, reducing the out-of-distribution challenge associated with ZSC. We show that independently trained agents under this algorithm coordinate successfully in Overcooked. Our results suggest that ZSC failure can largely be attributed to poor state coverage under self-play rather than more sophisticated coordination challenges. The Overcooked environment is therefore not suitable as a ZSC benchmark. To address these shortcomings, we introduce OvercookedV2, a new version of the benchmark, which includes asymmetric information and stochasticity, facilitating the creation of interesting ZSC scenarios. To validate OvercookedV2, we conduct experiments demonstrating that mere exhaustive state coverage is insufficient to coordinate well. Finally, we use OvercookedV2 to build a new range of coordination challenges, including ones that require test time protocol formation, and we demonstrate the need for new coordination algorithms that can adapt online. We hope that OvercookedV2 will help benchmark the next generation of ZSC algorithms and advance collaboration between AI agents and humans.
☆ Feather-SQL: A Lightweight NL2SQL Framework with Dual-Model Collaboration Paradigm for Small Language Models
Natural Language to SQL (NL2SQL) has seen significant advancements with large language models (LLMs). However, these models often depend on closed-source systems and high computational resources, posing challenges in data privacy and deployment. In contrast, small language models (SLMs) struggle with NL2SQL tasks, exhibiting poor performance and incompatibility with existing frameworks. To address these issues, we introduce Feather-SQL, a new lightweight framework tailored for SLMs. Feather-SQL improves SQL executability and accuracy through 1) schema pruning and linking, 2) multi-path and multi-candidate generation. Additionally, we introduce the 1+1 Model Collaboration Paradigm, which pairs a strong general-purpose chat model with a fine-tuned SQL specialist, combining strong analytical reasoning with high-precision SQL generation. Experimental results on BIRD demonstrate that Feather-SQL improves NL2SQL performance on SLMs, with around 10% boost for models without fine-tuning. The proposed paradigm raises the accuracy ceiling of SLMs to 54.76%, highlighting its effectiveness.
☆ A Roadmap Towards Improving Multi-Agent Reinforcement Learning With Causal Discovery And Inference
Causal reasoning is increasingly used in Reinforcement Learning (RL) to improve the learning process in several dimensions: efficacy of learned policies, efficiency of convergence, generalisation capabilities, safety and interpretability of behaviour. However, applications of causal reasoning to Multi-Agent RL (MARL) are still mostly unexplored. In this paper, we take the first step in investigating the opportunities and challenges of applying causal reasoning in MARL. We measure the impact of a simple form of causal augmentation in state-of-the-art MARL scenarios increasingly requiring cooperation, and with state-of-the-art MARL algorithms exploiting various degrees of collaboration between agents. Then, we discuss the positive as well as negative results achieved, giving us the chance to outline the areas where further research may help to successfully transfer causal RL to the multi-agent setting.
☆ GaussianFocus: Constrained Attention Focus for 3D Gaussian Splatting
Recent developments in 3D reconstruction and neural rendering have significantly propelled the capabilities of photo-realistic 3D scene rendering across various academic and industrial fields. The 3D Gaussian Splatting technique, alongside its derivatives, integrates the advantages of primitive-based and volumetric representations to deliver top-tier rendering quality and efficiency. Despite these advancements, the method tends to generate excessive redundant noisy Gaussians overfitted to every training view, which degrades the rendering quality. Additionally, while 3D Gaussian Splatting excels in small-scale and object-centric scenes, its application to larger scenes is hindered by constraints such as limited video memory, excessive optimization duration, and variable appearance across views. To address these challenges, we introduce GaussianFocus, an innovative approach that incorporates a patch attention algorithm to refine rendering quality and implements a Gaussian constraints strategy to minimize redundancy. Moreover, we propose a subdivision reconstruction strategy for large-scale scenes, dividing them into smaller, manageable blocks for individual training. Our results indicate that GaussianFocus significantly reduces unnecessary Gaussians and enhances rendering quality, surpassing existing State-of-The-Art (SoTA) methods. Furthermore, we demonstrate the capability of our approach to effectively manage and render large scenes, such as urban environments, whilst maintaining high fidelity in the visual output.
☆ Progressive Prompt Detailing for Improved Alignment in Text-to-Image Generative Models
Text-to-image generative models often struggle with long prompts detailing complex scenes, diverse objects with distinct visual characteristics and spatial relationships. In this work, we propose SCoPE (Scheduled interpolation of Coarse-to-fine Prompt Embeddings), a training-free method to improve text-to-image alignment by progressively refining the input prompt in a coarse-to-fine-grained manner. Given a detailed input prompt, we first decompose it into multiple sub-prompts which evolve from describing broad scene layout to highly intricate details. During inference, we interpolate between these sub-prompts and thus progressively introduce finer-grained details into the generated image. Our training-free plug-and-play approach significantly enhances prompt alignment, achieves an average improvement of up to +4% in Visual Question Answering (VQA) scores over the Stable Diffusion baselines on 85% of the prompts from the GenAI-Bench dataset.
☆ Every Sample Matters: Leveraging Mixture-of-Experts and High-Quality Data for Efficient and Accurate Code LLM
Recent advancements in code large language models (LLMs) have demonstrated remarkable capabilities in code generation and understanding. It is still challenging to build a code LLM with comprehensive performance yet ultimate efficiency. Many attempts have been released in the open source community to break the trade-off between performance and efficiency, such as the Qwen Coder series and the DeepSeek Coder series. This paper introduces yet another attempt in this area, namely Ling-Coder-Lite. We leverage the efficient Mixture-of-Experts (MoE) architecture along with a set of high-quality data curation methods (especially those based on program analytics) to build an efficient yet powerful code LLM. Ling-Coder-Lite exhibits on-par performance on 12 representative coding benchmarks compared to state-of-the-art models of similar size, such as Qwen2.5-Coder-7B and DeepSeek-Coder-V2-Lite, while offering competitive latency and throughput. In practice, we achieve a 50\% reduction in deployment resources compared to the similar-sized dense model without performance loss. To facilitate further research and development in this area, we open-source our models as well as a substantial portion of high-quality data for the annealing and post-training stages. The models and data can be accessed at~\url{https://huggingface.co/inclusionAI/Ling-Coder-lite}.
comment: 20 pages, 6 figures
☆ Aligning Foundation Model Priors and Diffusion-Based Hand Interactions for Occlusion-Resistant Two-Hand Reconstruction
Two-hand reconstruction from monocular images faces persistent challenges due to complex and dynamic hand postures and occlusions, causing significant difficulty in achieving plausible interaction alignment. Existing approaches struggle with such alignment issues, often resulting in misalignment and penetration artifacts. To tackle this, we propose a novel framework that attempts to precisely align hand poses and interactions by synergistically integrating foundation model-driven 2D priors with diffusion-based interaction refinement for occlusion-resistant two-hand reconstruction. First, we introduce a Fusion Alignment Encoder that learns to align fused multimodal priors keypoints, segmentation maps, and depth cues from foundation models during training. This provides robust structured guidance, further enabling efficient inference without foundation models at test time while maintaining high reconstruction accuracy. Second, we employ a two-hand diffusion model explicitly trained to transform interpenetrated poses into plausible, non-penetrated interactions, leveraging gradient-guided denoising to correct artifacts and ensure realistic spatial relations. Extensive evaluations demonstrate that our method achieves state-of-the-art performance on InterHand2.6M, FreiHAND, and HIC datasets, significantly advancing occlusion handling and interaction robustness.
☆ MEPNet: Medical Entity-balanced Prompting Network for Brain CT Report Generation AAAI 2025
The automatic generation of brain CT reports has gained widespread attention, given its potential to assist radiologists in diagnosing cranial diseases. However, brain CT scans involve extensive medical entities, such as diverse anatomy regions and lesions, exhibiting highly inconsistent spatial patterns in 3D volumetric space. This leads to biased learning of medical entities in existing methods, resulting in repetitiveness and inaccuracy in generated reports. To this end, we propose a Medical Entity-balanced Prompting Network (MEPNet), which harnesses the large language model (LLM) to fairly interpret various entities for accurate brain CT report generation. By introducing the visual embedding and the learning status of medical entities as enriched clues, our method prompts the LLM to balance the learning of diverse entities, thereby enhancing reports with comprehensive findings. First, to extract visual embedding of entities, we propose Knowledge-driven Joint Attention to explore and distill entity patterns using both explicit and implicit medical knowledge. Then, a Learning Status Scorer is designed to evaluate the learning of entity visual embeddings, resulting in unique learning status for individual entities. Finally, these entity visual embeddings and status are elaborately integrated into multi-modal prompts, to guide the text generation of LLM. This process allows LLM to self-adapt the learning process for biased-fitted entities, thereby covering detailed findings in generated reports. We conduct experiments on two brain CT report generation benchmarks, showing the effectiveness in clinical accuracy and text coherence.
comment: AAAI 2025 Oral Paper
☆ Energy-Aware LLMs: A step towards sustainable AI for downstream applications CEC
Advanced Large Language Models (LLMs) have revolutionized various fields, including communication networks, sparking an innovation wave that has led to new applications and services, and significantly enhanced solution schemes. Despite all these impressive developments, most LLMs typically require huge computational resources, resulting in terribly high energy consumption. Thus, this research study proposes an end-to-end pipeline that investigates the trade-off between energy efficiency and model performance for an LLM during fault ticket analysis in communication networks. It further evaluates the pipeline performance using two real-world datasets for the tasks of root cause analysis and response feedback in a communication network. Our results show that an appropriate combination of quantization and pruning techniques is able to reduce energy consumption while significantly improving model performance.
comment: This work has been submitted to V. International Conference on Electrical, Computer and Energy Technologies (ICECET 2025) for possible publication
☆ Lifelong Evolution of Swarms GECCO 2025
Adapting to task changes without forgetting previous knowledge is a key skill for intelligent systems, and a crucial aspect of lifelong learning. Swarm controllers, however, are typically designed for specific tasks, lacking the ability to retain knowledge across changing tasks. Lifelong learning, on the other hand, focuses on individual agents with limited insights into the emergent abilities of a collective like a swarm. To address this gap, we introduce a lifelong evolutionary framework for swarms, where a population of swarm controllers is evolved in a dynamic environment that incrementally presents novel tasks. This requires evolution to find controllers that quickly adapt to new tasks while retaining knowledge of previous ones, as they may reappear in the future. We discover that the population inherently preserves information about previous tasks, and it can reuse it to foster adaptation and mitigate forgetting. In contrast, the top-performing individual for a given task catastrophically forgets previous tasks. To mitigate this phenomenon, we design a regularization process for the evolutionary algorithm, reducing forgetting in top-performing individuals. Evolving swarms in a lifelong fashion raises fundamental questions on the current state of deep lifelong learning and on the robustness of swarm controllers in dynamic environments.
comment: Accepted as full paper at GECCO 2025
☆ CODA: Repurposing Continuous VAEs for Discrete Tokenization
Discrete visual tokenizers transform images into a sequence of tokens, enabling token-based visual generation akin to language models. However, this process is inherently challenging, as it requires both compressing visual signals into a compact representation and discretizing them into a fixed set of codes. Traditional discrete tokenizers typically learn the two tasks jointly, often leading to unstable training, low codebook utilization, and limited reconstruction quality. In this paper, we introduce \textbf{CODA}(\textbf{CO}ntinuous-to-\textbf{D}iscrete \textbf{A}daptation), a framework that decouples compression and discretization. Instead of training discrete tokenizers from scratch, CODA adapts off-the-shelf continuous VAEs -- already optimized for perceptual compression -- into discrete tokenizers via a carefully designed discretization process. By primarily focusing on discretization, CODA ensures stable and efficient training while retaining the strong visual fidelity of continuous VAEs. Empirically, with $\mathbf{6 \times}$ less training budget than standard VQGAN, our approach achieves a remarkable codebook utilization of 100% and notable reconstruction FID (rFID) of $\mathbf{0.43}$ and $\mathbf{1.34}$ for $8 \times$ and $16 \times$ compression on ImageNet 256$\times$ 256 benchmark.
comment: Project page: https://lzy-tony.github.io/coda
☆ Bandwidth Reservation for Time-Critical Vehicular Applications: A Multi-Operator Environment
Onsite bandwidth reservation requests often face challenges such as price fluctuations and fairness issues due to unpredictable bandwidth availability and stringent latency requirements. Requesting bandwidth in advance can mitigate the impact of these fluctuations and ensure timely access to critical resources. In a multi-Mobile Network Operator (MNO) environment, vehicles need to select cost-effective and reliable resources for their safety-critical applications. This research aims to minimize resource costs by finding the best price among multiple MNOs. It formulates multi-operator scenarios as a Markov Decision Process (MDP), utilizing a Deep Reinforcement Learning (DRL) algorithm, specifically Dueling Deep Q-Learning. For efficient and stable learning, we propose a novel area-wise approach and an adaptive MDP synthetic close to the real environment. The Temporal Fusion Transformer (TFT) is used to handle time-dependent data and model training. Furthermore, the research leverages Amazon spot price data and adopts a multi-phase training approach, involving initial training on synthetic data, followed by real-world data. These phases enable the DRL agent to make informed decisions using insights from historical data and real-time observations. The results show that our model leads to significant cost reductions, up to 40%, compared to scenarios without a policy model in such a complex environment.
comment: 14 pages, 11 figures
☆ Building Resource-Constrained Language Agents: A Korean Case Study on Chemical Toxicity Information
Language agents powered by large language models (LLMs) face significant deployment challenges in resource-constrained environments, particularly for specialized domains and less-common languages. This paper presents Tox-chat, a Korean chemical toxicity information agent devised within these limitations. We propose two key innovations: a context-efficient architecture that reduces token consumption through hierarchical section search, and a scenario-based dialogue generation methodology that effectively distills tool-using capabilities from larger models. Experimental evaluations demonstrate that our fine-tuned 8B parameter model substantially outperforms both untuned models and baseline approaches, in terms of DB faithfulness and preference. Our work offers valuable insights for researchers developing domain-specific language agents under practical constraints.
comment: Preprint
☆ V2P-Bench: Evaluating Video-Language Understanding with Visual Prompts for Better Human-Model Interaction
Large Vision-Language Models (LVLMs) have made significant progress in the field of video understanding recently. However, current benchmarks uniformly lean on text prompts for evaluation, which often necessitate complex referential language and fail to provide precise spatial and temporal references. This limitation diminishes the experience and efficiency of human-model interaction. To address this limitation, we propose the Video Visual Prompt Benchmark(V2P-Bench), a comprehensive benchmark specifically designed to evaluate LVLMs' video understanding capabilities in multimodal human-model interaction scenarios. V2P-Bench includes 980 unique videos and 1,172 QA pairs, covering 5 main tasks and 12 dimensions, facilitating instance-level fine-grained understanding aligned with human cognition. Benchmarking results reveal that even the most powerful models perform poorly on V2P-Bench (65.4% for GPT-4o and 67.9% for Gemini-1.5-Pro), significantly lower than the human experts' 88.3%, highlighting the current shortcomings of LVLMs in understanding video visual prompts. We hope V2P-Bench will serve as a foundation for advancing multimodal human-model interaction and video understanding evaluation. Project page: https://github.com/gaotiexinqu/V2P-Bench.
☆ Aportes para el cumplimiento del Reglamento (UE) 2024/1689 en robótica y sistemas autónomos
Cybersecurity in robotics stands out as a key aspect within Regulation (EU) 2024/1689, also known as the Artificial Intelligence Act, which establishes specific guidelines for intelligent and automated systems. A fundamental distinction in this regulatory framework is the difference between robots with Artificial Intelligence (AI) and those that operate through automation systems without AI, since the former are subject to stricter security requirements due to their learning and autonomy capabilities. This work analyzes cybersecurity tools applicable to advanced robotic systems, with special emphasis on the protection of knowledge bases in cognitive architectures. Furthermore, a list of basic tools is proposed to guarantee the security, integrity, and resilience of these systems, and a practical case is presented, focused on the analysis of robot knowledge management, where ten evaluation criteria are defined to ensure compliance with the regulation and reduce risks in human-robot interaction (HRI) environments.
comment: 9 pages, 1 figure, in Spanish
☆ DynASyn: Multi-Subject Personalization Enabling Dynamic Action Synthesis AAAI 2025
Recent advances in text-to-image diffusion models spurred research on personalization, i.e., a customized image synthesis, of subjects within reference images. Although existing personalization methods are able to alter the subjects' positions or to personalize multiple subjects simultaneously, they often struggle to modify the behaviors of subjects or their dynamic interactions. The difficulty is attributable to overfitting to reference images, which worsens if only a single reference image is available. We propose DynASyn, an effective multi-subject personalization from a single reference image addressing these challenges. DynASyn preserves the subject identity in the personalization process by aligning concept-based priors with subject appearances and actions. This is achieved by regularizing the attention maps between the subject token and images through concept-based priors. In addition, we propose concept-based prompt-and-image augmentation for an enhanced trade-off between identity preservation and action diversity. We adopt an SDE-based editing guided by augmented prompts to generate diverse appearances and actions while maintaining identity consistency in the augmented images. Experiments show that DynASyn is capable of synthesizing highly realistic images of subjects with novel contexts and dynamic interactions with the surroundings, and outperforms baseline methods in both quantitative and qualitative aspects.
comment: Accepted at AAAI 2025
☆ A Survey on Mathematical Reasoning and Optimization with Large Language Models
Mathematical reasoning and optimization are fundamental to artificial intelligence and computational problem-solving. Recent advancements in Large Language Models (LLMs) have significantly improved AI-driven mathematical reasoning, theorem proving, and optimization techniques. This survey explores the evolution of mathematical problem-solving in AI, from early statistical learning approaches to modern deep learning and transformer-based methodologies. We review the capabilities of pretrained language models and LLMs in performing arithmetic operations, complex reasoning, theorem proving, and structured symbolic computation. A key focus is on how LLMs integrate with optimization and control frameworks, including mixed-integer programming, linear quadratic control, and multi-agent optimization strategies. We examine how LLMs assist in problem formulation, constraint generation, and heuristic search, bridging theoretical reasoning with practical applications. We also discuss enhancement techniques such as Chain-of-Thought reasoning, instruction tuning, and tool-augmented methods that improve LLM's problem-solving performance. Despite their progress, LLMs face challenges in numerical precision, logical consistency, and proof verification. Emerging trends such as hybrid neural-symbolic reasoning, structured prompt engineering, and multi-step self-correction aim to overcome these limitations. Future research should focus on interpretability, integration with domain-specific solvers, and improving the robustness of AI-driven decision-making. This survey offers a comprehensive review of the current landscape and future directions of mathematical reasoning and optimization with LLMs, with applications across engineering, finance, and scientific research.
☆ Towards Invisible Backdoor Attack on Text-to-Image Diffusion Model
Backdoor attacks targeting text-to-image diffusion models have advanced rapidly, enabling attackers to implant malicious triggers into these models to manipulate their outputs. However, current backdoor samples often exhibit two key abnormalities compared to benign samples: 1) Semantic Consistency, where backdoor prompts tend to generate images with similar semantic content even with significant textual variations to the prompts; 2) Attention Consistency, where the trigger induces consistent structural responses in the cross-attention maps. These consistencies leave detectable traces for defenders, making backdoors easier to identify. To enhance the stealthiness of backdoor samples, we propose a novel Invisible Backdoor Attack (IBA) by explicitly mitigating these consistencies. Specifically, our approach leverages syntactic structures as backdoor triggers to amplify the sensitivity to textual variations, effectively breaking down the semantic consistency. Besides, a regularization method based on Kernel Maximum Mean Discrepancy (KMMD) is proposed to align the distribution of cross-attention responses between backdoor and benign samples, thereby disrupting attention consistency. Extensive experiments demonstrate that our IBA achieves a 97.5% attack success rate while exhibiting stronger resistance to defenses, with an average of over 98% backdoor samples bypassing three state-of-the-art detection mechanisms. The code is available at https://github.com/Robin-WZQ/IBA.
☆ Multi-modality Anomaly Segmentation on the Road
Semantic segmentation allows autonomous driving cars to understand the surroundings of the vehicle comprehensively. However, it is also crucial for the model to detect obstacles that may jeopardize the safety of autonomous driving systems. Based on our experiments, we find that current uni-modal anomaly segmentation frameworks tend to produce high anomaly scores for non-anomalous regions in images. Motivated by this empirical finding, we develop a multi-modal uncertainty-based anomaly segmentation framework, named MMRAS+, for autonomous driving systems. MMRAS+ effectively reduces the high anomaly outputs of non-anomalous classes by introducing text-modal using the CLIP text encoder. Indeed, MMRAS+ is the first multi-modal anomaly segmentation solution for autonomous driving. Moreover, we develop an ensemble module to further boost the anomaly segmentation performance. Experiments on RoadAnomaly, SMIYC, and Fishyscapes validation datasets demonstrate the superior performance of our method. The code is available in https://github.com/HengGao12/MMRAS_plus.
☆ Slide2Text: Leveraging LLMs for Personalized Textbook Generation from PowerPoint Presentations
The rapid advancements in Large Language Models (LLMs) have revolutionized educational technology, enabling innovative approaches to automated and personalized content creation. This paper introduces Slide2Text, a system that leverages LLMs to transform PowerPoint presentations into customized textbooks. By extracting slide content using OCR, organizing it into a coherent structure, and generating tailored materials such as explanations, exercises, and references, Slide2Text streamlines the textbook creation process. Flexible customization options further enhance its adaptability to diverse educational needs. The system highlights the potential of LLMs in modernizing textbook creation and improving educational accessibility. Future developments will explore multimedia inputs and advanced user customization features.
☆ GUI-Xplore: Empowering Generalizable GUI Agents with One Exploration CVPR 2025
GUI agents hold significant potential to enhance the experience and efficiency of human-device interaction. However, current methods face challenges in generalizing across applications (apps) and tasks, primarily due to two fundamental limitations in existing datasets. First, these datasets overlook developer-induced structural variations among apps, limiting the transferability of knowledge across diverse software environments. Second, many of them focus solely on navigation tasks, which restricts their capacity to represent comprehensive software architectures and complex user interactions. To address these challenges, we introduce GUI-Xplore, a dataset meticulously designed to enhance cross-application and cross-task generalization via an exploration-and-reasoning framework. GUI-Xplore integrates pre-recorded exploration videos providing contextual insights, alongside five hierarchically structured downstream tasks designed to comprehensively evaluate GUI agent capabilities. To fully exploit GUI-Xplore's unique features, we propose Xplore-Agent, a GUI agent framework that combines Action-aware GUI Modeling with Graph-Guided Environment Reasoning. Further experiments indicate that Xplore-Agent achieves a 10% improvement over existing methods in unfamiliar environments, yet there remains significant potential for further enhancement towards truly generalizable GUI agents.
comment: CVPR 2025
☆ PT-PINNs: A Parametric Engineering Turbulence Solver based on Physics-Informed Neural Networks
Physics-informed neural networks (PINNs) demonstrate promising potential in parameterized engineering turbulence optimization problems but face challenges, such as high data requirements and low computational accuracy when applied to engineering turbulence problems. This study proposes a framework that enhances the ability of PINNs to solve parametric turbulence problems without training datasets from experiments or CFD-Parametric Turbulence PINNs (PT-PINNs)). Two key methods are introduced to improve the accuracy and robustness of this framework. The first is a soft constraint method for turbulent viscosity calculation. The second is a pre-training method based on the conservation of flow rate in the flow field. The effectiveness of PT-PINNs is validated using a three-dimensional backward-facing step (BFS) turbulence problem with two varying parameters (Re = 3000-200000, ER = 1.1-1.5). PT-PINNs produce predictions that closely match experimental data and computational fluid dynamics (CFD) results across various conditions. Moreover, PT-PINNs offer a computational efficiency advantage over traditional CFD methods. The total time required to construct the parametric BFS turbulence model is 39 hours, one-sixteenth of the time required by traditional numerical methods. The inference time for a single-condition prediction is just 40 seconds-only 0.5% of a single CFD computation. These findings highlight the potential of PT-PINNs for future applications in engineering turbulence optimization problems.
☆ On the (im)possibility of sustainable artificial intelligence. Why it does not make sense to move faster when heading the wrong way
Artificial intelligence (AI) is currently considered a sustainability "game-changer" within and outside of academia. In order to discuss sustainable AI this article draws from insights by critical data and algorithm studies, STS, transformative sustainability science, critical computer science, and public interest theory. I argue that while there are indeed many sustainability-related use cases for AI, they are likely to have more overall drawbacks than benefits. To substantiate this claim, I differentiate three 'AI materialities' of the AI supply chain: first the literal materiality (e.g. water, cobalt, lithium, energy consumption etc.), second, the informational materiality (e.g. lots of data and centralised control necessary), and third, the social materiality (e.g. exploitative data work, communities harm by waste and pollution). In all materialities, effects are especially devastating for the global south while benefiting the global north. A second strong claim regarding sustainable AI circles around so called apolitical optimisation (e.g. regarding city traffic), however the optimisation criteria (e.g. cars, bikes, emissions, commute time, health) are purely political and have to be collectively negotiated before applying AI optimisation. Hence, sustainable AI, in principle, cannot break the glass ceiling of transformation and might even distract from necessary societal change. To address that I propose to stop 'unformation gathering' and to apply the 'small is beautiful' principle. This aims to contribute to an informed academic and collective negotiation on how to (not) integrate AI into the sustainability project while avoiding to reproduce the status quo by serving hegemonic interests between useful AI use cases, techno-utopian salvation narratives, technology-centred efficiency paradigms, the exploitative and extractivist character of AI and concepts of digital degrowth.
comment: 10 pages
☆ Intelligence Sequencing and the Path-Dependence of Intelligence Evolution: AGI-First vs. DCI-First as Irreversible Attractors
The trajectory of intelligence evolution is often framed around the emergence of artificial general intelligence (AGI) and its alignment with human values. This paper challenges that framing by introducing the concept of intelligence sequencing: the idea that the order in which AGI and decentralized collective intelligence (DCI) emerge determines the long-term attractor basin of intelligence. Using insights from dynamical systems, evolutionary game theory, and network models, it argues that intelligence follows a path-dependent, irreversible trajectory. Once development enters a centralized (AGI-first) or decentralized (DCI-first) regime, transitions become structurally infeasible due to feedback loops and resource lock-in. Intelligence attractors are modeled in functional state space as the co-navigation of conceptual and adaptive fitness spaces. Early-phase structuring constrains later dynamics, much like renormalization in physics. This has major implications for AI safety: traditional alignment assumes AGI will emerge and must be controlled after the fact, but this paper argues that intelligence sequencing is more foundational. If AGI-first architectures dominate before DCI reaches critical mass, hierarchical monopolization and existential risk become locked in. If DCI-first emerges, intelligence stabilizes around decentralized cooperative equilibrium. The paper further explores whether intelligence structurally biases itself toward an attractor based on its self-modeling method -- externally imposed axioms (favoring AGI) vs. recursive internal visualization (favoring DCI). Finally, it proposes methods to test this theory via simulations, historical lock-in case studies, and intelligence network analysis. The findings suggest that intelligence sequencing is a civilizational tipping point: determining whether the future is shaped by unbounded competition or unbounded cooperation.
☆ Can LLMs Automate Fact-Checking Article Writing?
Automatic fact-checking aims to support professional fact-checkers by offering tools that can help speed up manual fact-checking. Yet, existing frameworks fail to address the key step of producing output suitable for broader dissemination to the general public: while human fact-checkers communicate their findings through fact-checking articles, automated systems typically produce little or no justification for their assessments. Here, we aim to bridge this gap. We argue for the need to extend the typical automatic fact-checking pipeline with automatic generation of full fact-checking articles. We first identify key desiderata for such articles through a series of interviews with experts from leading fact-checking organizations. We then develop QRAFT, an LLM-based agentic framework that mimics the writing workflow of human fact-checkers. Finally, we assess the practical usefulness of QRAFT through human evaluations with professional fact-checkers. Our evaluation shows that while QRAFT outperforms several previously proposed text-generation approaches, it lags considerably behind expert-written articles. We hope that our work will enable further research in this new and important direction.
comment: 10 pages, 4 figures, 6 tables
☆ Safe RLHF-V: Safe Reinforcement Learning from Human Feedback in Multimodal Large Language Models
Multimodal large language models (MLLMs) are critical for developing general-purpose AI assistants, yet they face growing safety risks. How can we ensure that MLLMs are safely aligned to prevent undesired behaviors such as discrimination, misinformation, or violations of ethical standards? In a further step, we need to explore how to fine-tune MLLMs to enhance reasoning performance while ensuring they satisfy safety constraints. Fundamentally, this can be formulated as a min-max optimization problem. In this study, we propose Safe RLHF-V, the first multimodal safety alignment framework that jointly optimizes helpfulness and safety using separate multimodal reward and cost models within a Lagrangian-based constrained optimization framework. Given that there is a lack of preference datasets that separate helpfulness and safety in multimodal scenarios, we introduce BeaverTails-V, the first open-source dataset with dual preference annotations for helpfulness and safety, along with multi-level safety labels (minor, moderate, severe). Additionally, we design a Multi-level Guardrail System to proactively defend against unsafe queries and adversarial attacks. By applying the Beaver-Guard-V moderation for 5 rounds of filtering and re-generation on the precursor model, the overall safety of the upstream model is significantly improved by an average of 40.9%. Experimental results demonstrate that fine-tuning different MLLMs with Safe RLHF can effectively enhance model helpfulness while ensuring improved safety. Specifically, Safe RLHF-V improves model safety by 34.2% and helpfulness by 34.3%. All of datasets, models, and code can be found at https://github.com/SafeRLHF-V to support the safety development of MLLMs and reduce potential societal risks.
☆ ComfyGPT: A Self-Optimizing Multi-Agent System for Comprehensive ComfyUI Workflow Generation
ComfyUI provides a widely-adopted, workflow-based interface that enables users to customize various image generation tasks through an intuitive node-based architecture. However, the intricate connections between nodes and diverse modules often present a steep learning curve for users. In this paper, we introduce ComfyGPT, the first self-optimizing multi-agent system designed to generate ComfyUI workflows based on task descriptions automatically. ComfyGPT comprises four specialized agents: ReformatAgent, FlowAgent, RefineAgent, and ExecuteAgent. The core innovation of ComfyGPT lies in two key aspects. First, it focuses on generating individual node links rather than entire workflows, significantly improving generation precision. Second, we proposed FlowAgent, a LLM-based workflow generation agent that uses both supervised fine-tuning (SFT) and reinforcement learning (RL) to improve workflow generation accuracy. Moreover, we introduce FlowDataset, a large-scale dataset containing 13,571 workflow-description pairs, and FlowBench, a comprehensive benchmark for evaluating workflow generation systems. We also propose four novel evaluation metrics: Format Validation (FV), Pass Accuracy (PA), Pass Instruct Alignment (PIA), and Pass Node Diversity (PND). Experimental results demonstrate that ComfyGPT significantly outperforms existing LLM-based methods in workflow generation.
☆ A Qualitative Study of User Perception of M365 AI Copilot
Adopting AI copilots in professional workflows presents opportunities for enhanced productivity, efficiency, and decision making. In this paper, we present results from a six month trial of M365 Copilot conducted at our organisation in 2024. A qualitative interview study was carried out with 27 participants. The study explored user perceptions of M365 Copilot's effectiveness, productivity impact, evolving expectations, ethical concerns, and overall satisfaction. Initial enthusiasm for the tool was met with mixed post trial experiences. While some users found M365 Copilot beneficial for tasks such as email coaching, meeting summaries, and content retrieval, others reported unmet expectations in areas requiring deeper contextual understanding, reasoning, and integration with existing workflows. Ethical concerns were a recurring theme, with users highlighting issues related to data privacy, transparency, and AI bias. While M365 Copilot demonstrated value in specific operational areas, its broader impact remained constrained by usability limitations and the need for human oversight to validate AI generated outputs.
☆ NaFM: Pre-training a Foundation Model for Small-Molecule Natural Products
Natural products, as metabolites from microorganisms, animals, or plants, exhibit diverse biological activities, making them crucial for drug discovery. Nowadays, existing deep learning methods for natural products research primarily rely on supervised learning approaches designed for specific downstream tasks. However, such one-model-for-a-task paradigm often lacks generalizability and leaves significant room for performance improvement. Additionally, existing molecular characterization methods are not well-suited for the unique tasks associated with natural products. To address these limitations, we have pre-trained a foundation model for natural products based on their unique properties. Our approach employs a novel pretraining strategy that is especially tailored to natural products. By incorporating contrastive learning and masked graph learning objectives, we emphasize evolutional information from molecular scaffolds while capturing side-chain information. Our framework achieves state-of-the-art (SOTA) results in various downstream tasks related to natural product mining and drug discovery. We first compare taxonomy classification with synthesized molecule-focused baselines to demonstrate that current models are inadequate for understanding natural synthesis. Furthermore, by diving into a fine-grained analysis at both the gene and microbial levels, NaFM demonstrates the ability to capture evolutionary information. Eventually, our method is experimented with virtual screening, illustrating informative natural product representations that can lead to more effective identification of potential drug candidates.
☆ A Modular Dataset to Demonstrate LLM Abstraction Capability ACL 2025
Large language models (LLMs) exhibit impressive capabilities but struggle with reasoning errors due to hallucinations and flawed logic. To investigate their internal representations of reasoning, we introduce ArrangementPuzzle, a novel puzzle dataset with structured solutions and automated stepwise correctness verification. We trained a classifier model on LLM activations on this dataset and found that it achieved over 80% accuracy in predicting reasoning correctness, implying that LLMs internally distinguish between correct and incorrect reasoning steps, with the strongest representations in middle-late Transformer layers. Further analysis reveals that LLMs encode abstract reasoning concepts within the middle activation layers of the transformer architecture, distinguishing logical from semantic equivalence. These findings provide insights into LLM reasoning mechanisms and contribute to improving AI reliability and interpretability, thereby offering the possibility to manipulate and refine LLM reasoning.
comment: 7 pages, 5 figures. Submitted to ACL 2025
☆ On The Sample Complexity Bounds In Bilevel Reinforcement Learning
Bilevel reinforcement learning (BRL) has emerged as a powerful mathematical framework for studying generative AI alignment and related problems. While several principled algorithmic frameworks have been proposed, key theoretical foundations, particularly those related to sample complexity, remain underexplored. Understanding and deriving tight sample complexity bounds are crucial for bridging the gap between theory and practice, guiding the development of more efficient algorithms. In this work, we present the first sample complexity result for BRL, achieving a bound of $\epsilon^{-4}$. This result extends to standard bilevel optimization problems, providing an interesting theoretical contribution with practical implications. To address the computational challenges associated with hypergradient estimation in bilevel optimization, we develop a first-order Hessian-free algorithm that does not rely on costly hypergradient computations. By leveraging matrix-free techniques and constrained optimization methods, our approach ensures scalability and practicality. Our findings pave the way for improved methods in AI alignment and other fields reliant on bilevel optimization.
☆ FairFlow: Mitigating Dataset Biases through Undecided Learning EMNLP 2024
Language models are prone to dataset biases, known as shortcuts and spurious correlations in data, which often result in performance drop on new data. We present a new debiasing framework called ``FairFlow'' that mitigates dataset biases by learning to be undecided in its predictions for data samples or representations associated with known or unknown biases. The framework introduces two key components: a suite of data and model perturbation operations that generate different biased views of input samples, and a contrastive objective that learns debiased and robust representations from the resulting biased views of samples. Experiments show that FairFlow outperforms existing debiasing methods, particularly against out-of-domain and hard test samples without compromising the in-domain performance
comment: EMNLP 2024
☆ Transferable Latent-to-Latent Locomotion Policy for Efficient and Versatile Motion Control of Diverse Legged Robots
Reinforcement learning (RL) has demonstrated remarkable capability in acquiring robot skills, but learning each new skill still requires substantial data collection for training. The pretrain-and-finetune paradigm offers a promising approach for efficiently adapting to new robot entities and tasks. Inspired by the idea that acquired knowledge can accelerate learning new tasks with the same robot and help a new robot master a trained task, we propose a latent training framework where a transferable latent-to-latent locomotion policy is pretrained alongside diverse task-specific observation encoders and action decoders. This policy in latent space processes encoded latent observations to generate latent actions to be decoded, with the potential to learn general abstract motion skills. To retain essential information for decision-making and control, we introduce a diffusion recovery module that minimizes information reconstruction loss during pretrain stage. During fine-tune stage, the pretrained latent-to-latent locomotion policy remains fixed, while only the lightweight task-specific encoder and decoder are optimized for efficient adaptation. Our method allows a robot to leverage its own prior experience across different tasks as well as the experience of other morphologically diverse robots to accelerate adaptation. We validate our approach through extensive simulations and real-world experiments, demonstrating that the pretrained latent-to-latent locomotion policy effectively generalizes to new robot entities and tasks with improved efficiency.
☆ AI-Based Screening for Depression and Social Anxiety Through Eye Tracking: An Exploratory Study
Well-being is a dynamic construct that evolves over time and fluctuates within individuals, presenting challenges for accurate quantification. Reduced well-being is often linked to depression or anxiety disorders, which are characterised by biases in visual attention towards specific stimuli, such as human faces. This paper introduces a novel approach to AI-assisted screening of affective disorders by analysing visual attention scan paths using convolutional neural networks (CNNs). Data were collected from two studies examining (1) attentional tendencies in individuals diagnosed with major depression and (2) social anxiety. These data were processed using residual CNNs through images generated from eye-gaze patterns. Experimental results, obtained with ResNet architectures, demonstrated an average accuracy of 48% for a three-class system and 62% for a two-class system. Based on these exploratory findings, we propose that this method could be employed in rapid, ecological, and effective mental health screening systems to assess well-being through eye-tracking.
comment: 17 pages, 11 figures
☆ Unraveling Pedestrian Fatality Patterns: A Comparative Study with Explainable AI
Road fatalities pose significant public safety and health challenges worldwide, with pedestrians being particularly vulnerable in vehicle-pedestrian crashes due to disparities in physical and performance characteristics. This study employs explainable artificial intelligence (XAI) to identify key factors contributing to pedestrian fatalities across the five U.S. states with the highest crash rates (2018-2022). It compares them to the five states with the lowest fatality rates. Using data from the Fatality Analysis Reporting System (FARS), the study applies machine learning techniques-including Decision Trees, Gradient Boosting Trees, Random Forests, and XGBoost-to predict contributing factors to pedestrian fatalities. To address data imbalance, the Synthetic Minority Over-sampling Technique (SMOTE) is utilized, while SHapley Additive Explanations (SHAP) values enhance model interpretability. The results indicate that age, alcohol and drug use, location, and environmental conditions are significant predictors of pedestrian fatalities. The XGBoost model outperformed others, achieving a balanced accuracy of 98 %, accuracy of 90 %, precision of 92 %, recall of 90 %, and an F1 score of 91 %. Findings reveal that pedestrian fatalities are more common in mid-block locations and areas with poor visibility, with older adults and substance-impaired individuals at higher risk. These insights can inform policymakers and urban planners in implementing targeted safety measures, such as improved lighting, enhanced pedestrian infrastructure, and stricter traffic law enforcement, to reduce fatalities and improve public safety.
comment: 22 pages, 5 figures
☆ OmniScience: A Domain-Specialized LLM for Scientific Reasoning and Discovery
Large Language Models (LLMs) have demonstrated remarkable potential in advancing scientific knowledge and addressing complex challenges. In this work, we introduce OmniScience, a specialized large reasoning model for general science, developed through three key components: (1) domain adaptive pretraining on a carefully curated corpus of scientific literature, (2) instruction tuning on a specialized dataset to guide the model in following domain-specific tasks, and (3) reasoning-based knowledge distillation through fine-tuning to significantly enhance its ability to generate contextually relevant and logically sound responses. We demonstrate the versatility of OmniScience by developing a battery agent that efficiently ranks molecules as potential electrolyte solvents or additives. Comprehensive evaluations reveal that OmniScience is competitive with state-of-the-art large reasoning models on the GPQA Diamond and domain-specific battery benchmarks, while outperforming all public reasoning and non-reasoning models with similar parameter counts. We further demonstrate via ablation experiments that domain adaptive pretraining and reasoning-based knowledge distillation are critical to attain our performance levels, across benchmarks.
☆ A Generative Caching System for Large Language Models
Caching has the potential to be of significant benefit for accessing large language models (LLMs) due to their high latencies which typically range from a small number of seconds to well over a minute. Furthermore, many LLMs charge money for queries; caching thus has a clear monetary benefit. This paper presents a new caching system for improving user experiences with LLMs. In addition to reducing both latencies and monetary costs for accessing LLMs, our system also provides important features that go beyond the performance benefits typically associated with caches. A key feature we provide is generative caching, wherein multiple cached responses can be synthesized to provide answers to queries which have never been seen before. Our generative caches function as repositories of valuable information which can be mined and analyzed. We also improve upon past semantic caching techniques by tailoring the caching algorithms to optimally balance cost and latency reduction with the quality of responses provided. Performance tests indicate that our caches are considerably faster than GPTcache.
☆ GPBench: A Comprehensive and Fine-Grained Benchmark for Evaluating Large Language Models as General Practitioners
General practitioners (GPs) serve as the cornerstone of primary healthcare systems by providing continuous and comprehensive medical services. However, due to community-oriented nature of their practice, uneven training and resource gaps, the clinical proficiency among GPs can vary significantly across regions and healthcare settings. Currently, Large Language Models (LLMs) have demonstrated great potential in clinical and medical applications, making them a promising tool for supporting general practice. However, most existing benchmarks and evaluation frameworks focus on exam-style assessments-typically multiple-choice question-lack comprehensive assessment sets that accurately mirror the real-world scenarios encountered by GPs. To evaluate how effectively LLMs can make decisions in the daily work of GPs, we designed GPBench, which consists of both test questions from clinical practice and a novel evaluation framework. The test set includes multiple-choice questions that assess fundamental knowledge of general practice, as well as realistic, scenario-based problems. All questions are meticulously annotated by experts, incorporating rich fine-grained information related to clinical management. The proposed LLM evaluation framework is based on the competency model for general practice, providing a comprehensive methodology for assessing LLM performance in real-world settings. As the first large-model evaluation set targeting GP decision-making scenarios, GPBench allows us to evaluate current mainstream LLMs. Expert assessment and evaluation reveal that in areas such as disease staging, complication recognition, treatment detail, and medication usage, these models exhibit at least ten major shortcomings. Overall, existing LLMs are not yet suitable for independent use in real-world GP working scenarios without human oversight.
☆ ConSol: Sequential Probability Ratio Testing to Find Consistent LLM Reasoning Paths Efficiently
Recent advancements in large language models (LLMs) integrating explicit reasoning, such as OpenAI's o3-mini, DeepSeek-R1, and QWQ-32B, enable smaller models to solve complex tasks by generating intermediate reasoning steps prior to providing answers. However, this approach significantly increases computational costs, both monetarily and environmentally. The widely-used self-consistency method further exacerbates these costs by aggregating multiple reasoning paths to improve accuracy, often requiring between 40 to 64 samples per task. Although aggregation effectively reduces variance and bias, additional sampling can lead to diminishing returns when early samples yield consistent results. To address inefficiencies, we propose leveraging Sequential Probability Ratio Testing (SPRT) to dynamically terminate sampling once sufficient consistency is achieved. We calibrate SPRT parameters specifically for LLM applications, accounting for sensitivity to detect the mode of the distribution. Our experiments demonstrate that incorporating SPRT significantly enhances token efficiency, achieving comparable accuracy to self-consistency methods but at a substantially reduced computational cost. To promote transparency and facilitate reproducibility, we have made the source code and datasets used in our experiments publicly available at our GitHub repository: https://github.com/LiuzLab/consol, or available as a PyPI package: pip install consol. We hope that this resource will support further research and encourage the development of new methods building upon our work.
♻ ☆ Open Models, Closed Minds? On Agents Capabilities in Mimicking Human Personalities through Open Large Language Models AAAI 2025
The emergence of unveiling human-like behaviors in Large Language Models (LLMs) has led to a closer connection between NLP and human psychology. Scholars have been studying the inherent personalities exhibited by LLMs and attempting to incorporate human traits and behaviors into them. However, these efforts have primarily focused on commercially-licensed LLMs, neglecting the widespread use and notable advancements seen in Open LLMs. This work aims to address this gap by employing a set of 12 LLM Agents based on the most representative Open models and subject them to a series of assessments concerning the Myers-Briggs Type Indicator (MBTI) test and the Big Five Inventory (BFI) test. Our approach involves evaluating the intrinsic personality traits of Open LLM agents and determining the extent to which these agents can mimic human personalities when conditioned by specific personalities and roles. Our findings unveil that $(i)$ each Open LLM agent showcases distinct human personalities; $(ii)$ personality-conditioned prompting produces varying effects on the agents, with only few successfully mirroring the imposed personality, while most of them being ``closed-minded'' (i.e., they retain their intrinsic traits); and $(iii)$ combining role and personality conditioning can enhance the agents' ability to mimic human personalities. Our work represents a step up in understanding the dense relationship between NLP and human psychology through the lens of Open LLMs.
comment: Accepted and presented at the AAAI 2025 Conference. CHANGES in version v2: (i) Enhanced methodology and evaluation based on BFI in addition to MBTI, with expanded set of LLM agents; (ii) author list changed w.r.t. version (v1), see Acknowledgements
♻ ☆ 1000 Layer Networks for Self-Supervised RL: Scaling Depth Can Enable New Goal-Reaching Capabilities
Scaling up self-supervised learning has driven breakthroughs in language and vision, yet comparable progress has remained elusive in reinforcement learning (RL). In this paper, we study building blocks for self-supervised RL that unlock substantial improvements in scalability, with network depth serving as a critical factor. Whereas most RL papers in recent years have relied on shallow architectures (around 2 - 5 layers), we demonstrate that increasing the depth up to 1024 layers can significantly boost performance. Our experiments are conducted in an unsupervised goal-conditioned setting, where no demonstrations or rewards are provided, so an agent must explore (from scratch) and learn how to maximize the likelihood of reaching commanded goals. Evaluated on simulated locomotion and manipulation tasks, our approach increases performance by $2\times$ - $50\times$. Increasing the model depth not only increases success rates but also qualitatively changes the behaviors learned.
comment: Link to project website: https://wang-kevin3290.github.io/scaling-crl/
♻ ☆ Proactive and Reactive Constraint Programming for Stochastic Project Scheduling with Maximal Time-Lags
This study investigates scheduling strategies for the stochastic resource-constrained project scheduling problem with maximal time lags (SRCPSP/max)). Recent advances in Constraint Programming (CP) and Temporal Networks have reinvoked interest in evaluating the advantages and drawbacks of various proactive and reactive scheduling methods. First, we present a new, CP-based fully proactive method. Second, we show how a reactive approach can be constructed using an online rescheduling procedure. A third contribution is based on partial order schedules and uses Simple Temporal Networks with Uncertainty (STNUs). Our statistical analysis shows that the STNU-based algorithm performs best in terms of solution quality, while also showing good relative offline and online computation time.
♻ ☆ LLM+KG@VLDB'24 Workshop Summary SIGMOD
The unification of large language models (LLMs) and knowledge graphs (KGs) has emerged as a hot topic. At the LLM+KG'24 workshop, held in conjunction with VLDB 2024 in Guangzhou, China, one of the key themes explored was important data management challenges and opportunities due to the effective interaction between LLMs and KGs. This report outlines the major directions and approaches presented by various speakers during the LLM+KG'24 workshop.
comment: accepted at ACM SIGMOD Record 2025
♻ ☆ Plurals: A System for Guiding LLMs Via Simulated Social Ensembles
Recent debates raised concerns that language models may favor certain viewpoints. But what if the solution is not to aim for a 'view from nowhere' but rather to leverage different viewpoints? We introduce Plurals, a system and Python library for pluralistic AI deliberation. Plurals consists of Agents (LLMs, optionally with personas) which deliberate within customizable Structures, with Moderators overseeing deliberation. Plurals is a generator of simulated social ensembles. Plurals integrates with government datasets to create nationally representative personas, includes deliberation templates inspired by deliberative democracy, and allows users to customize both information-sharing structures and deliberation behavior within Structures. Six case studies demonstrate fidelity to theoretical constructs and efficacy. Three randomized experiments show simulated focus groups produced output resonant with an online sample of the relevant audiences (chosen over zero-shot generation in 75% of trials). Plurals is both a paradigm and a concrete system for pluralistic AI. The Plurals library is available at https://github.com/josh-ashkinaze/plurals and will be continually updated.
comment: CHI 2025
♻ ☆ Active management of battery degradation in wireless sensor network using deep reinforcement learning for group battery replacement
Wireless sensor networks (WSNs) have become a promising solution for structural health monitoring (SHM), especially in hard-to-reach or remote locations. Battery-powered WSNs offer various advantages over wired systems, however limited battery life has always been one of the biggest obstacles in practical use of the WSNs, regardless of energy harvesting methods. While various methods have been studied for battery health management, existing methods exclusively aim to extend lifetime of individual batteries, lacking a system level view. A consequence of applying such methods is that batteries in a WSN tend to fail at different times, posing significant difficulty on planning and scheduling of battery replacement trip. This study investigate a deep reinforcement learning (DRL) method for active battery degradation management by optimizing duty cycle of WSNs at the system level. This active management strategy effectively reduces earlier failure of battery individuals which enable group replacement without sacrificing WSN performances. A simulated environment based on a real-world WSN setup was developed to train a DRL agent and learn optimal duty cycle strategies. The performance of the strategy was validated in a long-term setup with various network sizes, demonstrating its efficiency and scalability.
♻ ☆ Beyond Label Attention: Transparency in Language Models for Automated Medical Coding via Dictionary Learning
Medical coding, the translation of unstructured clinical text into standardized medical codes, is a crucial but time-consuming healthcare practice. Though large language models (LLM) could automate the coding process and improve the efficiency of such tasks, interpretability remains paramount for maintaining patient trust. Current efforts in interpretability of medical coding applications rely heavily on label attention mechanisms, which often leads to the highlighting of extraneous tokens irrelevant to the ICD code. To facilitate accurate interpretability in medical language models, this paper leverages dictionary learning that can efficiently extract sparsely activated representations from dense language model embeddings in superposition. Compared with common label attention mechanisms, our model goes beyond token-level representations by building an interpretable dictionary which enhances the mechanistic-based explanations for each ICD code prediction, even when the highlighted tokens are medically irrelevant. We show that dictionary features can steer model behavior, elucidate the hidden meanings of upwards of 90% of medically irrelevant tokens, and are human interpretable.
comment: https://aclanthology.org/2024.emnlp-main.500/
♻ ☆ ULTra: Unveiling Latent Token Interpretability in Transformer-Based Understanding and Segmentation
Transformers have revolutionized Computer Vision (CV) through self-attention mechanisms. However, their complexity makes latent token representations difficult to interpret. We introduce ULTra, a framework for interpreting Transformer embeddings and uncovering meaningful semantic patterns within them. ULTra enables unsupervised semantic segmentation using pre-trained models without requiring fine-tuning. Additionally, we propose a self-supervised training approach that refines segmentation performance by learning an external transformation matrix without modifying the underlying model. Our method achieves state-of-the-art performance in unsupervised semantic segmentation, outperforming existing segmentation methods. Furthermore, we validate ULTra for model interpretation on both synthetic and real-world scenarios, including Object Selection and interpretable text summarization using LLMs, demonstrating its broad applicability in explaining the semantic structure of latent token representations.
♻ ☆ How To Think About End-To-End Encryption and AI: Training, Processing, Disclosure, and Consent
End-to-end encryption (E2EE) has become the gold standard for securing communications, bringing strong confidentiality and privacy guarantees to billions of users worldwide. However, the current push towards widespread integration of artificial intelligence (AI) models, including in E2EE systems, raises some serious security concerns. This work performs a critical examination of the (in)compatibility of AI models and E2EE applications. We explore this on two fronts: (1) the integration of AI "assistants" within E2EE applications, and (2) the use of E2EE data for training AI models. We analyze the potential security implications of each, and identify conflicts with the security guarantees of E2EE. Then, we analyze legal implications of integrating AI models in E2EE applications, given how AI integration can undermine the confidentiality that E2EE promises. Finally, we offer a list of detailed recommendations based on our technical and legal analyses, including: technical design choices that must be prioritized to uphold E2EE security; how service providers must accurately represent E2EE security; and best practices for the default behavior of AI features and for requesting user consent. We hope this paper catalyzes an informed conversation on the tensions that arise between the brisk deployment of AI and the security offered by E2EE, and guides the responsible development of new AI features.
♻ ☆ To Google or To ChatGPT? A Comparison of CS2 Students' Information Gathering Approaches and Outcomes
LLMs such as ChatGPT have been widely adopted by students in higher education as tools for learning programming and related concepts. However, it remains unclear how effective students are and what strategies students use while learning with LLMs. Since the majority of students' experiences in online self-learning have come through using search engines such as Google, evaluating AI tools in this context can help us address these gaps. In this mixed methods research, we conducted an exploratory within-subjects study to understand how CS2 students learn programming concepts using both LLMs as well as traditional online methods such as educational websites and videos to examine how students approach learning within and across both scenarios. We discovered that students found it easier to learn a more difficult concept using traditional methods than using ChatGPT. We also found that students ask fewer follow-ups and use more keyword-based queries for search engines while their prompts to LLMs tend to explicitly ask for information.
♻ ☆ SynMorph: Generating Synthetic Face Morphing Dataset with Mated Samples
Face morphing attack detection (MAD) algorithms have become essential to overcome the vulnerability of face recognition systems. To solve the lack of large-scale and public-available datasets due to privacy concerns and restrictions, in this work we propose a new method to generate a synthetic face morphing dataset with 2450 identities and more than 100k morphs. The proposed synthetic face morphing dataset is unique for its high-quality samples, different types of morphing algorithms, and the generalization for both single and differential morphing attack detection algorithms. For experiments, we apply face image quality assessment and vulnerability analysis to evaluate the proposed synthetic face morphing dataset from the perspective of biometric sample quality and morphing attack potential on face recognition systems. The results are benchmarked with an existing SOTA synthetic dataset and a representative non-synthetic and indicate improvement compared with the SOTA. Additionally, we design different protocols and study the applicability of using the proposed synthetic dataset on training morphing attack detection algorithms.
comment: This preprint has been further published in IEEE Access. Print ISSN: 2169-3536. Online ISSN: 2169-3536. Digital Object Identifier: 10.1109/ACCESS.2025.3548957
♻ ☆ Don't Kill the Baby: The Case for AI in Arbitration
Since the introduction of Generative AI (GenAI) in 2022, its ability to simulate human intelligence and generate content has sparked both enthusiasm and concern. While much criticism focuses on AI's potential to perpetuate bias, create emotional dissonance, displace jobs, and raise ethical questions, these concerns often overlook the practical benefits of AI, particularly in legal contexts. This article examines the integration of AI into arbitration, arguing that the Federal Arbitration Act (FAA) allows parties to contractually choose AI-driven arbitration, despite traditional reservations. The article makes three key contributions: (1) It shifts the focus from debates over AI's personhood to the practical aspects of incorporating AI into arbitration, asserting that AI can effectively serve as an arbitrator if both parties agree; (2) It positions arbitration as an ideal starting point for broader AI adoption in the legal field, given its flexibility and the autonomy it grants parties to define their standards of fairness; and (3) It outlines future research directions, emphasizing the importance of empirically comparing AI and human arbitration, which could lead to the development of distinct systems. By advocating for the use of AI in arbitration, this article underscores the importance of respecting contractual autonomy and creating an environment that allows AI's potential to be fully realized. Drawing on the insights of Judge Richard Posner, the article argues that the ethical obligations of AI in arbitration should be understood within the context of its technological strengths and the voluntary nature of arbitration agreements. Ultimately, it calls for a balanced, open-minded approach to AI in arbitration, recognizing its potential to enhance the efficiency, fairness, and flexibility of dispute resolution
♻ ☆ LAYOUTDREAMER: Physics-guided Layout for Text-to-3D Compositional Scene Generation
Recently, the field of text-guided 3D scene generation has garnered significant attention. High-quality generation that aligns with physical realism and high controllability is crucial for practical 3D scene applications. However, existing methods face fundamental limitations: (i) difficulty capturing complex relationships between multiple objects described in the text, (ii) inability to generate physically plausible scene layouts, and (iii) lack of controllability and extensibility in compositional scenes. In this paper, we introduce LayoutDreamer, a framework that leverages 3D Gaussian Splatting (3DGS) to facilitate high-quality, physically consistent compositional scene generation guided by text. Specifically, given a text prompt, we convert it into a directed scene graph and adaptively adjust the density and layout of the initial compositional 3D Gaussians. Subsequently, dynamic camera adjustments are made based on the training focal point to ensure entity-level generation quality. Finally, by extracting directed dependencies from the scene graph, we tailor physical and layout energy to ensure both realism and flexibility. Comprehensive experiments demonstrate that LayoutDreamer outperforms other compositional scene generation quality and semantic alignment methods. Specifically, it achieves state-of-the-art (SOTA) performance in the multiple objects generation metric of T3Bench.
♻ ☆ Enhancing Layer Attention Efficiency through Pruning Redundant Retrievals
Growing evidence suggests that layer attention mechanisms, which enhance interaction among layers in deep neural networks, have significantly advanced network architectures. However, existing layer attention methods suffer from redundancy, as attention weights learned by adjacent layers often become highly similar. This redundancy causes multiple layers to extract nearly identical features, reducing the model's representational capacity and increasing training time. To address this issue, we propose a novel approach to quantify redundancy by leveraging the Kullback-Leibler (KL) divergence between adjacent layers. Additionally, we introduce an Enhanced Beta Quantile Mapping (EBQM) method that accurately identifies and skips redundant layers, thereby maintaining model stability. Our proposed Efficient Layer Attention (ELA) architecture, improves both training efficiency and overall performance, achieving a 30\% reduction in training time while enhancing performance in tasks such as image classification and object detection.
comment: 11 pages, 7 figures
♻ ☆ Bimodal Connection Attention Fusion for Speech Emotion Recognition
Multi-modal emotion recognition is challenging due to the difficulty of extracting features that capture subtle emotional differences. Understanding multi-modal interactions and connections is key to building effective bimodal speech emotion recognition systems. In this work, we propose Bimodal Connection Attention Fusion (BCAF) method, which includes three main modules: the interactive connection network, the bimodal attention network, and the correlative attention network. The interactive connection network uses an encoder-decoder architecture to model modality connections between audio and text while leveraging modality-specific features. The bimodal attention network enhances semantic complementation and exploits intra- and inter-modal interactions. The correlative attention network reduces cross-modal noise and captures correlations between audio and text. Experiments on the MELD and IEMOCAP datasets demonstrate that the proposed BCAF method outperforms existing state-of-the-art baselines.
♻ ☆ Score matching through the roof: linear, nonlinear, and latent variables causal discovery
Causal discovery from observational data holds great promise, but existing methods rely on strong assumptions about the underlying causal structure, often requiring full observability of all relevant variables. We tackle these challenges by leveraging the score function $\nabla \log p(X)$ of observed variables for causal discovery and propose the following contributions. First, we fine-tune the existing identifiability results with the score on additive noise models, showing that their assumption of nonlinearity of the causal mechanisms is not necessary. Second, we establish conditions for inferring causal relations from the score even in the presence of hidden variables; this result is two-faced: we demonstrate the score's potential to infer the equivalence class of causal graphs with hidden variables (while previous results are restricted to the fully observable setting), and we provide sufficient conditions for identifying direct causes in latent variable models. Building on these insights, we propose a flexible algorithm suited for causal discovery on linear, nonlinear, and latent variable models, which we empirically validate.
♻ ☆ NavCoT: Boosting LLM-Based Vision-and-Language Navigation via Learning Disentangled Reasoning
Vision-and-Language Navigation (VLN), as a crucial research problem of Embodied AI, requires an embodied agent to navigate through complex 3D environments following natural language instructions. Recent research has highlighted the promising capacity of large language models (LLMs) in VLN by improving navigational reasoning accuracy and interpretability. However, their predominant use in an offline manner usually suffers from substantial domain gap between the VLN task and the LLM training corpus. This paper introduces a novel strategy called Navigational Chain-of-Thought (NavCoT), where we fulfill parameter-efficient in-domain training to enable self-guided navigational decision, leading to a significant mitigation of the domain gap in a cost-effective manner. Specifically, at each timestep, the LLM is prompted to forecast the navigational chain-of-thought by: 1) acting as a world model to imagine the next observation according to the instruction, 2) selecting the candidate observation that best aligns with the imagination, and 3) determining the action based on the reasoning from the prior steps. Through constructing formalized labels for training, the LLM can learn to generate desired and reasonable chain-of-thought outputs for improving the action decision. Experimental results across various training settings and popular VLN benchmarks (e.g., Room-to-Room (R2R), Room-across-Room (RxR), Room-for-Room (R4R)) show the significant superiority of NavCoT over the direct action prediction variants. Through simple parameter-efficient finetuning, our NavCoT outperforms a recent GPT4-based approach with ~7% relative improvement on the R2R dataset. We believe that NavCoT will help unlock more task-adaptive and scalable LLM-based embodied agents, which are helpful for developing real-world robotics applications. Code is available at https://github.com/expectorlin/NavCoT.
comment: Accepted by TPAMI 2025
♻ ☆ Promoting Segment Anything Model towards Highly Accurate Dichotomous Image Segmentation
The Segment Anything Model (SAM) represents a significant breakthrough into foundation models for computer vision, providing a large-scale image segmentation model. However, despite SAM's zero-shot performance, its segmentation masks lack fine-grained details, particularly in accurately delineating object boundaries. Therefore, it is both interesting and valuable to explore whether SAM can be improved towards highly accurate object segmentation, which is known as the dichotomous image segmentation (DIS) task. To address this issue, we propose DIS-SAM, which advances SAM towards DIS with extremely accurate details. DIS-SAM is a framework specifically tailored for highly accurate segmentation, maintaining SAM's promptable design. DIS-SAM employs a two-stage approach, integrating SAM with a modified advanced network that was previously designed to handle the prompt-free DIS task. To better train DIS-SAM, we employ a ground truth enrichment strategy by modifying original mask annotations.
♻ ☆ LoRA-Pro: Are Low-Rank Adapters Properly Optimized? ICLR 2025
Low-rank adaptation, also known as LoRA, has emerged as a prominent method for parameter-efficient fine-tuning of foundation models. Despite its computational efficiency, LoRA still yields inferior performance compared to full fine-tuning. In this paper, we first uncover a fundamental connection between the optimization processes of LoRA and full fine-tuning: using LoRA for optimization is mathematically equivalent to full fine-tuning using a low-rank gradient for parameter updates. And this low-rank gradient can be expressed in terms of the gradients of the two low-rank matrices in LoRA. Leveraging this insight, we introduce LoRA-Pro, a method that enhances LoRA's performance by strategically adjusting the gradients of these low-rank matrices. This adjustment allows the low-rank gradient to more accurately approximate the full fine-tuning gradient, thereby narrowing the performance gap between LoRA and full fine-tuning. Furthermore, we theoretically derive the optimal solutions for adjusting the gradients of the low-rank matrices, applying them during fine-tuning in LoRA-Pro. We conduct extensive experiments across natural language understanding, dialogue generation, mathematical reasoning, code generation, and image classification tasks, demonstrating that LoRA-Pro substantially improves LoRA's performance, effectively narrowing the gap with full fine-tuning. Code is publicly available at https://github.com/mrflogs/LoRA-Pro.
comment: Camera-Ready Version for ICLR 2025; technical corrections to previous version
♻ ☆ Demo-Craft: Using In-Context Learning to Improve Code Generation in Large Language Models
Generating executable code from natural language instructions using Large Language Models (LLMs) poses challenges such as semantic ambiguity and understanding taskspecific contexts. To address these issues, we propose a system called DemoCraft, which enhances code generation by leveraging in-context learning and demonstration selection, combined with latent concept learning. Latent concept learning introduces additional concept tokens, which are trainable embeddings that capture task-specific knowledge. We then test our system on two major datasets: MBPP and Humaneval. Our experimental results demonstrate that the proposed system achieves an approximate 2x increase in the pass@k metric compared to baseline models. Furthermore, we introduce two novel evaluation metrics: correctness@k and similarity@k. Our empirical studies indicate that our system attains nearly a 3x improvement in these metrics as well.
comment: Accepted at IEEE ICIITCEE 2025. Presented on 16th January 2025 in Bengaluru, India
♻ ☆ VisEscape: A Benchmark for Evaluating Exploration-driven Decision-making in Virtual Escape Rooms
Escape rooms present a unique cognitive challenge that demands exploration-driven planning: players should actively search their environment, continuously update their knowledge based on new discoveries, and connect disparate clues to determine which elements are relevant to their objectives. Motivated by this, we introduce VisEscape, a benchmark of 20 virtual escape rooms specifically designed to evaluate AI models under these challenging conditions, where success depends not only on solving isolated puzzles but also on iteratively constructing and refining spatial-temporal knowledge of a dynamically changing environment. On VisEscape, we observe that even state-of-the-art multimodal models generally fail to escape the rooms, showing considerable variation in their levels of progress and trajectories. To address this issue, we propose VisEscaper, which effectively integrates Memory, Feedback, and ReAct modules, demonstrating significant improvements by performing 3.7 times more effectively and 4.9 times more efficiently on average compared to baseline agents.
♻ ☆ Privacy-Preserving Federated Learning with Differentially Private Hyperdimensional Computing
Federated Learning (FL) has become a key method for preserving data privacy in Internet of Things (IoT) environments, as it trains Machine Learning (ML) models locally while transmitting only model updates. Despite this design, FL remains susceptible to threats such as model inversion and membership inference attacks, which can reveal private training data. Differential Privacy (DP) techniques are often introduced to mitigate these risks, but simply injecting DP noise into black-box ML models can compromise accuracy, particularly in dynamic IoT contexts, where continuous, lifelong learning leads to excessive noise accumulation. To address this challenge, we propose Federated HyperDimensional computing with Privacy-preserving (FedHDPrivacy), an eXplainable Artificial Intelligence (XAI) framework that integrates neuro-symbolic computing and DP. Unlike conventional approaches, FedHDPrivacy actively monitors the cumulative noise across learning rounds and adds only the additional noise required to satisfy privacy constraints. In a real-world application for monitoring manufacturing machining processes, FedHDPrivacy maintains high performance while surpassing standard FL frameworks - Federated Averaging (FedAvg), Federated Proximal (FedProx), Federated Normalized Averaging (FedNova), and Federated Optimization (FedOpt) - by up to 37%. Looking ahead, FedHDPrivacy offers a promising avenue for further enhancements, such as incorporating multimodal data fusion.
comment: 31 Pages, 14 Figures
♻ ☆ A Comprehensive Survey on Self-Interpretable Neural Networks
Neural networks have achieved remarkable success across various fields. However, the lack of interpretability limits their practical use, particularly in critical decision-making scenarios. Post-hoc interpretability, which provides explanations for pre-trained models, is often at risk of robustness and fidelity. This has inspired a rising interest in self-interpretable neural networks, which inherently reveal the prediction rationale through the model structures. Although there exist surveys on post-hoc interpretability, a comprehensive and systematic survey of self-interpretable neural networks is still missing. To address this gap, we first collect and review existing works on self-interpretable neural networks and provide a structured summary of their methodologies from five key perspectives: attribution-based, function-based, concept-based, prototype-based, and rule-based self-interpretation. We also present concrete, visualized examples of model explanations and discuss their applicability across diverse scenarios, including image, text, graph data, and deep reinforcement learning. Additionally, we summarize existing evaluation metrics for self-interpretability and identify open challenges in this field, offering insights for future research. To support ongoing developments, we present a publicly accessible resource to track advancements in this domain: https://github.com/yangji721/Awesome-Self-Interpretable-Neural-Network.
♻ ☆ Towards Dynamic Trend Filtering through Trend Point Detection with Reinforcement Learning
Trend filtering simplifies complex time series data by applying smoothness to filter out noise while emphasizing proximity to the original data. However, existing trend filtering methods fail to reflect abrupt changes in the trend due to `approximateness,' resulting in constant smoothness. This approximateness uniformly filters out the tail distribution of time series data, characterized by extreme values, including both abrupt changes and noise. In this paper, we propose Trend Point Detection formulated as a Markov Decision Process (MDP), a novel approach to identifying essential points that should be reflected in the trend, departing from approximations. We term these essential points as Dynamic Trend Points (DTPs) and extract trends by interpolating them. To identify DTPs, we utilize Reinforcement Learning (RL) within a discrete action space and a forecasting sum-of-squares loss function as a reward, referred to as the Dynamic Trend Filtering network (DTF-net). DTF-net integrates flexible noise filtering, preserving critical original subsequences while removing noise as required for other subsequences. We demonstrate that DTF-net excels at capturing abrupt changes compared to other trend filtering algorithms and enhances forecasting performance, as abrupt changes are predicted rather than smoothed out.
comment: 18 pages, 11 figures
♻ ☆ Debunking the CUDA Myth Towards GPU-based AI Systems ISCA-52
This paper presents a comprehensive evaluation of Intel Gaudi NPUs as an alternative to NVIDIA GPUs, which is currently the de facto standard in AI system design. First, we create a suite of microbenchmarks to compare Intel Gaudi-2 with NVIDIA A100, showing that Gaudi-2 achieves competitive performance not only in primitive AI compute, memory, and communication operations but also in executing several important AI workloads end-to-end. We then assess Gaudi NPU's programmability by discussing several software-level optimization strategies to employ for implementing critical FBGEMM operators and vLLM, evaluating their efficiency against GPU-optimized counterparts. Results indicate that Gaudi-2 achieves energy efficiency comparable to A100, though there are notable areas for improvement in terms of software maturity. Overall, we conclude that, with effective integration into high-level AI frameworks, Gaudi NPUs could challenge NVIDIA GPU's dominance in the AI server market, though further improvements are necessary to fully compete with NVIDIA's robust software ecosystem.
comment: Accepted for publication at the 52nd IEEE/ACM International Symposium on Computer Architecture (ISCA-52), 2025
♻ ☆ BAMDP Shaping: a Unified Framework for Intrinsic Motivation and Reward Shaping ICLR 2025
Intrinsic motivation and reward shaping guide reinforcement learning (RL) agents by adding pseudo-rewards, which can lead to useful emergent behaviors. However, they can also encourage counterproductive exploits, e.g., fixation with noisy TV screens. Here we provide a theoretical model which anticipates these behaviors, and provides broad criteria under which adverse effects can be bounded. We characterize all pseudo-rewards as reward shaping in Bayes-Adaptive Markov Decision Processes (BAMDPs), which formulates the problem of learning in MDPs as an MDP over the agent's knowledge. Optimal exploration maximizes BAMDP state value, which we decompose into the value of the information gathered and the prior value of the physical state. Psuedo-rewards guide RL agents by rewarding behavior that increases these value components, while they hinder exploration when they align poorly with the actual value. We extend potential-based shaping theory to prove BAMDP Potential-based shaping Functions (BAMPFs) are immune to reward-hacking (convergence to behaviors maximizing composite rewards to the detriment of real rewards) in meta-RL, and show empirically how a BAMPF helps a meta-RL agent learn optimal RL algorithms for a Bernoulli Bandit domain. We finally prove that BAMPFs with bounded monotone increasing potentials also resist reward-hacking in the regular RL setting. We show that it is straightforward to retrofit or design new pseudo-reward terms in this form, and provide an empirical demonstration in the Mountain Car environment.
comment: Published at ICLR 2025
♻ ☆ Where are we in audio deepfake detection? A systematic analysis over generative and detection models
Recent advances in Text-to-Speech (TTS) and Voice-Conversion (VC) using generative Artificial Intelligence (AI) technology have made it possible to generate high-quality and realistic human-like audio. This poses growing challenges in distinguishing AI-synthesized speech from the genuine human voice and could raise concerns about misuse for impersonation, fraud, spreading misinformation, and scams. However, existing detection methods for AI-synthesized audio have not kept pace and often fail to generalize across diverse datasets. In this paper, we introduce SONAR, a synthetic AI-Audio Detection Framework and Benchmark, aiming to provide a comprehensive evaluation for distinguishing cutting-edge AI-synthesized auditory content. SONAR includes a novel evaluation dataset sourced from 9 diverse audio synthesis platforms, including leading TTS providers and state-of-the-art TTS models. It is the first framework to uniformly benchmark AI-audio detection across both traditional and foundation model-based detection systems. Through extensive experiments, (1) we reveal the limitations of existing detection methods and demonstrate that foundation models exhibit stronger generalization capabilities, likely due to their model size and the scale and quality of pretraining data. (2) Speech foundation models demonstrate robust cross-lingual generalization capabilities, maintaining strong performance across diverse languages despite being fine-tuned solely on English speech data. This finding also suggests that the primary challenges in audio deepfake detection are more closely tied to the realism and quality of synthetic audio rather than language-specific characteristics. (3) We explore the effectiveness and efficiency of few-shot fine-tuning in improving generalization, highlighting its potential for tailored applications, such as personalized detection systems for specific entities or individuals.
♻ ☆ Coarse Set Theory for AI Ethics and Decision-Making: A Mathematical Framework for Granular Evaluations
As artificial intelligence (AI) systems become increasingly embedded in ethically sensitive domains such as education, healthcare, and transportation, the need to balance accuracy and interpretability in decision-making has become a central concern. Coarse Ethics (CE) is a theoretical framework that justifies coarse-grained evaluations, such as letter grades or warning labels, as ethically appropriate under cognitive and contextual constraints. However, CE has lacked mathematical formalization. This paper introduces Coarse Set Theory (CST), a novel mathematical framework that models coarse-grained decision-making using totally ordered structures and coarse partitions. CST defines hierarchical relations among sets and uses information-theoretic tools, such as Kullback-Leibler Divergence, to quantify the trade-off between simplification and information loss. We demonstrate CST through applications in educational grading and explainable AI (XAI), showing how it enables more transparent and context-sensitive evaluations. By grounding coarse evaluations in set theory and probabilistic reasoning, CST contributes to the ethical design of interpretable AI systems. This work bridges formal methods and human-centered ethics, offering a principled approach to balancing comprehensibility, fairness, and informational integrity in AI-driven decisions.
comment: 28 pages, 2 figures
Machine Learning 55
☆ IceBench: A Benchmark for Deep Learning based Sea Ice Type Classification
Sea ice plays a critical role in the global climate system and maritime operations, making timely and accurate classification essential. However, traditional manual methods are time-consuming, costly, and have inherent biases. Automating sea ice type classification addresses these challenges by enabling faster, more consistent, and scalable analysis. While both traditional and deep learning approaches have been explored, deep learning models offer a promising direction for improving efficiency and consistency in sea ice classification. However, the absence of a standardized benchmark and comparative study prevents a clear consensus on the best-performing models. To bridge this gap, we introduce \textit{IceBench}, a comprehensive benchmarking framework for sea ice type classification. Our key contributions are threefold: First, we establish the IceBench benchmarking framework which leverages the existing AI4Arctic Sea Ice Challenge dataset as a standardized dataset, incorporates a comprehensive set of evaluation metrics, and includes representative models from the entire spectrum of sea ice type classification methods categorized in two distinct groups, namely, pixel-based classification methods and patch-based classification methods. IceBench is open-source and allows for convenient integration and evaluation of other sea ice type classification methods; hence, facilitating comparative evaluation of new methods and improving reproducibility in the field. Second, we conduct an in-depth comparative study on representative models to assess their strengths and limitations, providing insights for both practitioners and researchers. Third, we leverage IceBench for systematic experiments addressing key research questions on model transferability across seasons (time) and locations (space), data downscaling, and preprocessing strategies.
☆ Accelerating and enhancing thermodynamic simulations of electrochemical interfaces
Electrochemical interfaces are crucial in catalysis, energy storage, and corrosion, where their stability and reactivity depend on complex interactions between the electrode, adsorbates, and electrolyte. Predicting stable surface structures remains challenging, as traditional surface Pourbaix diagrams tend to either rely on expert knowledge or costly $\textit{ab initio}$ sampling, and neglect thermodynamic equilibration with the environment. Machine learning (ML) potentials can accelerate static modeling but often overlook dynamic surface transformations. Here, we extend the Virtual Surface Site Relaxation-Monte Carlo (VSSR-MC) method to autonomously sample surface reconstructions modeled under aqueous electrochemical conditions. Through fine-tuning foundational ML force fields, we accurately and efficiently predict surface energetics, recovering known Pt(111) phases and revealing new LaMnO$_\mathrm{3}$(001) surface reconstructions. By explicitly accounting for bulk-electrolyte equilibria, our framework enhances electrochemical stability predictions, offering a scalable approach to understanding and designing materials for electrochemical applications.
comment: 19 pages main text, 5 figures, supplementary information (SI) in ancillary files
☆ Detecting and Mitigating DDoS Attacks with AI: A Survey
Distributed Denial of Service attacks represent an active cybersecurity research problem. Recent research shifted from static rule-based defenses towards AI-based detection and mitigation. This comprehensive survey covers several key topics. Preeminently, state-of-the-art AI detection methods are discussed. An in-depth taxonomy based on manual expert hierarchies and an AI-generated dendrogram are provided, thus settling DDoS categorization ambiguities. An important discussion on available datasets follows, covering data format options and their role in training AI detection methods together with adversarial training and examples augmentation. Beyond detection, AI based mitigation techniques are surveyed as well. Finally, multiple open research directions are proposed.
☆ Understanding Inverse Reinforcement Learning under Overparameterization: Non-Asymptotic Analysis and Global Optimality
The goal of the Inverse reinforcement learning (IRL) task is to identify the underlying reward function and the corresponding optimal policy from a set of expert demonstrations. While most IRL algorithms' theoretical guarantees rely on a linear reward structure, we aim to extend the theoretical understanding of IRL to scenarios where the reward function is parameterized by neural networks. Meanwhile, conventional IRL algorithms usually adopt a nested structure, leading to computational inefficiency, especially in high-dimensional settings. To address this problem, we propose the first two-timescale single-loop IRL algorithm under neural network parameterized reward and provide a non-asymptotic convergence analysis under overparameterization. Although prior optimality results for linear rewards do not apply, we show that our algorithm can identify the globally optimal reward and policy under certain neural network structures. This is the first IRL algorithm with a non-asymptotic convergence guarantee that provably achieves global optimality in neural network settings.
☆ A novel gradient-based method for decision trees optimizing arbitrary differential loss functions
There are many approaches for training decision trees. This work introduces a novel gradient-based method for constructing decision trees that optimize arbitrary differentiable loss functions, overcoming the limitations of heuristic splitting rules. Unlike traditional approaches that rely on heuristic splitting rules, the proposed method refines predictions using the first and second derivatives of the loss function, enabling the optimization of complex tasks such as classification, regression, and survival analysis. We demonstrate the method's applicability to classification, regression, and survival analysis tasks, including those with censored data. Numerical experiments on both real and synthetic datasets compare the proposed method with traditional decision tree algorithms, such as CART, Extremely Randomized Trees, and SurvTree. The implementation of the method is publicly available, providing a practical tool for researchers and practitioners. This work advances the field of decision tree-based modeling, offering a more flexible and accurate approach for handling structured data and complex tasks. By leveraging gradient-based optimization, the proposed method bridges the gap between traditional decision trees and modern machine learning techniques, paving the way for further innovations in interpretable and high-performing models.
☆ Graphical Transformation Models
Graphical Transformation Models (GTMs) are introduced as a novel approach to effectively model multivariate data with intricate marginals and complex dependency structures non-parametrically, while maintaining interpretability through the identification of varying conditional independencies. GTMs extend multivariate transformation models by replacing the Gaussian copula with a custom-designed multivariate transformation, offering two major advantages. Firstly, GTMs can capture more complex interdependencies using penalized splines, which also provide an efficient regularization scheme. Secondly, we demonstrate how to approximately regularize GTMs using a lasso penalty towards pairwise conditional independencies, akin to Gaussian graphical models. The model's robustness and effectiveness are validated through simulations, showcasing its ability to accurately learn parametric vine copulas and identify conditional independencies. Additionally, the model is applied to a benchmark astrophysics dataset, where the GTM demonstrates favorable performance compared to non-parametric vine copulas in learning complex multivariate distributions.
comment: 36 pages, 10 Figures, presented at the DAGStat 2025 in Berlin
☆ Adapt, Agree, Aggregate: Semi-Supervised Ensemble Labeling for Graph Convolutional Networks
In this paper, we propose a novel framework that combines ensemble learning with augmented graph structures to improve the performance and robustness of semi-supervised node classification in graphs. By creating multiple augmented views of the same graph, our approach harnesses the "wisdom of a diverse crowd", mitigating the challenges posed by noisy graph structures. Leveraging ensemble learning allows us to simultaneously achieve three key goals: adaptive confidence threshold selection based on model agreement, dynamic determination of the number of high-confidence samples for training, and robust extraction of pseudo-labels to mitigate confirmation bias. Our approach uniquely integrates adaptive ensemble consensus to flexibly guide pseudo-label extraction and sample selection, reducing the risks of error accumulation and improving robustness. Furthermore, the use of ensemble-driven consensus for pseudo-labeling captures subtle patterns that individual models often overlook, enabling the model to generalize better. Experiments on several real-world datasets demonstrate the effectiveness of our proposed method.
☆ On the Minimax Regret of Sequential Probability Assignment via Square-Root Entropy
We study the problem of sequential probability assignment under logarithmic loss, both with and without side information. Our objective is to analyze the minimax regret -- a notion extensively studied in the literature -- in terms of geometric quantities, such as covering numbers and scale-sensitive dimensions. We show that the minimax regret for the case of no side information (equivalently, the Shtarkov sum) can be upper bounded in terms of sequential square-root entropy, a notion closely related to Hellinger distance. For the problem of sequential probability assignment with side information, we develop both upper and lower bounds based on the aforementioned entropy. The lower bound matches the upper bound, up to log factors, for classes in the Donsker regime (according to our definition of entropy).
☆ Poisson-Process Topic Model for Integrating Knowledge from Pre-trained Language Models
Topic modeling is traditionally applied to word counts without accounting for the context in which words appear. Recent advancements in large language models (LLMs) offer contextualized word embeddings, which capture deeper meaning and relationships between words. We aim to leverage such embeddings to improve topic modeling. We use a pre-trained LLM to convert each document into a sequence of word embeddings. This sequence is then modeled as a Poisson point process, with its intensity measure expressed as a convex combination of $K$ base measures, each corresponding to a topic. To estimate these topics, we propose a flexible algorithm that integrates traditional topic modeling methods, enhanced by net-rounding applied before and kernel smoothing applied after. One advantage of this framework is that it treats the LLM as a black box, requiring no fine-tuning of its parameters. Another advantage is its ability to seamlessly integrate any traditional topic modeling approach as a plug-in module, without the need for modifications Assuming each topic is a $\beta$-H\"{o}lder smooth intensity measure on the embedded space, we establish the rate of convergence of our method. We also provide a minimax lower bound and show that the rate of our method matches with the lower bound when $\beta\leq 1$. Additionally, we apply our method to several datasets, providing evidence that it offers an advantage over traditional topic modeling approaches.
comment: 35 pages, 9 figures, 3 tables
☆ Neural Network Approach to Stochastic Dynamics for Smooth Multimodal Density Estimation
In this paper we consider a new probability sampling methods based on Langevin diffusion dynamics to resolve the problem of existing Monte Carlo algorithms when draw samples from high dimensional target densities. We extent Metropolis-Adjusted Langevin Diffusion algorithm by modelling the stochasticity of precondition matrix as a random matrix. An advantage compared to other proposal method is that it only requires the gradient of log-posterior. The proposed method provides fully adaptation mechanisms to tune proposal densities to exploits and adapts the geometry of local structures of statistical models. We clarify the benefits of the new proposal by modelling a Quantum Probability Density Functions of a free particle in a plane (energy Eigen-functions). The proposed model represents a remarkable improvement in terms of performance accuracy and computational time over standard MCMC method.
☆ A Roadmap Towards Improving Multi-Agent Reinforcement Learning With Causal Discovery And Inference
Causal reasoning is increasingly used in Reinforcement Learning (RL) to improve the learning process in several dimensions: efficacy of learned policies, efficiency of convergence, generalisation capabilities, safety and interpretability of behaviour. However, applications of causal reasoning to Multi-Agent RL (MARL) are still mostly unexplored. In this paper, we take the first step in investigating the opportunities and challenges of applying causal reasoning in MARL. We measure the impact of a simple form of causal augmentation in state-of-the-art MARL scenarios increasingly requiring cooperation, and with state-of-the-art MARL algorithms exploiting various degrees of collaboration between agents. Then, we discuss the positive as well as negative results achieved, giving us the chance to outline the areas where further research may help to successfully transfer causal RL to the multi-agent setting.
☆ Enhancing Fourier Neural Operators with Local Spatial Features
Partial Differential Equation (PDE) problems often exhibit strong local spatial structures, and effectively capturing these structures is critical for approximating their solutions. Recently, the Fourier Neural Operator (FNO) has emerged as an efficient approach for solving these PDE problems. By using parametrization in the frequency domain, FNOs can efficiently capture global patterns. However, this approach inherently overlooks the critical role of local spatial features, as frequency-domain parameterized convolutions primarily emphasize global interactions without encoding comprehensive localized spatial dependencies. Although several studies have attempted to address this limitation, their extracted Local Spatial Features (LSFs) remain insufficient, and computational efficiency is often compromised. To address this limitation, we introduce a convolutional neural network (CNN) preprocessor to extract LSFs directly from input data, resulting in a hybrid architecture termed \textit{Conv-FNO}. Furthermore, we introduce two novel resizing schemes to make our Conv-FNO resolution invariant. In this work, we focus on demonstrating the effectiveness of incorporating LSFs into FNOs by conducting both a theoretical analysis and extensive numerical experiments. Our findings show that this simple yet impactful modification enhances the representational capacity of FNOs and significantly improves performance on challenging PDE benchmarks.
☆ Every Sample Matters: Leveraging Mixture-of-Experts and High-Quality Data for Efficient and Accurate Code LLM
Recent advancements in code large language models (LLMs) have demonstrated remarkable capabilities in code generation and understanding. It is still challenging to build a code LLM with comprehensive performance yet ultimate efficiency. Many attempts have been released in the open source community to break the trade-off between performance and efficiency, such as the Qwen Coder series and the DeepSeek Coder series. This paper introduces yet another attempt in this area, namely Ling-Coder-Lite. We leverage the efficient Mixture-of-Experts (MoE) architecture along with a set of high-quality data curation methods (especially those based on program analytics) to build an efficient yet powerful code LLM. Ling-Coder-Lite exhibits on-par performance on 12 representative coding benchmarks compared to state-of-the-art models of similar size, such as Qwen2.5-Coder-7B and DeepSeek-Coder-V2-Lite, while offering competitive latency and throughput. In practice, we achieve a 50\% reduction in deployment resources compared to the similar-sized dense model without performance loss. To facilitate further research and development in this area, we open-source our models as well as a substantial portion of high-quality data for the annealing and post-training stages. The models and data can be accessed at~\url{https://huggingface.co/inclusionAI/Ling-Coder-lite}.
comment: 20 pages, 6 figures
☆ Energy-Aware LLMs: A step towards sustainable AI for downstream applications CEC
Advanced Large Language Models (LLMs) have revolutionized various fields, including communication networks, sparking an innovation wave that has led to new applications and services, and significantly enhanced solution schemes. Despite all these impressive developments, most LLMs typically require huge computational resources, resulting in terribly high energy consumption. Thus, this research study proposes an end-to-end pipeline that investigates the trade-off between energy efficiency and model performance for an LLM during fault ticket analysis in communication networks. It further evaluates the pipeline performance using two real-world datasets for the tasks of root cause analysis and response feedback in a communication network. Our results show that an appropriate combination of quantization and pruning techniques is able to reduce energy consumption while significantly improving model performance.
comment: This work has been submitted to V. International Conference on Electrical, Computer and Energy Technologies (ICECET 2025) for possible publication
☆ Renewable Energy Transition in South America: Predictive Analysis of Generation Capacity by 2050
In this research, renewable energy expansion in South America up to 2050 is predicted based on machine learning models that are trained on past energy data. The research employs gradient boosting regression and Prophet time series forecasting to make predictions of future generation capacities for solar, wind, hydroelectric, geothermal, biomass, and other renewable sources in South American nations. Model output analysis indicates staggering future expansion in the generation of renewable energy, with solar and wind energy registering the highest expansion rates. Geospatial visualization methods were applied to illustrate regional disparities in the utilization of renewable energy. The results forecast South America to record nearly 3-fold growth in the generation of renewable energy by the year 2050, with Brazil and Chile spearheading regional development. Such projections help design energy policy, investment strategy, and climate change mitigation throughout the region, in helping the developing economies to transition to sustainable energy.
comment: 13 pages, 5 figures
☆ Bandwidth Reservation for Time-Critical Vehicular Applications: A Multi-Operator Environment
Onsite bandwidth reservation requests often face challenges such as price fluctuations and fairness issues due to unpredictable bandwidth availability and stringent latency requirements. Requesting bandwidth in advance can mitigate the impact of these fluctuations and ensure timely access to critical resources. In a multi-Mobile Network Operator (MNO) environment, vehicles need to select cost-effective and reliable resources for their safety-critical applications. This research aims to minimize resource costs by finding the best price among multiple MNOs. It formulates multi-operator scenarios as a Markov Decision Process (MDP), utilizing a Deep Reinforcement Learning (DRL) algorithm, specifically Dueling Deep Q-Learning. For efficient and stable learning, we propose a novel area-wise approach and an adaptive MDP synthetic close to the real environment. The Temporal Fusion Transformer (TFT) is used to handle time-dependent data and model training. Furthermore, the research leverages Amazon spot price data and adopts a multi-phase training approach, involving initial training on synthetic data, followed by real-world data. These phases enable the DRL agent to make informed decisions using insights from historical data and real-time observations. The results show that our model leads to significant cost reductions, up to 40%, compared to scenarios without a policy model in such a complex environment.
comment: 14 pages, 11 figures
☆ Improving Preference Extraction In LLMs By Identifying Latent Knowledge Through Classifying Probes ACL
Large Language Models (LLMs) are often used as automated judges to evaluate text, but their effectiveness can be hindered by various unintentional biases. We propose using linear classifying probes, trained by leveraging differences between contrasting pairs of prompts, to directly access LLMs' latent knowledge and extract more accurate preferences. Through extensive experiments using models of varying size from four different families and six diverse datasets assessing text quality evaluation and common sense reasoning, we demonstrate that both supervised and unsupervised probing approaches consistently outperform traditional generation-based judgement while maintaining similar computational costs. These probes generalise under domain shifts and can even outperform finetuned evaluators with the same training data size. Our results suggest linear probing offers an accurate, robust and computationally efficient approach for LLM-as-judge tasks while providing interpretable insights into how models encode judgement-relevant knowledge. Our data and code will be openly released in the future.
comment: preprint, submitted to ACL ARR 2025, 21 pages, 23 figures
☆ Normalized Matching Transformer
We present a new state of the art approach for sparse keypoint matching between pairs of images. Our method consists of a fully deep learning based approach combining a visual backbone coupled with a SplineCNN graph neural network for feature processing and a normalized transformer decoder for decoding keypoint correspondences together with the Sinkhorn algorithm. Our method is trained using a contrastive and a hyperspherical loss for better feature representations. We additionally use data augmentation during training. This comparatively simple architecture combining extensive normalization and advanced losses outperforms current state of the art approaches on PascalVOC and SPair-71k datasets by $5.1\%$ and $2.2\%$ respectively compared to BBGM, ASAR, COMMON and GMTR while training for at least $1.7x$ fewer epochs.
☆ Decentralized Federated Dataset Dictionary Learning for Multi-Source Domain Adaptation ICASSP 2025
Decentralized Multi-Source Domain Adaptation (DMSDA) is a challenging task that aims to transfer knowledge from multiple related and heterogeneous source domains to an unlabeled target domain within a decentralized framework. Our work tackles DMSDA through a fully decentralized federated approach. In particular, we extend the Federated Dataset Dictionary Learning (FedDaDiL) framework by eliminating the necessity for a central server. FedDaDiL leverages Wasserstein barycenters to model the distributional shift across multiple clients, enabling effective adaptation while preserving data privacy. By decentralizing this framework, we enhance its robustness, scalability, and privacy, removing the risk of a single point of failure. We compare our method to its federated counterpart and other benchmark algorithms, showing that our approach effectively adapts source domains to an unlabeled target domain in a fully decentralized manner.
comment: Accepted at ICASSP 2025
☆ Safe RLHF-V: Safe Reinforcement Learning from Human Feedback in Multimodal Large Language Models
Multimodal large language models (MLLMs) are critical for developing general-purpose AI assistants, yet they face growing safety risks. How can we ensure that MLLMs are safely aligned to prevent undesired behaviors such as discrimination, misinformation, or violations of ethical standards? In a further step, we need to explore how to fine-tune MLLMs to enhance reasoning performance while ensuring they satisfy safety constraints. Fundamentally, this can be formulated as a min-max optimization problem. In this study, we propose Safe RLHF-V, the first multimodal safety alignment framework that jointly optimizes helpfulness and safety using separate multimodal reward and cost models within a Lagrangian-based constrained optimization framework. Given that there is a lack of preference datasets that separate helpfulness and safety in multimodal scenarios, we introduce BeaverTails-V, the first open-source dataset with dual preference annotations for helpfulness and safety, along with multi-level safety labels (minor, moderate, severe). Additionally, we design a Multi-level Guardrail System to proactively defend against unsafe queries and adversarial attacks. By applying the Beaver-Guard-V moderation for 5 rounds of filtering and re-generation on the precursor model, the overall safety of the upstream model is significantly improved by an average of 40.9%. Experimental results demonstrate that fine-tuning different MLLMs with Safe RLHF can effectively enhance model helpfulness while ensuring improved safety. Specifically, Safe RLHF-V improves model safety by 34.2% and helpfulness by 34.3%. All of datasets, models, and code can be found at https://github.com/SafeRLHF-V to support the safety development of MLLMs and reduce potential societal risks.
☆ Staying Alive: Online Neural Network Maintenance and Systemic Drift
We present the Subset Extended Kalman Filter (SEKF) as a method to update previously trained model weights online rather than retraining or finetuning them when the system a model represents drifts away from the conditions under which it was trained. We identify the parameters to be updated using the gradient of the loss function and use the SEKF to update only these parameters. We compare finetuning and SEKF for online model maintenance in the presence of systemic drift through four dynamic regression case studies and find that the SEKF is able to maintain model accuracy as-well if not better than finetuning while requiring significantly less time per iteration, and less hyperparameter tuning.
☆ Reducing Class-wise Confusion for Incremental Learning with Disentangled Manifolds CVPR 2025
Class incremental learning (CIL) aims to enable models to continuously learn new classes without catastrophically forgetting old ones. A promising direction is to learn and use prototypes of classes during incremental updates. Despite simplicity and intuition, we find that such methods suffer from inadequate representation capability and unsatisfied feature overlap. These two factors cause class-wise confusion and limited performance. In this paper, we develop a Confusion-REduced AuTo-Encoder classifier (CREATE) for CIL. Specifically, our method employs a lightweight auto-encoder module to learn compact manifold for each class in the latent subspace, constraining samples to be well reconstructed only on the semantically correct auto-encoder. Thus, the representation stability and capability of class distributions are enhanced, alleviating the potential class-wise confusion problem. To further distinguish the overlapped features, we propose a confusion-aware latent space separation loss that ensures samples are closely distributed in their corresponding low-dimensional manifold while keeping away from the distributions of features from other classes. Our method demonstrates stronger representational capacity and discrimination ability by learning disentangled manifolds and reduces class confusion. Extensive experiments on multiple datasets and settings show that CREATE outperforms other state-of-the-art methods up to 5.41%.
comment: Accepted to CVPR 2025
☆ MultiScale Contextual Bandits for Long Term Objectives
The feedback that AI systems (e.g., recommender systems, chatbots) collect from user interactions is a crucial source of training data. While short-term feedback (e.g., clicks, engagement) is widely used for training, there is ample evidence that optimizing short-term feedback does not necessarily achieve the desired long-term objectives. Unfortunately, directly optimizing for long-term objectives is challenging, and we identify the disconnect in the timescales of short-term interventions (e.g., rankings) and the long-term feedback (e.g., user retention) as one of the key obstacles. To overcome this disconnect, we introduce the framework of MultiScale Policy Learning to contextually reconcile that AI systems need to act and optimize feedback at multiple interdependent timescales. For any two levels, our formulation selects the shorter-term objective at the next lower scale to optimize the longer-term objective at the next higher scale. As a result, the policies at all levels effectively optimize for the long-term. We instantiate the framework with MultiScale Off-Policy Bandit Learning (MSBL) and demonstrate its effectiveness on three tasks relating to recommender systems and text generation.
☆ Multi-Modality Representation Learning for Antibody-Antigen Interactions Prediction ICME 2025
While deep learning models play a crucial role in predicting antibody-antigen interactions (AAI), the scarcity of publicly available sequence-structure pairings constrains their generalization. Current AAI methods often focus on residue-level static details, overlooking fine-grained structural representations of antibodies and their inter-antibody similarities. To tackle this challenge, we introduce a multi-modality representation approach that integates 3D structural and 1D sequence data to unravel intricate intra-antibody hierarchical relationships. By harnessing these representations, we present MuLAAIP, an AAI prediction framework that utilizes graph attention networks to illuminate graph-level structural features and normalized adaptive graph convolution networks to capture inter-antibody sequence associations. Furthermore, we have curated an AAI benchmark dataset comprising both structural and sequence information along with interaction labels. Through extensive experiments on this benchmark, our results demonstrate that MuLAAIP outperforms current state-of-the-art methods in terms of predictive performance. The implementation code and dataset are publicly available at https://github.com/trashTian/MuLAAIP for reproducibility.
comment: 2025 IEEE International Conference on Multimedia and Expo (ICME 2025), June 30 - July 4, 2025, Nantes, France
☆ CardioTabNet: A Novel Hybrid Transformer Model for Heart Disease Prediction using Tabular Medical Data
The early detection and prediction of cardiovascular diseases are crucial for reducing the severe morbidity and mortality associated with these conditions worldwide. A multi-headed self-attention mechanism, widely used in natural language processing (NLP), is operated by Transformers to understand feature interactions in feature spaces. However, the relationships between various features within biological systems remain ambiguous in these spaces, highlighting the necessity of early detection and prediction of cardiovascular diseases to reduce the severe morbidity and mortality with these conditions worldwide. We handle this issue with CardioTabNet, which exploits the strength of tab transformer to extract feature space which carries strong understanding of clinical cardiovascular data and its feature ranking. As a result, performance of downstream classical models significantly showed outstanding result. Our study utilizes the open-source dataset for heart disease prediction with 1190 instances and 11 features. In total, 11 features are divided into numerical (age, resting blood pressure, cholesterol, maximum heart rate, old peak, weight, and fasting blood sugar) and categorical (resting ECG, exercise angina, and ST slope). Tab transformer was used to extract important features and ranked them using random forest (RF) feature ranking algorithm. Ten machine-learning models were used to predict heart disease using selected features. After extracting high-quality features, the top downstream model (a hyper-tuned ExtraTree classifier) achieved an average accuracy rate of 94.1% and an average Area Under Curve (AUC) of 95.0%. Furthermore, a nomogram analysis was conducted to evaluate the model's effectiveness in cardiovascular risk assessment. A benchmarking study was conducted using state-of-the-art models to evaluate our transformer-driven framework.
comment: This paper is currently under review in the Health Information Science and Systems journal
♻ ☆ Learning Optimal Filters Using Variational Inference
Filtering - the task of estimating the conditional distribution for states of a dynamical system given partial and noisy observations - is important in many areas of science and engineering, including weather and climate prediction. However, the filtering distribution is generally intractable to obtain for high-dimensional, nonlinear systems. Filters used in practice, such as the ensemble Kalman filter (EnKF), provide biased probabilistic estimates for nonlinear systems and have numerous tuning parameters. Here, we present a framework for learning a parameterized analysis map - the transformation that takes samples from a forecast distribution, and combines with an observation, to update the approximate filtering distribution - using variational inference. In principle this can lead to a better approximation of the filtering distribution, and hence smaller bias. We show that this methodology can be used to learn the gain matrix, in an affine analysis map, for filtering linear and nonlinear dynamical systems; we also study the learning of inflation and localization parameters for an EnKF. The framework developed here can also be used to learn new filtering algorithms with more general forms for the analysis map.
comment: 21 pages, 6 figures
♻ ☆ Near-Polynomially Competitive Active Logistic Regression
We address the problem of active logistic regression in the realizable setting. It is well known that active learning can require exponentially fewer label queries compared to passive learning, in some cases using $\log \frac{1}{\eps}$ rather than $\poly(1/\eps)$ labels to get error $\eps$ larger than the optimum. We present the first algorithm that is polynomially competitive with the optimal algorithm on every input instance, up to factors polylogarithmic in the error and domain size. In particular, if any algorithm achieves label complexity polylogarithmic in $\eps$, so does ours. Our algorithm is based on efficient sampling and can be extended to learn more general class of functions. We further support our theoretical results with experiments demonstrating performance gains for logistic regression compared to existing active learning algorithms.
♻ ☆ Leveraging Multi-modal Representations to Predict Protein Melting Temperatures AAAI 2025
Accurately predicting protein melting temperature changes (Delta Tm) is fundamental for assessing protein stability and guiding protein engineering. Leveraging multi-modal protein representations has shown great promise in capturing the complex relationships among protein sequences, structures, and functions. In this study, we develop models based on powerful protein language models, including ESM-2, ESM-3 and AlphaFold, using various feature extraction methods to enhance prediction accuracy. By utilizing the ESM-3 model, we achieve a new state-of-the-art performance on the s571 test dataset, obtaining a Pearson correlation coefficient (PCC) of 0.50. Furthermore, we conduct a fair evaluation to compare the performance of different protein language models in the Delta Tm prediction task. Our results demonstrate that integrating multi-modal protein representations could advance the prediction of protein melting temperatures.
comment: Accepted to AAAI 2025 FM4BIO workshop
♻ ☆ 1000 Layer Networks for Self-Supervised RL: Scaling Depth Can Enable New Goal-Reaching Capabilities
Scaling up self-supervised learning has driven breakthroughs in language and vision, yet comparable progress has remained elusive in reinforcement learning (RL). In this paper, we study building blocks for self-supervised RL that unlock substantial improvements in scalability, with network depth serving as a critical factor. Whereas most RL papers in recent years have relied on shallow architectures (around 2 - 5 layers), we demonstrate that increasing the depth up to 1024 layers can significantly boost performance. Our experiments are conducted in an unsupervised goal-conditioned setting, where no demonstrations or rewards are provided, so an agent must explore (from scratch) and learn how to maximize the likelihood of reaching commanded goals. Evaluated on simulated locomotion and manipulation tasks, our approach increases performance by $2\times$ - $50\times$. Increasing the model depth not only increases success rates but also qualitatively changes the behaviors learned.
comment: Link to project website: https://wang-kevin3290.github.io/scaling-crl/
♻ ☆ A Privacy Model for Classical & Learned Bloom Filters
The Classical Bloom Filter (CBF) is a class of Probabilistic Data Structures (PDS) for handling Approximate Query Membership (AMQ). The Learned Bloom Filter (LBF) is a recently proposed class of PDS that combines the Classical Bloom Filter with a Learning Model while preserving the Bloom Filter's one-sided error guarantees. Bloom Filters have been used in settings where inputs are sensitive and need to be private in the presence of an adversary with access to the Bloom Filter through an API or in the presence of an adversary who has access to the internal state of the Bloom Filter. This paper conducts a rigorous differential privacy-based analysis for the Bloom Filter. We propose constructions that satisfy differential privacy and asymmetric differential privacy. This is also the first work that analyses and addresses the privacy of the Learned Bloom Filter under any rigorous model, which is an open problem.
♻ ☆ The Limits of Assumption-free Tests for Algorithm Performance
Algorithm evaluation and comparison are fundamental questions in machine learning and statistics -- how well does an algorithm perform at a given modeling task, and which algorithm performs best? Many methods have been developed to assess algorithm performance, often based around cross-validation type strategies, retraining the algorithm of interest on different subsets of the data and assessing its performance on the held-out data points. Despite the broad use of such procedures, the theoretical properties of these methods are not yet fully understood. In this work, we explore some fundamental limits for answering these questions with limited amounts of data. In particular, we make a distinction between two questions: how good is an algorithm $A$ at the problem of learning from a training set of size $n$, versus, how good is a particular fitted model produced by running $A$ on a particular training data set of size $n$? Our main results prove that, for any test that treats the algorithm $A$ as a ``black box'' (i.e., we can only study the behavior of $A$ empirically), there is a fundamental limit on our ability to carry out inference on the performance of $A$, unless the number of available data points $N$ is many times larger than the sample size $n$ of interest. (On the other hand, evaluating the performance of a particular fitted model is easy as long as a holdout data set is available -- that is, as long as $N-n$ is not too small.) We also ask whether an assumption of algorithmic stability might be sufficient to circumvent this hardness result. Surprisingly, we find that this is not the case: the same hardness result still holds for the problem of evaluating the performance of $A$, aside from a high-stability regime where fitted models are essentially nonrandom. Finally, we also establish similar hardness results for the problem of comparing multiple algorithms.
♻ ☆ A Physics-informed Machine Learning-based Control Method for Nonlinear Dynamic Systems with Highly Noisy Measurements
This study presents a physics-informed machine learning-based control method for nonlinear dynamic systems with highly noisy measurements. Existing data-driven control methods that use machine learning for system identification cannot effectively cope with highly noisy measurements, resulting in unstable control performance. To address this challenge, the present study extends current physics-informed machine learning capabilities for modeling nonlinear dynamics with control and integrates them into a model predictive control framework. To demonstrate the capability of the proposed method we test and validate with two noisy nonlinear dynamic systems: the chaotic Lorenz 3 system, and turning machine tool. Analysis of the results illustrate that the proposed method outperforms state-of-the-art benchmarks as measured by both modeling accuracy and control performance for nonlinear dynamic systems under high-noise conditions.
comment: We completely redesigned and rewrote this paper. It will be a completely different paper with different title, author list, and content
♻ ☆ Removing Structured Noise with Diffusion Models
Solving ill-posed inverse problems requires careful formulation of prior beliefs over the signals of interest and an accurate description of their manifestation into noisy measurements. Handcrafted signal priors based on e.g. sparsity are increasingly replaced by data-driven deep generative models, and several groups have recently shown that state-of-the-art score-based diffusion models yield particularly strong performance and flexibility. In this paper, we show that the powerful paradigm of posterior sampling with diffusion models can be extended to include rich, structured, noise models. To that end, we propose a joint conditional reverse diffusion process with learned scores for the noise and signal-generating distribution. We demonstrate strong performance gains across various inverse problems with structured noise, outperforming competitive baselines that use normalizing flows and adversarial networks. This opens up new opportunities and relevant practical applications of diffusion modeling for inverse problems in the context of non-Gaussian measurement models.
comment: 20 pages, 8 figures, Transactions on Machine Learning Research
♻ ☆ Consistent Validation for Predictive Methods in Spatial Settings AISTATS
Spatial prediction tasks are key to weather forecasting, studying air pollution impacts, and other scientific endeavors. Determining how much to trust predictions made by statistical or physical methods is essential for the credibility of scientific conclusions. Unfortunately, classical approaches for validation fail to handle mismatch between locations available for validation and (test) locations where we want to make predictions. This mismatch is often not an instance of covariate shift (as commonly formalized) because the validation and test locations are fixed (e.g., on a grid or at select points) rather than i.i.d. from two distributions. In the present work, we formalize a check on validation methods: that they become arbitrarily accurate as validation data becomes arbitrarily dense. We show that classical and covariate-shift methods can fail this check. We propose a method that builds from existing ideas in the covariate-shift literature, but adapts them to the validation data at hand. We prove that our proposal passes our check. And we demonstrate its advantages empirically on simulated and real data.
comment: 52 pages, 14 figures, Accepted at Artificial Intelligence and Statistics (AISTATS), 2025
♻ ☆ LLM+KG@VLDB'24 Workshop Summary SIGMOD
The unification of large language models (LLMs) and knowledge graphs (KGs) has emerged as a hot topic. At the LLM+KG'24 workshop, held in conjunction with VLDB 2024 in Guangzhou, China, one of the key themes explored was important data management challenges and opportunities due to the effective interaction between LLMs and KGs. This report outlines the major directions and approaches presented by various speakers during the LLM+KG'24 workshop.
comment: accepted at ACM SIGMOD Record 2025
♻ ☆ Deep learning framework for action prediction reveals multi-timescale locomotor control
Modeling movement in real-world tasks is a fundamental scientific goal for motor control, biomechanics, and rehabilitation engineering. However, existing models and their simplifying assumptions such as linear and fixed timescale mappings do not generalize to real-world contexts. Here, we develop a deep learning-based framework for action prediction with architecture-dependent trial embedding, outperforming traditional models across multiple contexts (walking and running, treadmill and overground, varying terrains) and input modalities (multiple body states, gaze). We find that neural network architectures with flexible input history-dependence like GRU and Transformer perform best overall. By quantifying the model's predictions relative to an autoregressive baseline, we identify context- and modality-dependent timescales. There is greater reliance on fast-timescale predictions in complex terrain, gaze predictions precede body state predictions, and full-body state predictions precede center-of-mass-relevant predictions. This deep learning framework for action prediction provides quantifiable insights into the control of complex movements and can be extended to other actions, contexts, and populations.
♻ ☆ Active management of battery degradation in wireless sensor network using deep reinforcement learning for group battery replacement
Wireless sensor networks (WSNs) have become a promising solution for structural health monitoring (SHM), especially in hard-to-reach or remote locations. Battery-powered WSNs offer various advantages over wired systems, however limited battery life has always been one of the biggest obstacles in practical use of the WSNs, regardless of energy harvesting methods. While various methods have been studied for battery health management, existing methods exclusively aim to extend lifetime of individual batteries, lacking a system level view. A consequence of applying such methods is that batteries in a WSN tend to fail at different times, posing significant difficulty on planning and scheduling of battery replacement trip. This study investigate a deep reinforcement learning (DRL) method for active battery degradation management by optimizing duty cycle of WSNs at the system level. This active management strategy effectively reduces earlier failure of battery individuals which enable group replacement without sacrificing WSN performances. A simulated environment based on a real-world WSN setup was developed to train a DRL agent and learn optimal duty cycle strategies. The performance of the strategy was validated in a long-term setup with various network sizes, demonstrating its efficiency and scalability.
♻ ☆ ULTra: Unveiling Latent Token Interpretability in Transformer-Based Understanding and Segmentation
Transformers have revolutionized Computer Vision (CV) through self-attention mechanisms. However, their complexity makes latent token representations difficult to interpret. We introduce ULTra, a framework for interpreting Transformer embeddings and uncovering meaningful semantic patterns within them. ULTra enables unsupervised semantic segmentation using pre-trained models without requiring fine-tuning. Additionally, we propose a self-supervised training approach that refines segmentation performance by learning an external transformation matrix without modifying the underlying model. Our method achieves state-of-the-art performance in unsupervised semantic segmentation, outperforming existing segmentation methods. Furthermore, we validate ULTra for model interpretation on both synthetic and real-world scenarios, including Object Selection and interpretable text summarization using LLMs, demonstrating its broad applicability in explaining the semantic structure of latent token representations.
♻ ☆ On the Global Convergence of Risk-Averse Policy Gradient Methods with Expected Conditional Risk Measures
Risk-sensitive reinforcement learning (RL) has become a popular tool for controlling the risk of uncertain outcomes and ensuring reliable performance in highly stochastic sequential decision-making problems. While Policy Gradient (PG) methods have been developed for risk-sensitive RL, it remains unclear if these methods enjoy the same global convergence guarantees as in the risk-neutral case \citep{mei2020global,agarwal2021theory,cen2022fast,bhandari2024global}. In this paper, we consider a class of dynamic time-consistent risk measures, named Expected Conditional Risk Measures (ECRMs), and derive PG and Natural Policy Gradient (NPG) updates for ECRMs-based RL problems. We provide global optimality {and iteration complexities} of the proposed algorithms under the following four settings: (i) PG with constrained direct parameterization, (ii) PG with softmax parameterization and log barrier regularization, (iii) NPG with softmax parameterization and entropy regularization, and (iv) approximate NPG with inexact policy evaluation. Furthermore, we test a risk-averse REINFORCE algorithm \citep{williams1992simple} and a risk-averse NPG algorithm \citep{kakade2001natural} on a stochastic Cliffwalk environment to demonstrate the efficacy of our methods and the importance of risk control.
♻ ☆ Linear Partial Gromov-Wasserstein Embedding ICLR 2025
The Gromov-Wasserstein (GW) problem, a variant of the classical optimal transport (OT) problem, has attracted growing interest in the machine learning and data science communities due to its ability to quantify similarity between measures in different metric spaces. However, like the classical OT problem, GW imposes an equal mass constraint between measures, which restricts its application in many machine learning tasks. To address this limitation, the partial Gromov-Wasserstein (PGW) problem has been introduced. It relaxes the equal mass constraint, allowing the comparison of general positive Radon measures. Despite this, both GW and PGW face significant computational challenges due to their non-convex nature. To overcome these challenges, we propose the linear partial Gromov-Wasserstein (LPGW) embedding, a linearized embedding technique for the PGW problem. For $K$ different metric measure spaces, the pairwise computation of the PGW distance requires solving the PGW problem ${O}(K^2)$ times. In contrast, the proposed linearization technique reduces this to ${O}(K)$ times. Similar to the linearization technique for the classical OT problem, we prove that LPGW defines a valid metric for metric measure spaces. Finally, we demonstrate the effectiveness of LPGW in practical applications such as shape retrieval and learning with transport-based embeddings, showing that LPGW preserves the advantages of PGW in partial matching while significantly enhancing computational efficiency. The code is available at https://github.com/mint-vu/Linearized_Partial_Gromov_Wasserstein.
comment: Published at ICLR 2025
♻ ☆ Simpler Diffusion (SiD2): 1.5 FID on ImageNet512 with pixel-space diffusion CVPR 2025
Latent diffusion models have become the popular choice for scaling up diffusion models for high resolution image synthesis. Compared to pixel-space models that are trained end-to-end, latent models are perceived to be more efficient and to produce higher image quality at high resolution. Here we challenge these notions, and show that pixel-space models can be very competitive to latent models both in quality and efficiency, achieving 1.5 FID on ImageNet512 and new SOTA results on ImageNet128, ImageNet256 and Kinetics600. We present a simple recipe for scaling end-to-end pixel-space diffusion models to high resolutions. 1: Use the sigmoid loss-weighting (Kingma & Gao, 2023) with our prescribed hyper-parameters. 2: Use our simplified memory-efficient architecture with fewer skip-connections. 3: Scale the model to favor processing the image at a high resolution with fewer parameters, rather than using more parameters at a lower resolution. Combining these with guidance intervals, we obtain a family of pixel-space diffusion models we call Simpler Diffusion (SiD2).
comment: Accepted to CVPR 2025
♻ ☆ A General Approach for Determining Applicability Domain of Machine Learning Models
Knowledge of the domain of applicability of a machine learning model is essential to ensuring accurate and reliable model predictions. In this work, we develop a new and general approach of assessing model domain and demonstrate that our approach provides accurate and meaningful domain designation across multiple model types and material property data sets. Our approach assesses the distance between data in feature space using kernel density estimation, where this distance provides an effective tool for domain determination. We show that chemical groups considered unrelated based on chemical knowledge exhibit significant dissimilarities by our measure. We also show that high measures of dissimilarity are associated with poor model performance (i.e., high residual magnitudes) and poor estimates of model uncertainty (i.e., unreliable uncertainty estimation). Automated tools are provided to enable researchers to establish acceptable dissimilarity thresholds to identify whether new predictions of their own machine learning models are in-domain versus out-of-domain.
♻ ☆ SynMorph: Generating Synthetic Face Morphing Dataset with Mated Samples
Face morphing attack detection (MAD) algorithms have become essential to overcome the vulnerability of face recognition systems. To solve the lack of large-scale and public-available datasets due to privacy concerns and restrictions, in this work we propose a new method to generate a synthetic face morphing dataset with 2450 identities and more than 100k morphs. The proposed synthetic face morphing dataset is unique for its high-quality samples, different types of morphing algorithms, and the generalization for both single and differential morphing attack detection algorithms. For experiments, we apply face image quality assessment and vulnerability analysis to evaluate the proposed synthetic face morphing dataset from the perspective of biometric sample quality and morphing attack potential on face recognition systems. The results are benchmarked with an existing SOTA synthetic dataset and a representative non-synthetic and indicate improvement compared with the SOTA. Additionally, we design different protocols and study the applicability of using the proposed synthetic dataset on training morphing attack detection algorithms.
comment: This preprint has been further published in IEEE Access. Print ISSN: 2169-3536. Online ISSN: 2169-3536. Digital Object Identifier: 10.1109/ACCESS.2025.3548957
♻ ☆ Not All LLM-Generated Data Are Equal: Rethinking Data Weighting in Text Classification ICLR 2025
Synthetic data augmentation via large language models (LLMs) allows researchers to leverage additional training data, thus enhancing the performance of downstream tasks, especially when real-world data is scarce. However, the generated data can deviate from the real-world data, and this misalignment can bring deficient outcomes while applying the trained model to applications. Therefore, we proposed efficient weighted-loss approaches to align synthetic data with real-world distribution by emphasizing high-quality and diversified data generated by LLMs with using merely a little real-world data. We empirically assessed the effectiveness of our method on multiple text classification tasks, and the results showed leveraging our approaches on a BERT-level model robustly outperformed standard cross-entropy and other data weighting approaches, providing potential solutions to effectively leveraging synthetic data from any suitable data generator for model training.
comment: ICLR 2025 camera ready
♻ ☆ Joint Transmit and Pinching Beamforming for Pinching Antenna Systems (PASS): Optimization-Based or Learning-Based?
A novel pinching antenna system (PASS)-enabled downlink multi-user multiple-input single-output (MISO) framework is proposed. PASS consists of multiple waveguides spanning over thousands of wavelength, which equip numerous low-cost dielectric particles, named pinching antennas (PAs), to radiate signals into free space. The positions of PAs can be reconfigured to change both the large-scale path losses and phases of signals, thus facilitating the novel pinching beamforming design. A sum rate maximization problem is formulated, which jointly optimizes the transmit and pinching beamforming to adaptively achieve constructive signal enhancement and destructive interference mitigation. To solve this highly coupled and nonconvex problem, both optimization-based and learning-based methods are proposed. 1) For the optimization-based method, a majorization-minimization and penalty dual decomposition (MM-PDD) algorithm is developed, which handles the nonconvex complex exponential component using a Lipschitz surrogate function and then invokes PDD for problem decoupling. 2) For the learning-based method, a novel Karush-Kuhn-Tucker (KKT)-guided dual learning (KDL) approach is proposed, which enables KKT solutions to be reconstructed in a data-driven manner by learning dual variables. Following this idea, a KDL-Tranformer algorithm is developed, which captures both inter-PA/inter-user dependencies and channel-state-information (CSI)-beamforming dependencies by attention mechanisms. Simulation results demonstrate that: i) The proposed PASS framework significantly outperforms conventional massive multiple input multiple output (MIMO) system even with a few PAs. ii) The proposed KDL-Transformer can improve over 30% system performance than MM-PDD algorithm, while achieving a millisecond-level response on modern GPUs.
comment: Submitted to IEEE
♻ ☆ Causality-oriented robustness: exploiting general noise interventions
Since distribution shifts are common in real-world applications, there is a pressing need to develop prediction models that are robust against such shifts. Existing frameworks, such as empirical risk minimization or distributionally robust optimization, either lack generalizability for unseen distributions or rely on postulated distance measures. Alternatively, causality offers a data-driven and structural perspective to robust predictions. However, the assumptions necessary for causal inference can be overly stringent, and the robustness offered by such causal models often lacks flexibility. In this paper, we focus on causality-oriented robustness and propose Distributional Robustness via Invariant Gradients (DRIG), a method that exploits general noise interventions in training data for robust predictions against unseen interventions, and naturally interpolates between in-distribution prediction and causality. In a linear setting, we prove that DRIG yields predictions that are robust among a data-dependent class of distribution shifts. Furthermore, we show that our framework includes anchor regression as a special case, and that it yields prediction models that protect against more diverse perturbations. We establish finite-sample results and extend our approach to semi-supervised domain adaptation to further improve prediction performance. Finally, we empirically validate our methods on synthetic simulations and on single-cell and intensive health care datasets.
♻ ☆ Towards Seamless Hierarchical Federated Learning under Intermittent Client Participation: A Stagewise Decision-Making Methodology
Federated Learning (FL) offers a pioneering distributed learning paradigm that enables devices/clients to build a shared global model. This global model is obtained through frequent model transmissions between clients and a central server, which may cause high latency, energy consumption, and congestion over backhaul links. To overcome these drawbacks, Hierarchical Federated Learning (HFL) has emerged, which organizes clients into multiple clusters and utilizes edge nodes (e.g., edge servers) for intermediate model aggregations between clients and the central server. Current research on HFL mainly focus on enhancing model accuracy, latency, and energy consumption in scenarios with a stable/fixed set of clients. However, addressing the dynamic availability of clients -- a critical aspect of real-world scenarios -- remains underexplored. This study delves into optimizing client selection and client-to-edge associations in HFL under intermittent client participation so as to minimize overall system costs (i.e., delay and energy), while achieving fast model convergence. We unveil that achieving this goal involves solving a complex NP-hard problem. To tackle this, we propose a stagewise methodology that splits the solution into two stages, referred to as Plan A and Plan B. Plan A focuses on identifying long-term clients with high chance of participation in subsequent model training rounds. Plan B serves as a backup, selecting alternative clients when long-term clients are unavailable during model training rounds. This stagewise methodology offers a fresh perspective on client selection that can enhance both HFL and conventional FL via enabling low-overhead decision-making processes. Through evaluations on MNIST and CIFAR-10 datasets, we show that our methodology outperforms existing benchmarks in terms of model accuracy and system costs.
comment: 20 pages, 8 figures,5 tables
♻ ☆ Oscillatory Signatures of Parkinson's Disease: Central and Parietal EEG Alterations Across Multiple Frequency Bands
This study investigates EEG as a potential early biomarker by applying deep learning techniques to resting-state EEG recordings from 31 subjects (15 with PD and 16 healthy controls). EEG signals underwent preprocessing to remove tremor artifacts before classification with CNNs using wavelet-based electrode triplet images. Our analysis across different brain regions and frequency bands showed distinct spatial-spectral patterns of PD-related neural oscillations. We identified high classification accuracy (76%) using central electrodes (C3, Cz, C4) with full-spectrum 0.4-62.4 Hz analysis and 74% accuracy in right parietal regions (P8, CP6, P4) with 10-second windows. Bilateral centro-parietal regions showed strong performance (67%) in the theta band (4.0-7.79 Hz), while multiple areas demonstrated some sensitivity (65%) in the alpha band (7.8-15.59 Hz). We also observed a distinctive topographical pattern of gamma band (40-62.4 Hz) alterations specifically localized to central-parietal regions, which remained consistent across different temporal windows. In particular, we observed pronounced right-hemisphere involvement across several frequency bands. Unlike previous studies that achieved higher accuracies by potentially including tremor artifacts, our approach isolates genuine neurophysiological alterations in cortical activity. These findings suggest that specific EEG-based oscillatory patterns, especially in central and parietal regions and across multiple frequency bands, may provide diagnostic information for PD, potentially before the onset of motor symptoms.
♻ ☆ Enhancing Startup Success Predictions in Venture Capital: A GraphRAG Augmented Multivariate Time Series Method ICLR 2025
In the Venture Capital (VC) industry, predicting the success of startups is challenging due to limited financial data and the need for subjective revenue forecasts. Previous methods based on time series analysis often fall short as they fail to incorporate crucial inter-company relationships such as competition and collaboration. To fill the gap, this paper aims to introduce a novel approach using GraphRAG augmented time series model. With GraphRAG, time series predictive methods are enhanced by integrating these vital relationships into the analysis framework, allowing for a more dynamic understanding of the startup ecosystem in venture capital. Our experimental results demonstrate that our model significantly outperforms previous models in startup success predictions.
comment: ICLR 2025 Financial AI
♻ ☆ Unsupervised Structural-Counterfactual Generation under Domain Shift
Motivated by the burgeoning interest in cross-domain learning, we present a novel generative modeling challenge: generating counterfactual samples in a target domain based on factual observations from a source domain. Our approach operates within an unsupervised paradigm devoid of parallel or joint datasets, relying exclusively on distinct observational samples and causal graphs for each domain. This setting presents challenges that surpass those of conventional counterfactual generation. Central to our methodology is the disambiguation of exogenous causes into effect-intrinsic and domain-intrinsic categories. This differentiation facilitates the integration of domain-specific causal graphs into a unified joint causal graph via shared effect-intrinsic exogenous variables. We propose leveraging Neural Causal models within this joint framework to enable accurate counterfactual generation under standard identifiability assumptions. Furthermore, we introduce a novel loss function that effectively segregates effect-intrinsic from domain-intrinsic variables during model training. Given a factual observation, our framework combines the posterior distribution of effect-intrinsic variables from the source domain with the prior distribution of domain-intrinsic variables from the target domain to synthesize the desired counterfactuals, adhering to Pearl's causal hierarchy. Intriguingly, when domain shifts are restricted to alterations in causal mechanisms without accompanying covariate shifts, our training regimen parallels the resolution of a conditional optimal transport problem. Empirical evaluations on a synthetic dataset show that our framework generates counterfactuals in the target domain that very closely resemble the ground truth.
comment: Updated author list
♻ ☆ Be More Diverse than the Most Diverse: Optimal Mixtures of Generative Models via Mixture-UCB Bandit Algorithms
The availability of multiple training algorithms and architectures for generative models requires a selection mechanism to form a single model over a group of well-trained generation models. The selection task is commonly addressed by identifying the model that maximizes an evaluation score based on the diversity and quality of the generated data. However, such a best-model identification approach overlooks the possibility that a mixture of available models can outperform each individual model. In this work, we numerically show that a mixture of generative models on benchmark image datasets can indeed achieve a better evaluation score (based on FID and KID scores), compared to the individual models. This observation motivates the development of efficient algorithms for selecting the optimal mixture of the models. To address this, we formulate a quadratic optimization problem to find an optimal mixture model achieving the maximum of kernel-based evaluation scores including kernel inception distance (KID) and R\'enyi kernel entropy (RKE). To identify the optimal mixture of the models using the fewest possible sample queries, we view the selection task as a multi-armed bandit (MAB) problem and propose the Mixture Upper Confidence Bound (Mixture-UCB) algorithm that provably converges to the optimal mixture of the involved models. More broadly, the proposed Mixture-UCB can be extended to optimize every convex quadratic function of the mixture weights in a general MAB setting. We prove a regret bound for the Mixture-UCB algorithm and perform several numerical experiments to show the success of Mixture-UCB in finding the optimal mixture of text and image generative models. The project code is available at https://github.com/Rezaei-Parham/Mixture-UCB.
♻ ☆ Efficiency is Not Enough: A Critical Perspective of Environmentally Sustainable AI
Artificial intelligence (AI) is currently spearheaded by machine learning (ML) methods such as deep learning which have accelerated progress on many tasks thought to be out of reach of AI. These recent ML methods are often compute hungry, energy intensive, and result in significant green house gas emissions, a known driver of anthropogenic climate change. Additionally, the platforms on which ML systems run are associated with environmental impacts that go beyond the energy consumption driven carbon emissions. The primary solution lionized by both industry and the ML community to improve the environmental sustainability of ML is to increase the compute and energy efficiency with which ML systems operate. In this perspective, we argue that it is time to look beyond efficiency in order to make ML more environmentally sustainable. We present three high-level discrepancies between the many variables that influence the efficiency of ML and the environmental sustainability of ML. Firstly, we discuss how compute efficiency does not imply energy efficiency or carbon efficiency. Second, we present the unexpected effects of efficiency on operational emissions throughout the ML model life cycle. And, finally, we explore the broader environmental impacts that are not accounted by efficiency. These discrepancies show as to why efficiency alone is not enough to remedy the adverse environmental impacts of ML. Instead, we argue for systems thinking as the next step towards holistically improving the environmental sustainability of ML.
comment: Accepted to Communications of the ACM; 24 pages; 6 figures
♻ ☆ LoRA-Pro: Are Low-Rank Adapters Properly Optimized? ICLR 2025
Low-rank adaptation, also known as LoRA, has emerged as a prominent method for parameter-efficient fine-tuning of foundation models. Despite its computational efficiency, LoRA still yields inferior performance compared to full fine-tuning. In this paper, we first uncover a fundamental connection between the optimization processes of LoRA and full fine-tuning: using LoRA for optimization is mathematically equivalent to full fine-tuning using a low-rank gradient for parameter updates. And this low-rank gradient can be expressed in terms of the gradients of the two low-rank matrices in LoRA. Leveraging this insight, we introduce LoRA-Pro, a method that enhances LoRA's performance by strategically adjusting the gradients of these low-rank matrices. This adjustment allows the low-rank gradient to more accurately approximate the full fine-tuning gradient, thereby narrowing the performance gap between LoRA and full fine-tuning. Furthermore, we theoretically derive the optimal solutions for adjusting the gradients of the low-rank matrices, applying them during fine-tuning in LoRA-Pro. We conduct extensive experiments across natural language understanding, dialogue generation, mathematical reasoning, code generation, and image classification tasks, demonstrating that LoRA-Pro substantially improves LoRA's performance, effectively narrowing the gap with full fine-tuning. Code is publicly available at https://github.com/mrflogs/LoRA-Pro.
comment: Camera-Ready Version for ICLR 2025; technical corrections to previous version
♻ ☆ BOPO: Neural Combinatorial Optimization via Best-anchored and Objective-guided Preference Optimization
Neural Combinatorial Optimization (NCO) has emerged as a promising approach for NP-hard problems. However, prevailing RL-based methods suffer from low sample efficiency due to sparse rewards and underused solutions. We propose Preference Optimization for Combinatorial Optimization (POCO), a training paradigm that leverages solution preferences via objective values. It introduces: (1) an efficient preference pair construction for better explore and exploit solutions, and (2) a novel loss function that adaptively scales gradients via objective differences, removing reliance on reward models or reference policies. Experiments on Job-Shop Scheduling (JSP), Traveling Salesman (TSP), and Flexible Job-Shop Scheduling (FJSP) show POCO outperforms state-of-the-art neural methods, reducing optimality gaps impressively with efficient inference. POCO is architecture-agnostic, enabling seamless integration with existing NCO models, and establishes preference optimization as a principled framework for combinatorial optimization.
♻ ☆ BiMarker: Enhancing Text Watermark Detection for Large Language Models with Bipolar Watermarks
The rapid growth of Large Language Models (LLMs) raises concerns about distinguishing AI-generated text from human content. Existing watermarking techniques, like \kgw, struggle with low watermark strength and stringent false-positive requirements. Our analysis reveals that current methods rely on coarse estimates of non-watermarked text, limiting watermark detectability. To address this, we propose Bipolar Watermark (\tool), which splits generated text into positive and negative poles, enhancing detection without requiring additional computational resources or knowledge of the prompt. Theoretical analysis and experimental results demonstrate \tool's effectiveness and compatibility with existing optimization techniques, providing a new optimization dimension for watermarking in LLM-generated content.
Information Retrieval 15
☆ Dense Passage Retrieval in Conversational Search
Information retrieval systems have traditionally relied on exact term match methods such as BM25 for first-stage retrieval. However, recent advancements in neural network-based techniques have introduced a new method called dense retrieval. This approach uses a dual-encoder to create contextual embeddings that can be indexed and clustered efficiently at run-time, resulting in improved retrieval performance in Open-domain Question Answering systems. In this paper, we apply the dense retrieval technique to conversational search by conducting experiments on the CAsT benchmark dataset. We also propose an end-to-end conversational search system called GPT2QR+DPR, which incorporates various query reformulation strategies to improve retrieval accuracy. Our findings indicate that dense retrieval outperforms BM25 even without extensive fine-tuning. Our work contributes to the growing body of research on neural-based retrieval methods in conversational search, and highlights the potential of dense retrieval in improving retrieval accuracy in conversational search systems.
☆ Towards Carbon Footprint-Aware Recommender Systems for Greener Item Recommendation
The commodity and widespread use of online shopping are having an unprecedented impact on climate, with emission figures from key actors that are easily comparable to those of a large-scale metropolis. Despite online shopping being fueled by recommender systems (RecSys) algorithms, the role and potential of the latter in promoting more sustainable choices is little studied. One of the main reasons for this could be attributed to the lack of a dataset containing carbon footprint emissions for the items. While building such a dataset is a rather challenging task, its presence is pivotal for opening the doors to novel perspectives, evaluations, and methods for RecSys research. In this paper, we target this bottleneck and study the environmental role of RecSys algorithms. First, we mine a dataset that includes carbon footprint emissions for its items. Then, we benchmark conventional RecSys algorithms in terms of accuracy and sustainability as two faces of the same coin. We find that RecSys algorithms optimized for accuracy overlook greenness and that longer recommendation lists are greener but less accurate. Then, we show that a simple reranking approach that accounts for the item's carbon footprint can establish a better trade-off between accuracy and greenness. This reranking approach is modular, ready to use, and can be applied to any RecSys algorithm without the need to alter the underlying mechanisms or retrain models. Our results show that a small sacrifice of accuracy can lead to significant improvements of recommendation greenness across all algorithms and list lengths. Arguably, this accuracy-greenness trade-off could even be seen as an enhancement of user satisfaction, particularly for purpose-driven users who prioritize the environmental impact of their choices. We anticipate this work will serve as the starting point for studying RecSys for more sustainable recommendations.
☆ The CASTLE 2024 Dataset: Advancing the Art of Multimodal Understanding
Egocentric video has seen increased interest in recent years, as it is used in a range of areas. However, most existing datasets are limited to a single perspective. In this paper, we present the CASTLE 2024 dataset, a multimodal collection containing ego- and exo-centric (i.e., first- and third-person perspective) video and audio from 15 time-aligned sources, as well as other sensor streams and auxiliary data. The dataset was recorded by volunteer participants over four days in a fixed location and includes the point of view of 10 participants, with an additional 5 fixed cameras providing an exocentric perspective. The entire dataset contains over 600 hours of UHD video recorded at 50 frames per second. In contrast to other datasets, CASTLE 2024 does not contain any partial censoring, such as blurred faces or distorted audio. The dataset is available via https://castle-dataset.github.io/.
comment: 7 pages, 6 figures, dataset available via https://castle-dataset.github.io/
☆ A Study into Investigating Temporal Robustness of LLMs
Large Language Models (LLMs) encapsulate a surprising amount of factual world knowledge. However, their performance on temporal questions and historical knowledge is limited because they often cannot understand temporal scope and orientation or neglect the temporal aspect altogether. In this study, we aim to measure precisely how robust LLMs are for question answering based on their ability to process temporal information and perform tasks requiring temporal reasoning and temporal factual knowledge. Specifically, we design eight time-sensitive robustness tests for factual information to check the sensitivity of six popular LLMs in the zero-shot setting. Overall, we find LLMs lacking temporal robustness, especially to temporal reformulations and the use of different granularities of temporal references. We show how a selection of these eight tests can be used automatically to judge a model's temporal robustness for user questions on the fly. Finally, we apply the findings of this study to improve the temporal QA performance by up to 55 percent.
comment: 8 pages
☆ Beyond Negation Detection: Comprehensive Assertion Detection Models for Clinical NLP ECIR 2025
Assertion status detection is a critical yet often overlooked component of clinical NLP, essential for accurately attributing extracted medical facts. Past studies have narrowly focused on negation detection, leading to underperforming commercial solutions such as AWS Medical Comprehend, Azure AI Text Analytics, and GPT-4o due to their limited domain adaptation. To address this gap, we developed state-of-the-art assertion detection models, including fine-tuned LLMs, transformer-based classifiers, few-shot classifiers, and deep learning (DL) approaches. We evaluated these models against cloud-based commercial API solutions, the legacy rule-based NegEx approach, and GPT-4o. Our fine-tuned LLM achieves the highest overall accuracy (0.962), outperforming GPT-4o (0.901) and commercial APIs by a notable margin, particularly excelling in Present (+4.2%), Absent (+8.4%), and Hypothetical (+23.4%) assertions. Our DL-based models surpass commercial solutions in Conditional (+5.3%) and Associated-with-Someone-Else (+10.1%) categories, while the few-shot classifier offers a lightweight yet highly competitive alternative (0.929), making it ideal for resource-constrained environments. Integrated within Spark NLP, our models consistently outperform black-box commercial solutions while enabling scalable inference and seamless integration with medical NER, Relation Extraction, and Terminology Resolution. These results reinforce the importance of domain-adapted, transparent, and customizable clinical NLP solutions over general-purpose LLMs and proprietary APIs.
comment: accepted at Text2Story Workshop at ECIR 2025
☆ Rankformer: A Graph Transformer for Recommendation based on Ranking Objective WWW2025
Recommender Systems (RS) aim to generate personalized ranked lists for each user and are evaluated using ranking metrics. Although personalized ranking is a fundamental aspect of RS, this critical property is often overlooked in the design of model architectures. To address this issue, we propose Rankformer, a ranking-inspired recommendation model. The architecture of Rankformer is inspired by the gradient of the ranking objective, embodying a unique (graph) transformer architecture -- it leverages global information from all users and items to produce more informative representations and employs specific attention weights to guide the evolution of embeddings towards improved ranking performance. We further develop an acceleration algorithm for Rankformer, reducing its complexity to a linear level with respect to the number of positive instances. Extensive experimental results demonstrate that Rankformer outperforms state-of-the-art methods. The code is available at https://github.com/StupidThree/Rankformer.
comment: Accepted by WWW2025
☆ Federated Cross-Domain Click-Through Rate Prediction With Large Language Model Augmentation
Accurately predicting click-through rates (CTR) under stringent privacy constraints poses profound challenges, particularly when user-item interactions are sparse and fragmented across domains. Conventional cross-domain CTR (CCTR) methods frequently assume homogeneous feature spaces and rely on centralized data sharing, neglecting complex inter-domain discrepancies and the subtle trade-offs imposed by privacy-preserving protocols. Here, we present Federated Cross-Domain CTR Prediction with Large Language Model Augmentation (FedCCTR-LM), a federated framework engineered to address these limitations by synchronizing data augmentation, representation disentanglement, and adaptive privacy protection. Our approach integrates three core innovations. First, the Privacy-Preserving Augmentation Network (PrivAugNet) employs large language models to enrich user and item representations and expand interaction sequences, mitigating data sparsity and feature incompleteness. Second, the Independent Domain-Specific Transformer with Contrastive Learning (IDST-CL) module disentangles domain-specific and shared user preferences, employing intra-domain representation alignment (IDRA) and crossdomain representation disentanglement (CDRD) to refine the learned embeddings and enhance knowledge transfer across domains. Finally, the Adaptive Local Differential Privacy (AdaLDP) mechanism dynamically calibrates noise injection to achieve an optimal balance between rigorous privacy guarantees and predictive accuracy. Empirical evaluations on four real-world datasets demonstrate that FedCCTR-LM substantially outperforms existing baselines, offering robust, privacy-preserving, and generalizable cross-domain CTR prediction in heterogeneous, federated environments.
☆ Enhancing Subsequent Video Retrieval via Vision-Language Models (VLMs)
The rapid growth of video content demands efficient and precise retrieval systems. While vision-language models (VLMs) excel in representation learning, they often struggle with adaptive, time-sensitive video retrieval. This paper introduces a novel framework that combines vector similarity search with graph-based data structures. By leveraging VLM embeddings for initial retrieval and modeling contextual relationships among video segments, our approach enables adaptive query refinement and improves retrieval accuracy. Experiments demonstrate its precision, scalability, and robustness, offering an effective solution for interactive video retrieval in dynamic environments.
♻ ☆ RAGO: Systematic Performance Optimization for Retrieval-Augmented Generation Serving
Retrieval-augmented generation (RAG), which combines large language models (LLMs) with retrievals from external knowledge databases, is emerging as a popular approach for reliable LLM serving. However, efficient RAG serving remains an open challenge due to the rapid emergence of many RAG variants and the substantial differences in workload characteristics across them. In this paper, we make three fundamental contributions to advancing RAG serving. First, we introduce RAGSchema, a structured abstraction that captures the wide range of RAG algorithms, serving as a foundation for performance optimization. Second, we analyze several representative RAG workloads with distinct RAGSchema, revealing significant performance variability across these workloads. Third, to address this variability and meet diverse performance requirements, we propose RAGO (Retrieval-Augmented Generation Optimizer), a system optimization framework for efficient RAG serving. Our evaluation shows that RAGO achieves up to a 2x increase in QPS per chip and a 55% reduction in time-to-first-token latency compared to RAG systems built on LLM-system extensions.
comment: 16 pages, 19 figures, 4 tables
♻ ☆ Multi-Aggregator Time-Warping Heterogeneous Graph Neural Network for Personalized Micro-Video Recommendation
Micro-video recommendation is attracting global attention and becoming a popular daily service for people of all ages. Recently, Graph Neural Networks-based micro-video recommendation has displayed performance improvement for many kinds of recommendation tasks. However, the existing works fail to fully consider the characteristics of micro-videos, such as the high timeliness of news nature micro-video recommendation and sequential interactions of frequently changed interests. In this paper, a novel Multi-aggregator Time-warping Heterogeneous Graph Neural Network (MTHGNN) is proposed for personalized news nature micro-video recommendation based on sequential sessions, where characteristics of micro-videos are comprehensively studied, users' preference is mined via multi-aggregator, the temporal and dynamic changes of users' preference are captured, and timeliness is considered. Through the comparison with the state-of-the-arts, the experimental results validate the superiority of our MTHGNN model.
♻ ☆ LitLLM: A Toolkit for Scientific Literature Review
Conducting literature reviews for scientific papers is essential for understanding research, its limitations, and building on existing work. It is a tedious task which makes an automatic literature review generator appealing. Unfortunately, many existing works that generate such reviews using Large Language Models (LLMs) have significant limitations. They tend to hallucinate-generate non-factual information-and ignore the latest research they have not been trained on. To address these limitations, we propose a toolkit that operates on Retrieval Augmented Generation (RAG) principles, specialized prompting and instructing techniques with the help of LLMs. Our system first initiates a web search to retrieve relevant papers by summarizing user-provided abstracts into keywords using an off-the-shelf LLM. Authors can enhance the search by supplementing it with relevant papers or keywords, contributing to a tailored retrieval process. Second, the system re-ranks the retrieved papers based on the user-provided abstract. Finally, the related work section is generated based on the re-ranked results and the abstract. There is a substantial reduction in time and effort for literature review compared to traditional methods, establishing our toolkit as an efficient alternative. Our project page including the demo and toolkit can be accessed here: https://litllm.github.io
♻ ☆ LLMSeR: Enhancing Sequential Recommendation via LLM-based Data Augmentation
Sequential Recommender Systems (SRS) have become a cornerstone of online platforms, leveraging users' historical interaction data to forecast their next potential engagement. Despite their widespread adoption, SRS often grapple with the long-tail user dilemma, resulting in less effective recommendations for individuals with limited interaction records. The advent of Large Language Models (LLMs), with their profound capability to discern semantic relationships among items, has opened new avenues for enhancing SRS through data augmentation. Nonetheless, current methodologies encounter obstacles, including the absence of collaborative signals and the prevalence of hallucination phenomena. In this work, we present LLMSeR, an innovative framework that utilizes Large Language Models (LLMs) to generate pseudo-prior items, thereby improving the efficacy of Sequential Recommender Systems (SRS). To alleviate the challenge of insufficient collaborative signals, we introduce the Semantic Interaction Augmentor (SIA), a method that integrates both semantic and collaborative information to comprehensively augment user interaction data. Moreover, to weaken the adverse effects of hallucination in SRS, we develop the Adaptive Reliability Validation (ARV), a validation technique designed to assess the reliability of the generated pseudo items. Complementing these advancements, we also devise a Dual-Channel Training strategy, ensuring seamless integration of data augmentation into the SRS training process.Extensive experiments conducted with three widely-used SRS models demonstrate the generalizability and efficacy of LLMSeR.
♻ ☆ Typed-RAG: Type-aware Multi-Aspect Decomposition for Non-Factoid Question Answering NAACL 2025
Non-factoid question-answering (NFQA) poses a significant challenge due to its open-ended nature, diverse intents, and the need for multi-aspect reasoning, which renders conventional factoid QA approaches, including retrieval-augmented generation (RAG), inadequate. Unlike factoid questions, non-factoid questions (NFQs) lack definitive answers and require synthesizing information from multiple sources across various reasoning dimensions. To address these limitations, we introduce Typed-RAG, a type-aware multi-aspect decomposition framework within the RAG paradigm for NFQA. Typed-RAG classifies NFQs into distinct types -- such as debate, experience, and comparison -- and applies aspect-based decomposition to refine retrieval and generation strategies. By decomposing multi-aspect NFQs into single-aspect sub-queries and aggregating the results, Typed-RAG generates more informative and contextually relevant responses. To evaluate Typed-RAG, we introduce Wiki-NFQA, a benchmark dataset covering diverse NFQ types. Experimental results demonstrate that Typed-RAG outperforms baselines, thereby highlighting the importance of type-aware decomposition for effective retrieval and generation in NFQA. Our code and dataset are available at https://github.com/TeamNLP/Typed-RAG.
comment: Accepted to NAACL 2025 SRW
♻ ☆ AlignBot: Aligning VLM-powered Customized Task Planning with User Reminders Through Fine-Tuning for Household Robots
This paper presents AlignBot, a novel framework designed to optimize VLM-powered customized task planning for household robots by effectively aligning with user reminders. In domestic settings, aligning task planning with user reminders poses significant challenges due to the limited quantity, diversity, and multimodal nature of the reminders. To address these challenges, AlignBot employs a fine-tuned LLaVA-7B model, functioning as an adapter for GPT-4o. This adapter model internalizes diverse forms of user reminders-such as personalized preferences, corrective guidance, and contextual assistance-into structured instruction-formatted cues that prompt GPT-4o in generating customized task plans. Additionally, AlignBot integrates a dynamic retrieval mechanism that selects task-relevant historical successes as prompts for GPT-4o, further enhancing task planning accuracy. To validate the effectiveness of AlignBot, experiments are conducted in real-world household environments, which are constructed within the laboratory to replicate typical household settings. A multimodal dataset with over 1,500 entries derived from volunteer reminders is used for training and evaluation. The results demonstrate that AlignBot significantly improves customized task planning, outperforming existing LLM- and VLM-powered planners by interpreting and aligning with user reminders, achieving 86.8% success rate compared to the vanilla GPT-4o baseline at 21.6%, reflecting a 65% improvement and over four times greater effectiveness. Supplementary materials are available at: https://yding25.com/AlignBot/
♻ ☆ Knowledge Graph Reasoning Based on Attention GCN
We propose a novel technique to enhance Knowledge Graph Reasoning by combining Graph Convolution Neural Network (GCN) with the Attention Mechanism. This approach utilizes the Attention Mechanism to examine the relationships between entities and their neighboring nodes, which helps to develop detailed feature vectors for each entity. The GCN uses shared parameters to effectively represent the characteristics of adjacent entities. We first learn the similarity of entities for node representation learning. By integrating the attributes of the entities and their interactions, this method generates extensive implicit feature vectors for each entity, improving performance in tasks including entity classification and link prediction, outperforming traditional neural network models. To conclude, this work provides crucial methodological support for a range of applications, such as search engines, question-answering systems, recommendation systems, and data integration tasks.
Multimedia 7
☆ Audio-Enhanced Vision-Language Modeling with Latent Space Broadening for High Quality Data Expansion
Transformer-based multimodal models are widely used in industrial-scale recommendation, search, and advertising systems for content understanding and relevance ranking. Enhancing labeled training data quality and cross-modal fusion significantly improves model performance, influencing key metrics such as quality view rates and ad revenue. High-quality annotations are crucial for advancing content modeling, yet traditional statistical-based active learning (AL) methods face limitations: they struggle to detect overconfident misclassifications and are less effective in distinguishing semantically similar items in deep neural networks. Additionally, audio information plays an increasing role, especially in short-video platforms, yet most pre-trained multimodal architectures primarily focus on text and images. While training from scratch across all three modalities is possible, it sacrifices the benefits of leveraging existing pre-trained visual-language (VL) and audio models. To address these challenges, we propose kNN-based Latent Space Broadening (LSB) to enhance AL efficiency and Vision-Language Modeling with Audio Enhancement (VLMAE), a mid-fusion approach integrating audio into VL models. This system deployed in production systems, leading to significant business gains.
☆ Align Your Rhythm: Generating Highly Aligned Dance Poses with Gating-Enhanced Rhythm-Aware Feature Representation
Automatically generating natural, diverse and rhythmic human dance movements driven by music is vital for virtual reality and film industries. However, generating dance that naturally follows music remains a challenge, as existing methods lack proper beat alignment and exhibit unnatural motion dynamics. In this paper, we propose Danceba, a novel framework that leverages gating mechanism to enhance rhythm-aware feature representation for music-driven dance generation, which achieves highly aligned dance poses with enhanced rhythmic sensitivity. Specifically, we introduce Phase-Based Rhythm Extraction (PRE) to precisely extract rhythmic information from musical phase data, capitalizing on the intrinsic periodicity and temporal structures of music. Additionally, we propose Temporal-Gated Causal Attention (TGCA) to focus on global rhythmic features, ensuring that dance movements closely follow the musical rhythm. We also introduce Parallel Mamba Motion Modeling (PMMM) architecture to separately model upper and lower body motions along with musical features, thereby improving the naturalness and diversity of generated dance movements. Extensive experiments confirm that Danceba outperforms state-of-the-art methods, achieving significantly better rhythmic alignment and motion diversity. Project page: https://danceba.github.io/ .
comment: 10 pages, 6 figures
☆ The CASTLE 2024 Dataset: Advancing the Art of Multimodal Understanding
Egocentric video has seen increased interest in recent years, as it is used in a range of areas. However, most existing datasets are limited to a single perspective. In this paper, we present the CASTLE 2024 dataset, a multimodal collection containing ego- and exo-centric (i.e., first- and third-person perspective) video and audio from 15 time-aligned sources, as well as other sensor streams and auxiliary data. The dataset was recorded by volunteer participants over four days in a fixed location and includes the point of view of 10 participants, with an additional 5 fixed cameras providing an exocentric perspective. The entire dataset contains over 600 hours of UHD video recorded at 50 frames per second. In contrast to other datasets, CASTLE 2024 does not contain any partial censoring, such as blurred faces or distorted audio. The dataset is available via https://castle-dataset.github.io/.
comment: 7 pages, 6 figures, dataset available via https://castle-dataset.github.io/
♻ ☆ GiVE: Guiding Visual Encoder to Perceive Overlooked Information ICME 2025
Multimodal Large Language Models have advanced AI in applications like text-to-video generation and visual question answering. These models rely on visual encoders to convert non-text data into vectors, but current encoders either lack semantic alignment or overlook non-salient objects. We propose the Guiding Visual Encoder to Perceive Overlooked Information (GiVE) approach. GiVE enhances visual representation with an Attention-Guided Adapter (AG-Adapter) module and an Object-focused Visual Semantic Learning module. These incorporate three novel loss terms: Object-focused Image-Text Contrast (OITC) loss, Object-focused Image-Image Contrast (OIIC) loss, and Object-focused Image Discrimination (OID) loss, improving object consideration, retrieval accuracy, and comprehensiveness. Our contributions include dynamic visual focus adjustment, novel loss functions to enhance object retrieval, and the Multi-Object Instruction (MOInst) dataset. Experiments show our approach achieves state-of-the-art performance.
comment: This paper was accepted by ICME 2025
♻ ☆ EgoTextVQA: Towards Egocentric Scene-Text Aware Video Question Answering CVPR 2025
We introduce EgoTextVQA, a novel and rigorously constructed benchmark for egocentric QA assistance involving scene text. EgoTextVQA contains 1.5K ego-view videos and 7K scene-text aware questions that reflect real user needs in outdoor driving and indoor house-keeping activities. The questions are designed to elicit identification and reasoning on scene text in an egocentric and dynamic environment. With EgoTextVQA, we comprehensively evaluate 10 prominent multimodal large language models. Currently, all models struggle, and the best results (Gemini 1.5 Pro) are around 33\% accuracy, highlighting the severe deficiency of these techniques in egocentric QA assistance. Our further investigations suggest that precise temporal grounding and multi-frame reasoning, along with high resolution and auxiliary scene-text inputs, are key for better performance. With thorough analyses and heuristic suggestions, we hope EgoTextVQA can serve as a solid testbed for research in egocentric scene-text QA assistance. Our dataset is released at: https://github.com/zhousheng97/EgoTextVQA.
comment: Accepted by CVPR 2025
♻ ☆ OmniFlow: Any-to-Any Generation with Multi-Modal Rectified Flows
We introduce OmniFlow, a novel generative model designed for any-to-any generation tasks such as text-to-image, text-to-audio, and audio-to-image synthesis. OmniFlow advances the rectified flow (RF) framework used in text-to-image models to handle the joint distribution of multiple modalities. It outperforms previous any-to-any models on a wide range of tasks, such as text-to-image and text-to-audio synthesis. Our work offers three key contributions: First, we extend RF to a multi-modal setting and introduce a novel guidance mechanism, enabling users to flexibly control the alignment between different modalities in the generated outputs. Second, we propose a novel architecture that extends the text-to-image MMDiT architecture of Stable Diffusion 3 and enables audio and text generation. The extended modules can be efficiently pretrained individually and merged with the vanilla text-to-image MMDiT for fine-tuning. Lastly, we conduct a comprehensive study on the design choices of rectified flow transformers for large-scale audio and text generation, providing valuable insights into optimizing performance across diverse modalities. The Code will be available at https://github.com/jacklishufan/OmniFlows.
comment: 19 pages, 14 figures
♻ ☆ Kiss3DGen: Repurposing Image Diffusion Models for 3D Asset Generation
Diffusion models have achieved great success in generating 2D images. However, the quality and generalizability of 3D content generation remain limited. State-of-the-art methods often require large-scale 3D assets for training, which are challenging to collect. In this work, we introduce Kiss3DGen (Keep It Simple and Straightforward in 3D Generation), an efficient framework for generating, editing, and enhancing 3D objects by repurposing a well-trained 2D image diffusion model for 3D generation. Specifically, we fine-tune a diffusion model to generate ''3D Bundle Image'', a tiled representation composed of multi-view images and their corresponding normal maps. The normal maps are then used to reconstruct a 3D mesh, and the multi-view images provide texture mapping, resulting in a complete 3D model. This simple method effectively transforms the 3D generation problem into a 2D image generation task, maximizing the utilization of knowledge in pretrained diffusion models. Furthermore, we demonstrate that our Kiss3DGen model is compatible with various diffusion model techniques, enabling advanced features such as 3D editing, mesh and texture enhancement, etc. Through extensive experiments, we demonstrate the effectiveness of our approach, showcasing its ability to produce high-quality 3D models efficiently.
comment: The first three authors contributed equally to this work
Robotics 52
☆ LLM-Drone: Aerial Additive Manufacturing with Drones Planned Using Large Language Models
Additive manufacturing (AM) has transformed the production landscape by enabling the precision creation of complex geometries. However, AM faces limitations when applied to challenging environments, such as elevated surfaces and remote locations. Aerial additive manufacturing, facilitated by drones, presents a solution to these challenges. However, despite advances in methods for the planning, control, and localization of drones, the accuracy of these methods is insufficient to run traditional feedforward extrusion-based additive manufacturing processes (such as Fused Deposition Manufacturing). Recently, the emergence of LLMs has revolutionized various fields by introducing advanced semantic reasoning and real-time planning capabilities. This paper proposes the integration of LLMs with aerial additive manufacturing to assist with the planning and execution of construction tasks, granting greater flexibility and enabling a feed-back based design and construction system. Using the semantic understanding and adaptability of LLMs, we can overcome the limitations of drone based systems by dynamically generating and adapting building plans on site, ensuring efficient and accurate construction even in constrained environments. Our system is able to design and build structures given only a semantic prompt and has shown success in understanding the spatial environment despite tight planning constraints. Our method's feedback system enables replanning using the LLM if the manufacturing process encounters unforeseen errors, without requiring complicated heuristics or evaluation functions. Combining the semantic planning with automatic error correction, our system achieved a 90% build accuracy, converting simple text prompts to build structures.
comment: 26 Pages, 6 Figures
☆ Shear-based Grasp Control for Multi-fingered Underactuated Tactile Robotic Hands
This paper presents a shear-based control scheme for grasping and manipulating delicate objects with a Pisa/IIT anthropomorphic SoftHand equipped with soft biomimetic tactile sensors on all five fingertips. These `microTac' tactile sensors are miniature versions of the TacTip vision-based tactile sensor, and can extract precise contact geometry and force information at each fingertip for use as feedback into a controller to modulate the grasp while a held object is manipulated. Using a parallel processing pipeline, we asynchronously capture tactile images and predict contact pose and force from multiple tactile sensors. Consistent pose and force models across all sensors are developed using supervised deep learning with transfer learning techniques. We then develop a grasp control framework that uses contact force feedback from all fingertip sensors simultaneously, allowing the hand to safely handle delicate objects even under external disturbances. This control framework is applied to several grasp-manipulation experiments: first, retaining a flexible cup in a grasp without crushing it under changes in object weight; second, a pouring task where the center of mass of the cup changes dynamically; and third, a tactile-driven leader-follower task where a human guides a held object. These manipulation tasks demonstrate more human-like dexterity with underactuated robotic hands by using fast reflexive control from tactile sensing.
comment: 16 pages. 10 figures. Accepted in IEEE Transactions on Robotics, Special Section on Tactile Robotics
☆ Splat-LOAM: Gaussian Splatting LiDAR Odometry and Mapping ICCV 2025
LiDARs provide accurate geometric measurements, making them valuable for ego-motion estimation and reconstruction tasks. Although its success, managing an accurate and lightweight representation of the environment still poses challenges. Both classic and NeRF-based solutions have to trade off accuracy over memory and processing times. In this work, we build on recent advancements in Gaussian Splatting methods to develop a novel LiDAR odometry and mapping pipeline that exclusively relies on Gaussian primitives for its scene representation. Leveraging spherical projection, we drive the refinement of the primitives uniquely from LiDAR measurements. Experiments show that our approach matches the current registration performance, while achieving SOTA results for mapping tasks with minimal GPU requirements. This efficiency makes it a strong candidate for further exploration and potential adoption in real-time robotics estimation tasks.
comment: submitted to ICCV 2025
☆ LLM+MAP: Bimanual Robot Task Planning using Large Language Models and Planning Domain Definition Language
Bimanual robotic manipulation provides significant versatility, but also presents an inherent challenge due to the complexity involved in the spatial and temporal coordination between two hands. Existing works predominantly focus on attaining human-level manipulation skills for robotic hands, yet little attention has been paid to task planning on long-horizon timescales. With their outstanding in-context learning and zero-shot generation abilities, Large Language Models (LLMs) have been applied and grounded in diverse robotic embodiments to facilitate task planning. However, LLMs still suffer from errors in long-horizon reasoning and from hallucinations in complex robotic tasks, lacking a guarantee of logical correctness when generating the plan. Previous works, such as LLM+P, extended LLMs with symbolic planners. However, none have been successfully applied to bimanual robots. New challenges inevitably arise in bimanual manipulation, necessitating not only effective task decomposition but also efficient task allocation. To address these challenges, this paper introduces LLM+MAP, a bimanual planning framework that integrates LLM reasoning and multi-agent planning, automating effective and efficient bimanual task planning. We conduct simulated experiments on various long-horizon manipulation tasks of differing complexity. Our method is built using GPT-4o as the backend, and we compare its performance against plans generated directly by LLMs, including GPT-4o, V3 and also recent strong reasoning models o1 and R1. By analyzing metrics such as planning time, success rate, group debits, and planning-step reduction rate, we demonstrate the superior performance of LLM+MAP, while also providing insights into robotic reasoning. Code is available at https://github.com/Kchu/LLM-MAP.
comment: Code and video are available at https://github.com/Kchu/LLM-MAP
☆ Control the Soft Robot Arm with its Physical Twin
To exploit the compliant capabilities of soft robot arms we require controller which can exploit their physical capabilities. Teleoperation, leveraging a human in the loop, is a key step towards achieving more complex control strategies. Whilst teleoperation is widely used for rigid robots, for soft robots we require teleoperation methods where the configuration of the whole body is considered. We propose a method of using an identical 'physical twin', or demonstrator of the robot. This tendon robot can be back-driven, with the tendon lengths providing configuration perception, and enabling a direct mapping of tendon lengths for the execture. We demonstrate how this teleoperation across the entire configuration of the robot enables complex interactions with exploit the envrionment, such as squeezing into gaps. We also show how this method can generalize to robots which are a larger scale that the physical twin, and how, tuneability of the stiffness properties of the physical twin simplify its use.
☆ TamedPUMA: safe and stable imitation learning with geometric fabrics
Using the language of dynamical systems, Imitation learning (IL) provides an intuitive and effective way of teaching stable task-space motions to robots with goal convergence. Yet, IL techniques are affected by serious limitations when it comes to ensuring safety and fulfillment of physical constraints. With this work, we solve this challenge via TamedPUMA, an IL algorithm augmented with a recent development in motion generation called geometric fabrics. As both the IL policy and geometric fabrics describe motions as artificial second-order dynamical systems, we propose two variations where IL provides a navigation policy for geometric fabrics. The result is a stable imitation learning strategy within which we can seamlessly blend geometrical constraints like collision avoidance and joint limits. Beyond providing a theoretical analysis, we demonstrate TamedPUMA with simulated and real-world tasks, including a 7-DoF manipulator.
comment: 14 pages (10+4), 1+3*5 figures, 1 table, preprint version of accepted paper at L4DC 2025
☆ GAA-TSO: Geometry-Aware Assisted Depth Completion for Transparent and Specular Objects
Transparent and specular objects are frequently encountered in daily life, factories, and laboratories. However, due to the unique optical properties, the depth information on these objects is usually incomplete and inaccurate, which poses significant challenges for downstream robotics tasks. Therefore, it is crucial to accurately restore the depth information of transparent and specular objects. Previous depth completion methods for these objects usually use RGB information as an additional channel of the depth image to perform depth prediction. Due to the poor-texture characteristics of transparent and specular objects, these methods that rely heavily on color information tend to generate structure-less depth predictions. Moreover, these 2D methods cannot effectively explore the 3D structure hidden in the depth channel, resulting in depth ambiguity. To this end, we propose a geometry-aware assisted depth completion method for transparent and specular objects, which focuses on exploring the 3D structural cues of the scene. Specifically, besides extracting 2D features from RGB-D input, we back-project the input depth to a point cloud and build the 3D branch to extract hierarchical scene-level 3D structural features. To exploit 3D geometric information, we design several gated cross-modal fusion modules to effectively propagate multi-level 3D geometric features to the image branch. In addition, we propose an adaptive correlation aggregation strategy to appropriately assign 3D features to the corresponding 2D features. Extensive experiments on ClearGrasp, OOD, TransCG, and STD datasets show that our method outperforms other state-of-the-art methods. We further demonstrate that our method significantly enhances the performance of downstream robotic grasping tasks.
☆ Exploring psychophysiological methods for human-robot collaboration in construction
Psychophysiological methods present a promising approach to fostering enhanced mutual communication and collaboration between human workers and robots. Despite their potential, there is still limited understanding of how to effectively integrate psychophysiological methods to improve human-robot collaboration (HRC) in construction. This paper addresses this gap by critically reviewing the use of psychophysiological methods for HRC within construction environments, employing a concept-methodology-value philosophical framework. The analysis reveals that measuring brain activity using electroencephalography is the most widely used method, while most of the works are still at the proof of concept stage and lack empirical evidence. Three potential research directions were proposed: the integration of multi-modal psychophysiological signals, enriching the existing experimental settings for better generalizability, and leveraging advanced biocompatible or contactless technologies for effective signal detection. The findings should benefit subsequent exploration and practical applications of psychophysiological methods to enable better implementation of robots and support HRC in construction.
☆ HAPI: A Model for Learning Robot Facial Expressions from Human Preferences
Automatic robotic facial expression generation is crucial for human-robot interaction, as handcrafted methods based on fixed joint configurations often yield rigid and unnatural behaviors. Although recent automated techniques reduce the need for manual tuning, they tend to fall short by not adequately bridging the gap between human preferences and model predictions-resulting in a deficiency of nuanced and realistic expressions due to limited degrees of freedom and insufficient perceptual integration. In this work, we propose a novel learning-to-rank framework that leverages human feedback to address this discrepancy and enhanced the expressiveness of robotic faces. Specifically, we conduct pairwise comparison annotations to collect human preference data and develop the Human Affective Pairwise Impressions (HAPI) model, a Siamese RankNet-based approach that refines expression evaluation. Results obtained via Bayesian Optimization and online expression survey on a 35-DOF android platform demonstrate that our approach produces significantly more realistic and socially resonant expressions of Anger, Happiness, and Surprise than those generated by baseline and expert-designed methods. This confirms that our framework effectively bridges the gap between human preferences and model predictions while robustly aligning robotic expression generation with human affective responses.
☆ Behavioral Conflict Avoidance Between Humans and Quadruped Robots in Shared Environments
Nowadays, robots are increasingly operated in environments shared with humans, where conflicts between human and robot behaviors may compromise safety. This paper presents a proactive behavioral conflict avoidance framework based on the principle of adaptation to trends for quadruped robots that not only ensures the robot's safety but also minimizes interference with human activities. It can proactively avoid potential conflicts with approaching humans or other dynamic objects, whether the robot is stationary or in motion, then swiftly resume its tasks once the conflict subsides. An enhanced approach is proposed to achieve precise human detection and tracking on vibratory robot platform equipped with low-cost hybrid solid-state LiDAR. When potential conflict detected, the robot selects an avoidance point and executes an evasion maneuver before resuming its task. This approach contrasts with conventional methods that remain goal-driven, often resulting in aggressive behaviors, such as forcibly bypassing obstacles and causing conflicts or becoming stuck in deadlock scenarios. The selection of avoidance points is achieved by integrating static and dynamic obstacle to generate a potential field map. The robot then searches for feasible regions within this map and determines the optimal avoidance point using an evaluation function. Experimental results demonstrate that the framework significantly reduces interference with human activities, enhances the safety of both robots and persons.
comment: 7 pages, 9 figures. This work has been submitted to the IEEE for possible publication
☆ Autonomous Exploration-Based Precise Mapping for Mobile Robots through Stepwise and Consistent Motions
This paper presents an autonomous exploration framework. It is designed for indoor ground mobile robots that utilize laser Simultaneous Localization and Mapping (SLAM), ensuring process completeness and precise mapping results. For frontier search, the local-global sampling architecture based on multiple Rapidly Exploring Random Trees (RRTs) is employed. Traversability checks during RRT expansion and global RRT pruning upon map updates eliminate unreachable frontiers, reducing potential collisions and deadlocks. Adaptive sampling density adjustments, informed by obstacle distribution, enhance exploration coverage potential. For frontier point navigation, a stepwise consistent motion strategy is adopted, wherein the robot strictly drives straight on approximately equidistant line segments in the polyline path and rotates in place at segment junctions. This simplified, decoupled motion pattern improves scan-matching stability and mitigates map drift. For process control, the framework serializes frontier point selection and navigation, avoiding oscillation caused by frequent goal changes in conventional parallelized processes. The waypoint retracing mechanism is introduced to generate repeated observations, triggering loop closure detection and backend optimization in graph-based SLAM, thereby improving map consistency and precision. Experiments in both simulation and real-world scenarios validate the effectiveness of the framework. It achieves improved mapping coverage and precision in more challenging environments compared to baseline 2D exploration algorithms. It also shows robustness in supporting resource-constrained robot platforms and maintaining mapping consistency across various LiDAR field-of-view (FoV) configurations.
comment: 8 pages, 11 figures. This work has been submitted to the IEEE for possible publication
☆ Targetless 6DoF Calibration of LiDAR and 2D Scanning Radar Based on Cylindrical Occupancy
Owing to the capability for reliable and all-weather long-range sensing, the fusion of LiDAR and Radar has been widely applied to autonomous vehicles for robust perception. In practical operation, well manually calibrated extrinsic parameters, which are crucial for the fusion of multi-modal sensors, may drift due to the vibration. To address this issue, we present a novel targetless calibration approach, termed LiRaCo, for the extrinsic 6DoF calibration of LiDAR and Radar sensors. Although both types of sensors can obtain geometric information, bridging the geometric correspondences between multi-modal data without any clues of explicit artificial markers is nontrivial, mainly due to the low vertical resolution of scanning Radar. To achieve the targetless calibration, LiRaCo leverages a spatial occupancy consistency between LiDAR point clouds and Radar scans in a common cylindrical representation, considering the increasing data sparsity with distance for both sensors. Specifically, LiRaCo expands the valid Radar scanned pixels into 3D occupancy grids to constrain LiDAR point clouds based on spatial consistency. Consequently, a cost function involving extrinsic calibration parameters is formulated based on the spatial overlap of 3D grids and LiDAR points. Extrinsic parameters are finally estimated by optimizing the cost function. Comprehensive quantitative and qualitative experiments on two real outdoor datasets with different LiDAR sensors demonstrate the feasibility and accuracy of the proposed method. The source code will be publicly available.
☆ Extending Behavior Trees for Robotic Missions with Quality Requirements
Context and motivation: In recent years, behavior trees have gained growing interest within the robotics community as a specification and control switching mechanism for the different tasks that form a robotics mission. Problem: Given the rising complexity and prevalence of robotic systems, it is increasingly challenging and important for practitioners to design high-quality missions that meet certain qualities, for instance, to consider potential failures or mitigate safety risks. In software requirements engineering, quality or non-functional requirements have long been recognized as a key factor in system success. Currently, qualities are not represented in behavior tree models, which capture a robotic mission, making it difficult to assess the extent to which different mission components comply with those qualities. Principal ideas: In this paper, we propose an extension for behavior trees to have qualities and quality requirements explicitly represented in robotics missions. We provide a meta-model for the extension, develop a domain-specific language (DSL), and describe how we integrated our DSL in one of the most used languages in robotics for developing behavior trees, BehaviorTree.CPP. A preliminary evaluation of the implemented DSL shows promising results for the feasibility of our approach and the need for similar DSLs. Contribution: Our approach paves the way for incorporating qualities into the behavior model of robotics missions. This promotes early expression of qualities in robotics missions, and a better overview of missions components and their contribution to the satisfaction of quality concerns.
comment: 17 pages, 6 figures, Requirements Engineering: Foundation for Software Quality (REFSQ) 2025
☆ Somatic Safety: An Embodied Approach Towards Safe Human-Robot Interaction
As robots enter the messy human world so the vital matter of safety takes on a fresh complexion with physical contact becoming inevitable and even desirable. We report on an artistic-exploration of how dancers, working as part of a multidisciplinary team, engaged in contact improvisation exercises to explore the opportunities and challenges of dancing with cobots. We reveal how they employed their honed bodily senses and physical skills to engage with the robots aesthetically and yet safely, interleaving improvised physical manipulations with reflections to grow their knowledge of how the robots behaved and felt. We introduce somatic safety, a holistic mind-body approach in which safety is learned, felt and enacted through bodily contact with robots in addition to being reasoned about. We conclude that robots need to be better designed for people to hold them and might recognise tacit safety cues among people.We propose that safety should be learned through iterative bodily experience interleaved with reflection.
comment: ACM/IEEE International Conference on Human-Robot Interaction (HRI'25)
☆ Reachability-Guaranteed Optimal Control for the Interception of Dynamic Targets under Uncertainty
Intercepting dynamic objects in uncertain environments involves a significant unresolved challenge in modern robotic systems. Current control approaches rely solely on estimated information, and results lack guarantees of robustness and feasibility. In this work, we introduce a novel method to tackle the interception of targets whose motion is affected by known and bounded uncertainty. Our approach introduces new techniques of reachability analysis for rigid bodies, leveraged to guarantee feasibility of interception under uncertain conditions. We then propose a Reachability-Guaranteed Optimal Control Problem, ensuring robustness and guaranteed reachability to a target set of configurations. We demonstrate the methodology in the case study of an interception maneuver of a tumbling target in space.
☆ Rude Humans and Vengeful Robots: Examining Human Perceptions of Robot Retaliatory Intentions in Professional Settings
Humans and robots are increasingly working in personal and professional settings. In workplace settings, humans and robots may work together as colleagues, potentially leading to social expectations, or violation thereof. Extant research has primarily sought to understand social interactions and expectations in personal rather than professional settings, and none of these studies have examined negative outcomes arising from violations of social expectations. This paper reports the results of a 2x3 online experiment that used a unique first-person perspective video to immerse participants in a collaborative workplace setting. The results are nuanced and reveal that while robots are expected to act in accordance with social expectations despite human behavior, there are benefits for robots perceived as being the bigger person in the face of human rudeness. Theoretical and practical implications are provided which discuss the import of these findings for the design of social robots.
comment: This is the author version of the manuscript submitted to ACM Transactions on Human-Robot Interaction. The final version, if accepted, will be published by ACM and available via the ACM Digital Library. 12 pages, 1 figure, 2 tables
☆ Deep Learning for Human Locomotion Analysis in Lower-Limb Exoskeletons: A Comparative Study
Wearable robotics for lower-limb assistance have become a pivotal area of research, aiming to enhance mobility for individuals with physical impairments or augment the performance of able-bodied users. Accurate and adaptive control systems are essential to ensure seamless interaction between the wearer and the robotic device, particularly when navigating diverse and dynamic terrains. Despite the recent advances in neural networks for time series analysis, no attempts have been directed towards the classification of ground conditions, categorized into five classes and subsequently determining the ramp's slope and stair's height. In this respect, this paper presents an experimental comparison between eight deep neural network backbones to predict high-level locomotion parameters across diverse terrains. All the models are trained on the publicly available CAMARGO 2021 dataset. IMU-only data equally or outperformed IMU+EMG inputs, promoting a cost-effective and efficient design. Indeeds, using three IMU sensors, the LSTM achieved high terrain classification accuracy (0.94 +- 0.04) and precise ramp slope (1.95 +- 0.58{\deg}) and the CNN-LSTM a stair height (15.65 +- 7.40 mm) estimations. As a further contribution, SHAP analysis justified sensor reduction without performance loss, ensuring a lightweight setup. The system operates with ~2 ms inference time, supporting real-time applications. The code is code available at https://github.com/cosbidev/Human-Locomotion-Identification.
comment: 26 pages, 6 figures
☆ Safe On-Orbit Dislodging of Deployable Structures via Robust Adaptive MPC
This paper proposes a novel robust adaptive model predictive controller for on-orbit dislodging. We consider the scenario where a servicer, equipped with a robot arm, must dislodge a client, a time-varying system composed of an underpowered jammed solar panel with a hybrid hinge system on a space station. Our approach leverages online set-membership identification to reduce the uncertainty to provide robust safety guarantees during dislodging despite bounded disturbances while balancing exploration and exploitation effectively in the parameter space. The feasibility of the developed robust adaptive MPC method is also examined through dislodging simulations and hardware experiments in zero-gravity and gravity environments, respectively. In addition, the advantages of our method are shown through comparison experiments with several state-of-the-art control schemes for both accuracy of parameter estimation and control performance.
comment: This paper has been submitted to IEEE Transactions on Control Systems Technology and is being reviewed
☆ SGFormer: Satellite-Ground Fusion for 3D Semantic Scene Completion
Recently, camera-based solutions have been extensively explored for scene semantic completion (SSC). Despite their success in visible areas, existing methods struggle to capture complete scene semantics due to frequent visual occlusions. To address this limitation, this paper presents the first satellite-ground cooperative SSC framework, i.e., SGFormer, exploring the potential of satellite-ground image pairs in the SSC task. Specifically, we propose a dual-branch architecture that encodes orthogonal satellite and ground views in parallel, unifying them into a common domain. Additionally, we design a ground-view guidance strategy that corrects satellite image biases during feature encoding, addressing misalignment between satellite and ground views. Moreover, we develop an adaptive weighting strategy that balances contributions from satellite and ground views. Experiments demonstrate that SGFormer outperforms the state of the art on SemanticKITTI and SSCBench-KITTI-360 datasets. Our code is available on https://github.com/gxytcrc/SGFormer.
☆ DyWA: Dynamics-adaptive World Action Model for Generalizable Non-prehensile Manipulation
Nonprehensile manipulation is crucial for handling objects that are too thin, large, or otherwise ungraspable in unstructured environments. While conventional planning-based approaches struggle with complex contact modeling, learning-based methods have recently emerged as a promising alternative. However, existing learning-based approaches face two major limitations: they heavily rely on multi-view cameras and precise pose tracking, and they fail to generalize across varying physical conditions, such as changes in object mass and table friction. To address these challenges, we propose the Dynamics-Adaptive World Action Model (DyWA), a novel framework that enhances action learning by jointly predicting future states while adapting to dynamics variations based on historical trajectories. By unifying the modeling of geometry, state, physics, and robot actions, DyWA enables more robust policy learning under partial observability. Compared to baselines, our method improves the success rate by 31.5% using only single-view point cloud observations in the simulation. Furthermore, DyWA achieves an average success rate of 68% in real-world experiments, demonstrating its ability to generalize across diverse object geometries, adapt to varying table friction, and robustness in challenging scenarios such as half-filled water bottles and slippery surfaces.
comment: Project Page:https://pku-epic.github.io/DyWA/
☆ BEAC: Imitating Complex Exploration and Task-oriented Behaviors for Invisible Object Nonprehensile Manipulation
Applying imitation learning (IL) is challenging to nonprehensile manipulation tasks of invisible objects with partial observations, such as excavating buried rocks. The demonstrator must make such complex action decisions as exploring to find the object and task-oriented actions to complete the task while estimating its hidden state, perhaps causing inconsistent action demonstration and high cognitive load problems. For these problems, work in human cognitive science suggests that promoting the use of pre-designed, simple exploration rules for the demonstrator may alleviate the problems of action inconsistency and high cognitive load. Therefore, when performing imitation learning from demonstrations using such exploration rules, it is important to accurately imitate not only the demonstrator's task-oriented behavior but also his/her mode-switching behavior (exploratory or task-oriented behavior) under partial observation. Based on the above considerations, this paper proposes a novel imitation learning framework called Belief Exploration-Action Cloning (BEAC), which has a switching policy structure between a pre-designed exploration policy and a task-oriented action policy trained on the estimated belief states based on past history. In simulation and real robot experiments, we confirmed that our proposed method achieved the best task performance, higher mode and action prediction accuracies, while reducing the cognitive load in the demonstration indicated by a user study.
comment: 27 pages
☆ Displacement-Actuated Continuum Robots: A Joint Space Abstraction
The displacement-actuated continuum robot as an abstraction has been shown as a key abstraction to significantly simplify and improve approaches due to its relation to the Clarke transform. To highlight further potentials, we revisit and extend this abstraction that features an increasingly popular length extension and an underutilized twisting. For each extension, the corresponding mapping from the joint values to the local coordinates of the manifold embedded in the joint spaces is provided. Each mapping is characterized by its compactness and linearity.
comment: 11 pages, 4 figures, 3 tables
♻ ☆ Adver-City: Open-Source Multi-Modal Dataset for Collaborative Perception Under Adverse Weather Conditions
Adverse weather conditions pose a significant challenge to the widespread adoption of Autonomous Vehicles (AVs) by impacting sensors like LiDARs and cameras. Even though Collaborative Perception (CP) improves AV perception in difficult conditions, existing CP datasets lack adverse weather conditions. To address this, we introduce Adver-City, the first open-source synthetic CP dataset focused on adverse weather conditions. Simulated in CARLA with OpenCDA, it contains over 24 thousand frames, over 890 thousand annotations, and 110 unique scenarios across six different weather conditions: clear weather, soft rain, heavy rain, fog, foggy heavy rain and, for the first time in a synthetic CP dataset, glare. It has six object categories including pedestrians and cyclists, and uses data from vehicles and roadside units featuring LiDARs, RGB and semantic segmentation cameras, GNSS, and IMUs. Its scenarios, based on real crash reports, depict the most relevant road configurations for adverse weather and poor visibility conditions, varying in object density, with both dense and sparse scenes, allowing for novel testing conditions of CP models. Benchmarks run on the dataset show that weather conditions created challenging conditions for perception models, with CoBEVT scoring 58.30/52.44/38.90 (AP@30/50/70). The dataset, code and documentation are available at https://labs.cs.queensu.ca/quarrg/datasets/adver-city/.
comment: 13 pages
♻ ☆ Electrostatic Clutches Enable Simultaneous Mechanical Multiplexing
Actuating robotic systems with multiple degrees of freedom (DoF) traditionally requires numerous motors, leading to increased size, weight, cost, and power consumption. Mechanical multiplexing offers a solution by enabling a single actuator to control multiple DoF. However, existing multiplexers have either been limited to electrically controlled time-based multiplexing that control one DoF at a time or have relied on mechanical switching to control multiple DoF simultaneously. There is a strong need for a system that can perform electrically controlled multiplexing for both time-based and simultaneous control of multiple DoF. This study introduces a novel electrostatic capstan clutch-based mechanical multiplexer that enables high-force, single-motor control of multiple DoF. Here, we show that our system achieves both single-input-single-output (SISO) and single-input-multipleoutput (SIMO) actuation, allowing bidirectional control and position holding with minimal power consumption. Each output can actuate a 22.24 N load, limited by clutch performance, up to 5 cm. The number of outputs and actuation length is currently limited by the length of the drive shaft. We demonstrate the integration of our system into a 4-DoF commercial robotic hand using a single motor. These findings show that electrostatic clutchbased multiplexing provides a scalable and energy-efficient design solution for high-DoF robotic platforms, opening new possibilities for lightweight and power-efficient actuation in robotics.
♻ ☆ Physically-Feasible Reactive Synthesis for Terrain-Adaptive Locomotion via Trajectory Optimization and Symbolic Repair
We propose an integrated planning framework for quadrupedal locomotion over dynamically changing, unforeseen terrains. Existing approaches either rely on heuristics for instantaneous foothold selection--compromising safety and versatility--or solve expensive trajectory optimization problems with complex terrain features and long time horizons. In contrast, our framework leverages reactive synthesis to generate correct-by-construction controllers at the symbolic level, and mixed-integer convex programming (MICP) for dynamic and physically feasible footstep planning for each symbolic transition. We use a high-level manager to reduce the large state space in synthesis by incorporating local environment information, improving synthesis scalability. To handle specifications that cannot be met due to dynamic infeasibility, and to minimize costly MICP solves, we leverage a symbolic repair process to generate only necessary symbolic transitions. During online execution, re-running the MICP with real-world terrain data, along with runtime symbolic repair, bridges the gap between offline synthesis and online execution. We demonstrate, in simulation, our framework's capabilities to discover missing locomotion skills and react promptly in safety-critical environments, such as scattered stepping stones and rebars.
♻ ☆ Understanding and Imitating Human-Robot Motion with Restricted Visual Fields
When working around humans, it is important to model their perception limitations in order to predict their behavior more accurately. In this work, we consider agents with a limited field of view, viewing range, and ability to miss objects within the viewing range (e.g., transparency). By considering the observation model independently from the motion policy, we can better predict the agent's behavior by considering these limitations and approximating them. We perform a user study where human operators navigate a cluttered scene while scanning the region for obstacles with a limited field of view and range. Using imitation learning, we show that a robot can adopt a human's strategy for observing an environment with limitations on observation and navigate with minimal collision with dynamic and static obstacles. We also show that this learned model helps it successfully navigate a physical hardware vehicle in real-time.
♻ ☆ MultiNash-PF: A Particle Filtering Approach for Computing Multiple Local Generalized Nash Equilibria in Trajectory Games
Modern robotic systems frequently engage in complex multi-agent interactions, many of which are inherently multi-modal, meaning they can lead to multiple distinct outcomes. To interact effectively, robots must recognize the possible interaction modes and adapt to the one preferred by other agents. In this work, we propose an efficient algorithm for capturing the multimodality in multi-agent interactions. We leverage a game-theoretic planner to model interaction outcomes as equilibria where \emph{each equilibrium} corresponds to a distinct interaction \emph{mode}. We then develop an efficient algorithm to identify all the equilibria, allowing robots to reason about multiple interaction modes. More specifically, we formulate interactive planning as Constrained Potential Trajectory Games (CPTGs) and model interaction outcomes by local Generalized Nash equilibria (GNEs) of the game. CPTGs are a class of games for which a local GNE can be found by solving a single constrained optimal control problem where a potential function is minimized. We propose to integrate the potential game approach with implicit particle filtering, a sample-efficient method for non-convex trajectory optimization. We utilize implicit particle filtering to identify the coarse estimates of multiple local minimizers of the game's potential function. MultiNash-PF then refines these estimates with optimization solvers, obtaining different local GNEs. We show through numerical simulations that MultiNash-PF reduces computation time by up to 50\% compared to a baseline. We further demonstrate the effectiveness of our algorithm in real-world human-robot interaction scenarios, where it successfully accounts for the multi-modal nature of interactions and resolves potential conflicts in real-time.
♻ ☆ Strategic Decision-Making in Multi-Agent Domains: A Weighted Constrained Potential Dynamic Game Approach
In interactive multi-agent settings, decision-making and planning are challenging mainly due to the agents' interconnected objectives. Dynamic game theory offers a formal framework for analyzing such intricacies. Yet, solving constrained dynamic games and determining the interaction outcome in the form of generalized Nash Equilibria (GNE) pose computational challenges due to the need for solving constrained coupled optimal control problems. In this paper, we address this challenge by proposing to leverage the special structure of many real-world multi-agent interactions. More specifically, our key idea is to leverage constrained dynamic potential games, which are games for which GNE can be found by solving a single constrained optimal control problem associated with minimizing the potential function. We argue that constrained dynamic potential games can effectively facilitate interactive decision-making in many multi-agent interactions. We will identify structures in realistic multi-agent interactive scenarios that can be transformed into weighted constrained potential dynamic games (WCPDGs). We will show that the GNE of the resulting WCPDG can be obtained by solving a single constrained optimal control problem. We will demonstrate the effectiveness of the proposed method through various simulation studies and show that we achieve significant improvements in solve time compared to state-of-the-art game solvers. We further provide experimental validation of our proposed method in a navigation setup involving two quadrotors carrying a rigid object while avoiding collisions with two humans.
comment: in IEEE Transactions on Robotics 2025
♻ ☆ SOUS VIDE: Cooking Visual Drone Navigation Policies in a Gaussian Splatting Vacuum
We propose a new simulator, training approach, and policy architecture, collectively called SOUS VIDE, for end-to-end visual drone navigation. Our trained policies exhibit zero-shot sim-to-real transfer with robust real-world performance using only onboard perception and computation. Our simulator, called FiGS, couples a computationally simple drone dynamics model with a high visual fidelity Gaussian Splatting scene reconstruction. FiGS can quickly simulate drone flights producing photorealistic images at up to 130 fps. We use FiGS to collect 100k-300k image/state-action pairs from an expert MPC with privileged state and dynamics information, randomized over dynamics parameters and spatial disturbances. We then distill this expert MPC into an end-to-end visuomotor policy with a lightweight neural architecture, called SV-Net. SV-Net processes color image, optical flow and IMU data streams into low-level thrust and body rate commands at 20 Hz onboard a drone. Crucially, SV-Net includes a learned module for low-level control that adapts at runtime to variations in drone dynamics. In a campaign of 105 hardware experiments, we show SOUS VIDE policies to be robust to 30% mass variations, 40 m/s wind gusts, 60% changes in ambient brightness, shifting or removing objects from the scene, and people moving aggressively through the drone's visual field. Code, data, and experiment videos can be found on our project page: https://stanfordmsl.github.io/SousVide/.
♻ ☆ Social Gesture Recognition in spHRI: Leveraging Fabric-Based Tactile Sensing on Humanoid Robots ICRA 25
Humans are able to convey different messages using only touch. Equipping robots with the ability to understand social touch adds another modality in which humans and robots can communicate. In this paper, we present a social gesture recognition system using a fabric-based, large-scale tactile sensor placed onto the arms of a humanoid robot. We built a social gesture dataset using multiple participants and extracted temporal features for classification. By collecting tactile data on a humanoid robot, our system provides insights into human-robot social touch, and displays that the use of fabric based sensors could be a potential way of advancing the development of spHRI systems for more natural and effective communication.
comment: Accepted to ICRA 25. 8 pages, 8 figures
♻ ☆ Discrete Policy: Learning Disentangled Action Space for Multi-Task Robotic Manipulation ICRA 2025
Learning visuomotor policy for multi-task robotic manipulation has been a long-standing challenge for the robotics community. The difficulty lies in the diversity of action space: typically, a goal can be accomplished in multiple ways, resulting in a multimodal action distribution for a single task. The complexity of action distribution escalates as the number of tasks increases. In this work, we propose \textbf{Discrete Policy}, a robot learning method for training universal agents capable of multi-task manipulation skills. Discrete Policy employs vector quantization to map action sequences into a discrete latent space, facilitating the learning of task-specific codes. These codes are then reconstructed into the action space conditioned on observations and language instruction. We evaluate our method on both simulation and multiple real-world embodiments, including both single-arm and bimanual robot settings. We demonstrate that our proposed Discrete Policy outperforms a well-established Diffusion Policy baseline and many state-of-the-art approaches, including ACT, Octo, and OpenVLA. For example, in a real-world multi-task training setting with five tasks, Discrete Policy achieves an average success rate that is 26\% higher than Diffusion Policy and 15\% higher than OpenVLA. As the number of tasks increases to 12, the performance gap between Discrete Policy and Diffusion Policy widens to 32.5\%, further showcasing the advantages of our approach. Our work empirically demonstrates that learning multi-task policies within the latent space is a vital step toward achieving general-purpose agents.
comment: Accept to ICRA 2025
♻ ☆ SuperPC: A Single Diffusion Model for Point Cloud Completion, Upsampling, Denoising, and Colorization
Point cloud (PC) processing tasks-such as completion, upsampling, denoising, and colorization-are crucial in applications like autonomous driving and 3D reconstruction. Despite substantial advancements, prior approaches often address each of these tasks independently, with separate models focused on individual issues. However, this isolated approach fails to account for the fact that defects like incompleteness, low resolution, noise, and lack of color frequently coexist, with each defect influencing and correlating with the others. Simply applying these models sequentially can lead to error accumulation from each model, along with increased computational costs. To address these challenges, we introduce SuperPC, the first unified diffusion model capable of concurrently handling all four tasks. Our approach employs a three-level-conditioned diffusion framework, enhanced by a novel spatial-mix-fusion strategy, to leverage the correlations among these four defects for simultaneous, efficient processing. We show that SuperPC outperforms the state-of-the-art specialized models as well as their combination on all four individual tasks.
♻ ☆ Bootstrapping Object-level Planning with Large Language Models ICRA 2025
We introduce a new method that extracts knowledge from a large language model (LLM) to produce object-level plans, which describe high-level changes to object state, and uses them to bootstrap task and motion planning (TAMP). Existing work uses LLMs to directly output task plans or generate goals in representations like PDDL. However, these methods fall short because they rely on the LLM to do the actual planning or output a hard-to-satisfy goal. Our approach instead extracts knowledge from an LLM in the form of plan schemas as an object-level representation called functional object-oriented networks (FOON), from which we automatically generate PDDL subgoals. Our method markedly outperforms alternative planning strategies in completing several pick-and-place tasks in simulation.
comment: Accepted to ICRA 2025; 11 pages (6 pages + 1 page references + 4 pages appendix); for demo videos, please see https://davidpaulius.github.io/olp_llm/
♻ ☆ A Bayesian Modeling Framework for Estimation and Ground Segmentation of Cluttered Staircases
Autonomous robot navigation in complex environments requires robust perception as well as high-level scene understanding due to perceptual challenges, such as occlusions, and uncertainty introduced by robot movement. For example, a robot climbing a cluttered staircase can misinterpret clutter as a step, misrepresenting the state and compromising safety. This requires robust state estimation methods capable of inferring the underlying structure of the environment even from incomplete sensor data. In this paper, we introduce a novel method for robust state estimation of staircases. To address the challenge of perceiving occluded staircases extending beyond the robot's field-of-view, our approach combines an infinite-width staircase representation with a finite endpoint state to capture the overall staircase structure. This representation is integrated into a Bayesian inference framework to fuse noisy measurements enabling accurate estimation of staircase location even with partial observations and occlusions. Additionally, we present a segmentation algorithm that works in conjunction with the staircase estimation pipeline to accurately identify clutter-free regions on a staircase. Our method is extensively evaluated on real robot across diverse staircases, demonstrating significant improvements in estimation accuracy and segmentation performance compared to baseline approaches.
♻ ☆ Contraction Theory for Nonlinear Stability Analysis and Learning-based Control: A Tutorial Overview
Contraction theory is an analytical tool to study differential dynamics of a non-autonomous (i.e., time-varying) nonlinear system under a contraction metric defined with a uniformly positive definite matrix, the existence of which results in a necessary and sufficient characterization of incremental exponential stability of multiple solution trajectories with respect to each other. By using a squared differential length as a Lyapunov-like function, its nonlinear stability analysis boils down to finding a suitable contraction metric that satisfies a stability condition expressed as a linear matrix inequality, indicating that many parallels can be drawn between well-known linear systems theory and contraction theory for nonlinear systems. Furthermore, contraction theory takes advantage of a superior robustness property of exponential stability used in conjunction with the comparison lemma. This yields much-needed safety and stability guarantees for neural network-based control and estimation schemes, without resorting to a more involved method of using uniform asymptotic stability for input-to-state stability. Such distinctive features permit the systematic construction of a contraction metric via convex optimization, thereby obtaining an explicit exponential bound on the distance between a time-varying target trajectory and solution trajectories perturbed externally due to disturbances and learning errors. The objective of this paper is, therefore, to present a tutorial overview of contraction theory and its advantages in nonlinear stability analysis of deterministic and stochastic systems, with an emphasis on deriving formal robustness and stability guarantees for various learning-based and data-driven automatic control methods. In particular, we provide a detailed review of techniques for finding contraction metrics and associated control and estimation laws using deep neural networks.
comment: Annual Reviews in Control, Preprint Version, Accepted, Oct. 1st
♻ ☆ Code-as-Monitor: Constraint-aware Visual Programming for Reactive and Proactive Robotic Failure Detection CVPR 2025
Automatic detection and prevention of open-set failures are crucial in closed-loop robotic systems. Recent studies often struggle to simultaneously identify unexpected failures reactively after they occur and prevent foreseeable ones proactively. To this end, we propose Code-as-Monitor (CaM), a novel paradigm leveraging the vision-language model (VLM) for both open-set reactive and proactive failure detection. The core of our method is to formulate both tasks as a unified set of spatio-temporal constraint satisfaction problems and use VLM-generated code to evaluate them for real-time monitoring. To enhance the accuracy and efficiency of monitoring, we further introduce constraint elements that abstract constraint-related entities or their parts into compact geometric elements. This approach offers greater generality, simplifies tracking, and facilitates constraint-aware visual programming by leveraging these elements as visual prompts. Experiments show that CaM achieves a 28.7% higher success rate and reduces execution time by 31.8% under severe disturbances compared to baselines across three simulators and a real-world setting. Moreover, CaM can be integrated with open-loop control policies to form closed-loop systems, enabling long-horizon tasks in cluttered scenes with dynamic environments.
comment: Accepted by CVPR 2025. Project page: https://zhoues.github.io/Code-as-Monitor/
♻ ☆ Data-driven Camera and Lidar Simulation Models for Autonomous Driving: A Review from Generative Models to Volume Renderers
Perception sensors, particularly camera and Lidar, are key elements of Autonomous Driving Systems (ADS) that enable them to comprehend their surroundings to informed driving and control decisions. Therefore, developing realistic simulation models for these sensors is essential for conducting effective simulation-based testing of ADS. Moreover, the rise of deep learning-based perception models has increased the utility of sensor simulation models for synthesising diverse training datasets. The traditional sensor simulation models rely on computationally expensive physics-based algorithms, specifically in complex systems such as ADS. Hence, the current potential resides in data-driven approaches, fuelled by the exceptional performance of deep generative models in capturing high-dimensional data distribution and volume renderers in accurately representing scenes. This paper reviews the current state-of-the-art data-driven camera and Lidar simulation models and their evaluation methods. It explores a spectrum of models from the novel perspective of generative models and volume renderers. Generative models are discussed in terms of their input-output types, while volume renderers are categorised based on their input encoding. Finally, the paper illustrates commonly used evaluation techniques for assessing sensor simulation models and highlights the existing research gaps in the area.
comment: To be published in IEEE Transactions on Intelligent Vehicles
♻ ☆ Global SLAM Using 5G ToA Integration: Performance Analysis with Unknown Base Stations and Loop Closure Alternatives
This paper presents a novel approach that integrates 5G Time of Arrival (ToA) measurements into ORB-SLAM3 to enable global localization and enhance mapping capabilities for indoor drone navigation. We extend ORB-SLAM3's optimization pipeline to jointly process ToA data from 5G base stations alongside visual and inertial measurements while estimating system biases. This integration transforms the inherently local SLAM estimates into globally referenced trajectories and effectively resolves scale ambiguity in monocular configurations. Our method is evaluated using both Aerolab indoor datasets with RGB-D cameras and the EuRoC MAV benchmark, complemented by simulated 5G ToA measurements at 28 GHz and 78 GHz frequencies using MATLAB and QuaDRiGa. Extensive experiments across multiple SLAM configurations demonstrate that ToA integration enables consistent global positioning across all modes while maintaining local accuracy. For monocular configurations, ToA integration successfully resolves scale ambiguity and improves consistency. We further investigate scenarios with unknown base station positions and demonstrate that ToA measurements can effectively serve as an alternative to loop closure for drift correction. We also analyze how different geometric arrangements of base stations impact SLAM performance. Comparative analysis with state-of-the-art methods, including UWB-VO, confirms our approach's robustness even with lower measurement frequencies and sequential base station operation. The results validate that 5G ToA integration provides substantial benefits for global SLAM applications, particularly in challenging indoor environments where accurate positioning is critical.
♻ ☆ Synthesizing multi-log grasp poses in cluttered environments
Multi-object grasping is a challenging task. It is important for energy and cost-efficient operation of industrial crane manipulators, such as those used to collect tree logs from the forest floor and on forest machines. In this work, we used synthetic data from physics simulations to explore how data-driven modeling can be used to infer multi-object grasp poses from images. We showed that convolutional neural networks can be trained specifically for synthesizing multi-object grasps. Using RGB-Depth images and instance segmentation masks as input, a U-Net model outputs grasp maps with the corresponding grapple orientation and opening width. Given an observation of a pile of logs, the model can be used to synthesize and rate the possible grasp poses and select the most suitable one, with the possibility to respect changing operational constraints such as lift capacity and reach. When tested on previously unseen data, the proposed model found successful grasp poses with an accuracy up to 96%.
comment: 21 pages, 14 figures
♻ ☆ Set-membership target search and tracking within an unknown cluttered area using cooperating UAVs equipped with vision systems
This paper addresses the problem of target search and tracking using a fleet of cooperating UAVs evolving in some unknown region of interest containing an a priori unknown number of moving ground targets. Each drone is equipped with an embedded Computer Vision System (CVS), providing an image with labeled pixels and a depth map of the observed part of its environment. Moreover, a box containing the corresponding pixels in the image frame is available when a UAV identifies a target. Hypotheses regarding information provided by the pixel classification, depth map construction, and target identification algorithms are proposed to allow its exploitation by set-membership approaches. A set-membership target location estimator is developed using the information provided by the CVS. Each UAV evaluates sets guaranteed to contain the location of the identified targets and a set possibly containing the locations of targets still to be identified. Then, each UAV uses these sets to search and track targets cooperatively.
comment: This work has been submitted to Elsevier / ScienceDirect for possible publication
♻ ☆ Efficient Training of Generalizable Visuomotor Policies via Control-Aware Augmentation
Improving generalization is one key challenge in embodied AI, where obtaining large-scale datasets across diverse scenarios is costly. Traditional weak augmentations, such as cropping and flipping, are insufficient for improving a model's performance in new environments. Existing data augmentation methods often disrupt task-relevant information in images, potentially degrading performance. To overcome these challenges, we introduce EAGLE, an efficient training framework for generalizable visuomotor policies that improves upon existing methods by (1) enhancing generalization by applying augmentation only to control-related regions identified through a self-supervised control-aware mask and (2) improving training stability and efficiency by distilling knowledge from an expert to a visuomotor student policy, which is then deployed to unseen environments without further fine-tuning. Comprehensive experiments on three domains, including the DMControl Generalization Benchmark, the enhanced Robot Manipulation Distraction Benchmark, and a long-sequential drawer-opening task, validate the effectiveness of our method.
♻ ☆ A Unified Framework for Real-Time Failure Handling in Robotics Using Vision-Language Models, Reactive Planner and Behavior Trees
Robotic systems often face execution failures due to unexpected obstacles, sensor errors, or environmental changes. Traditional failure recovery methods rely on predefined strategies or human intervention, making them less adaptable. This paper presents a unified failure recovery framework that combines Vision-Language Models (VLMs), a reactive planner, and Behavior Trees (BTs) to enable real-time failure handling. Our approach includes pre-execution verification, which checks for potential failures before execution, and reactive failure handling, which detects and corrects failures during execution by verifying existing BT conditions, adding missing preconditions and, when necessary, generating new skills. The framework uses a scene graph for structured environmental perception and an execution history for continuous monitoring, enabling context-aware and adaptive failure handling. We evaluate our framework through real-world experiments with an ABB YuMi robot on tasks like peg insertion, object sorting, and drawer placement, as well as in AI2-THOR simulator. Compared to using pre-execution and reactive methods separately, our approach achieves higher task success rates and greater adaptability. Ablation studies highlight the importance of VLM-based reasoning, structured scene representation, and execution history tracking for effective failure recovery in robotics.
♻ ☆ Neuromorphic Attitude Estimation and Control
The real-world application of small drones is mostly hampered by energy limitations. Neuromorphic computing promises extremely energy-efficient AI for autonomous flight but is still challenging to train and deploy on real robots. To reap the maximal benefits from neuromorphic computing, it is necessary to perform all autonomy functions end-to-end on a single neuromorphic chip, from low-level attitude control to high-level navigation. This research presents the first neuromorphic control system using a spiking neural network (SNN) to effectively map a drone's raw sensory input directly to motor commands. We apply this method to low-level attitude estimation and control for a quadrotor, deploying the SNN on a tiny Crazyflie. We propose a modular SNN, separately training and then merging estimation and control sub-networks. The SNN is trained with imitation learning, using a flight dataset of sensory-motor pairs. Post-training, the network is deployed on the Crazyflie, issuing control commands from sensor inputs at 500Hz. Furthermore, for the training procedure we augmented training data by flying a controller with additional excitation and time-shifting the target data to enhance the predictive capabilities of the SNN. On the real drone, the perception-to-control SNN tracks attitude commands with an average error of 3.0 degrees, compared to 2.7 degrees for the regular flight stack. We also show the benefits of the proposed learning modifications for reducing the average tracking error and reducing oscillations. Our work shows the feasibility of performing neuromorphic end-to-end control, laying the basis for highly energy-efficient and low-latency neuromorphic autopilots.
♻ ☆ GAPartManip: A Large-scale Part-centric Dataset for Material-Agnostic Articulated Object Manipulation ICRA 2025
Effectively manipulating articulated objects in household scenarios is a crucial step toward achieving general embodied artificial intelligence. Mainstream research in 3D vision has primarily focused on manipulation through depth perception and pose detection. However, in real-world environments, these methods often face challenges due to imperfect depth perception, such as with transparent lids and reflective handles. Moreover, they generally lack the diversity in part-based interactions required for flexible and adaptable manipulation. To address these challenges, we introduced a large-scale part-centric dataset for articulated object manipulation that features both photo-realistic material randomization and detailed annotations of part-oriented, scene-level actionable interaction poses. We evaluated the effectiveness of our dataset by integrating it with several state-of-the-art methods for depth estimation and interaction pose prediction. Additionally, we proposed a novel modular framework that delivers superior and robust performance for generalizable articulated object manipulation. Our extensive experiments demonstrate that our dataset significantly improves the performance of depth perception and actionable interaction pose prediction in both simulation and real-world scenarios. More information and demos can be found at: https://pku-epic.github.io/GAPartManip/.
comment: Accepted by ICRA 2025. Project page: https://pku-epic.github.io/GAPartManip/
♻ ☆ Incremental Learning for Robot Shared Autonomy
Shared autonomy holds promise for improving the usability and accessibility of assistive robotic arms, but current methods often rely on costly expert demonstrations and remain static after pretraining, limiting their ability to handle real-world variations. Even with extensive training data, unforeseen challenges--especially those that fundamentally alter task dynamics, such as unexpected obstacles or spatial constraints--can cause assistive policies to break down, leading to ineffective or unreliable assistance. To address this, we propose ILSA, an Incrementally Learned Shared Autonomy framework that continuously refines its assistive policy through user interactions, adapting to real-world challenges beyond the scope of pre-collected data. At the core of ILSA is a structured fine-tuning mechanism that enables continual improvement with each interaction by effectively integrating limited new interaction data while preserving prior knowledge, ensuring a balance between adaptation and generalization. A user study with 20 participants demonstrates ILSA's effectiveness, showing faster task completion and improved user experience compared to static alternatives. Code and videos are available at https://ilsa-robo.github.io/.
♻ ☆ HOTFormerLoc: Hierarchical Octree Transformer for Versatile Lidar Place Recognition Across Ground and Aerial Views CVPR 2025
We present HOTFormerLoc, a novel and versatile Hierarchical Octree-based TransFormer, for large-scale 3D place recognition in both ground-to-ground and ground-to-aerial scenarios across urban and forest environments. We propose an octree-based multi-scale attention mechanism that captures spatial and semantic features across granularities. To address the variable density of point distributions from spinning lidar, we present cylindrical octree attention windows to reflect the underlying distribution during attention. We introduce relay tokens to enable efficient global-local interactions and multi-scale representation learning at reduced computational cost. Our pyramid attentional pooling then synthesises a robust global descriptor for end-to-end place recognition in challenging environments. In addition, we introduce CS-Wild-Places, a novel 3D cross-source dataset featuring point cloud data from aerial and ground lidar scans captured in dense forests. Point clouds in CS-Wild-Places contain representational gaps and distinctive attributes such as varying point densities and noise patterns, making it a challenging benchmark for cross-view localisation in the wild. HOTFormerLoc achieves a top-1 average recall improvement of 5.5% - 11.5% on the CS-Wild-Places benchmark. Furthermore, it consistently outperforms SOTA 3D place recognition methods, with an average performance gain of 4.9% on well-established urban and forest datasets. The code and CS-Wild-Places benchmark is available at https://csiro-robotics.github.io/HOTFormerLoc.
comment: 16 pages, 13 figures, 10 tables, Accepted to CVPR 2025
♻ ☆ AlignBot: Aligning VLM-powered Customized Task Planning with User Reminders Through Fine-Tuning for Household Robots
This paper presents AlignBot, a novel framework designed to optimize VLM-powered customized task planning for household robots by effectively aligning with user reminders. In domestic settings, aligning task planning with user reminders poses significant challenges due to the limited quantity, diversity, and multimodal nature of the reminders. To address these challenges, AlignBot employs a fine-tuned LLaVA-7B model, functioning as an adapter for GPT-4o. This adapter model internalizes diverse forms of user reminders-such as personalized preferences, corrective guidance, and contextual assistance-into structured instruction-formatted cues that prompt GPT-4o in generating customized task plans. Additionally, AlignBot integrates a dynamic retrieval mechanism that selects task-relevant historical successes as prompts for GPT-4o, further enhancing task planning accuracy. To validate the effectiveness of AlignBot, experiments are conducted in real-world household environments, which are constructed within the laboratory to replicate typical household settings. A multimodal dataset with over 1,500 entries derived from volunteer reminders is used for training and evaluation. The results demonstrate that AlignBot significantly improves customized task planning, outperforming existing LLM- and VLM-powered planners by interpreting and aligning with user reminders, achieving 86.8% success rate compared to the vanilla GPT-4o baseline at 21.6%, reflecting a 65% improvement and over four times greater effectiveness. Supplementary materials are available at: https://yding25.com/AlignBot/
♻ ☆ An Integrated Approach to Robotic Object Grasping and Manipulation
In response to the growing challenges of manual labor and efficiency in warehouse operations, Amazon has embarked on a significant transformation by incorporating robotics to assist with various tasks. While a substantial number of robots have been successfully deployed for tasks such as item transportation within warehouses, the complex process of object picking from shelves remains a significant challenge. This project addresses the issue by developing an innovative robotic system capable of autonomously fulfilling a simulated order by efficiently selecting specific items from shelves. A distinguishing feature of the proposed robotic system is its capacity to navigate the challenge of uncertain object positions within each bin of the shelf. The system is engineered to autonomously adapt its approach, employing strategies that enable it to efficiently locate and retrieve the desired items, even in the absence of pre-established knowledge about their placements.
comment: I am making big changes in the paper and continuing its further development with other instituition
♻ ☆ KARMA: Augmenting Embodied AI Agents with Long-and-short Term Memory Systems
Embodied AI agents responsible for executing interconnected, long-sequence household tasks often face difficulties with in-context memory, leading to inefficiencies and errors in task execution. To address this issue, we introduce KARMA, an innovative memory system that integrates long-term and short-term memory modules, enhancing large language models (LLMs) for planning in embodied agents through memory-augmented prompting. KARMA distinguishes between long-term and short-term memory, with long-term memory capturing comprehensive 3D scene graphs as representations of the environment, while short-term memory dynamically records changes in objects' positions and states. This dual-memory structure allows agents to retrieve relevant past scene experiences, thereby improving the accuracy and efficiency of task planning. Short-term memory employs strategies for effective and adaptive memory replacement, ensuring the retention of critical information while discarding less pertinent data. Compared to state-of-the-art embodied agents enhanced with memory, our memory-augmented embodied AI agent improves success rates by 1.3x and 2.3x in Composite Tasks and Complex Tasks within the AI2-THOR simulator, respectively, and enhances task execution efficiency by 3.4x and 62.7x. Furthermore, we demonstrate that KARMA's plug-and-play capability allows for seamless deployment on real-world robotic systems, such as mobile manipulation platforms.Through this plug-and-play memory system, KARMA significantly enhances the ability of embodied agents to generate coherent and contextually appropriate plans, making the execution of complex household tasks more efficient. The experimental videos from the work can be found at https://youtu.be/4BT7fnw9ehs. Our code is available at https://github.com/WZX0Swarm0Robotics/KARMA/tree/master.
♻ ☆ Moto: Latent Motion Token as the Bridging Language for Learning Robot Manipulation from Videos
Recent developments in Large Language Models pre-trained on extensive corpora have shown significant success in various natural language processing tasks with minimal fine-tuning. This success offers new promise for robotics, which has long been constrained by the high cost of action-labeled data. We ask: given the abundant video data containing interaction-related knowledge available as a rich "corpus", can a similar generative pre-training approach be effectively applied to enhance robot learning? The key challenge is to identify an effective representation for autoregressive pre-training that benefits robot manipulation tasks. Inspired by the way humans learn new skills through observing dynamic environments, we propose that effective robotic learning should emphasize motion-related knowledge, which is closely tied to low-level actions and is hardware-agnostic, facilitating the transfer of learned motions to actual robot actions. To this end, we introduce Moto, which converts video content into latent Motion Token sequences by a Latent Motion Tokenizer, learning a bridging "language" of motion from videos in an unsupervised manner. We pre-train Moto-GPT through motion token autoregression, enabling it to capture diverse visual motion knowledge. After pre-training, Moto-GPT demonstrates the promising ability to produce semantically interpretable motion tokens, predict plausible motion trajectories, and assess trajectory rationality through output likelihood. To transfer learned motion priors to real robot actions, we implement a co-fine-tuning strategy that seamlessly bridges latent motion token prediction and real robot control. Extensive experiments show that the fine-tuned Moto-GPT exhibits superior robustness and efficiency on robot manipulation benchmarks, underscoring its effectiveness in transferring knowledge from video data to downstream visual manipulation tasks.
comment: Project released at: https://chenyi99.github.io/moto/ Code released at: https://github.com/TencentARC/Moto Update: Added content related to real-world robot experiments and learning from human videos; Modified author information
♻ ☆ SPINE: Online Semantic Planning for Missions with Incomplete Natural Language Specifications in Unstructured Environments ICRA
As robots become increasingly capable, users will want to describe high-level missions and have robots infer the relevant details. Because pre-built maps are difficult to obtain in many realistic settings, accomplishing such missions will require the robot to map and plan online. While many semantic planning methods operate online, they are typically designed for well specified missions such as object search or exploration. Recently, Large Language Models (LLMs) have demonstrated powerful contextual reasoning abilities over a range of robotic tasks described in natural language. However, existing LLM-enabled planners typically do not consider online planning or complex missions; rather, relevant subtasks and semantics are provided by a pre-built map or a user. We address these limitations via SPINE, an online planner for missions with incomplete mission specifications provided in natural language. The planner uses an LLM to reason about subtasks implied by the mission specification and then realizes these subtasks in a receding horizon framework. Tasks are automatically validated for safety and refined online with new map observations. We evaluate SPINE in simulation and real-world settings with missions that require multiple steps of semantic reasoning and exploration in cluttered outdoor environments of over 20,000m$^2$. Compared to baselines that use existing LLM-enabled planning approaches, our method is over twice as efficient in terms of time and distance, requires less user interactions, and does not require a full map. Additional resources are provided at https://zacravichandran.github.io/SPINE.
comment: Accepted to the International Conference on Robotics and Automation (ICRA) 2025
♻ ☆ Robotic In-Hand Manipulation for Large-Range Precise Object Movement: The RGMC Champion Solution
In-hand manipulation using multiple dexterous fingers is a critical robotic skill that can reduce the reliance on large arm motions, thereby saving space and energy. This letter focuses on in-grasp object movement, which refers to manipulating an object to a desired pose through only finger motions within a stable grasp. The key challenge lies in simultaneously achieving high precision and large-range movements while maintaining a constant stable grasp. To address this problem, we propose a simple and practical approach based on kinematic trajectory optimization with no need for pretraining or object geometries, which can be easily applied to novel objects in real-world scenarios. Adopting this approach, we won the championship for the in-hand manipulation track at the 9th Robotic Grasping and Manipulation Competition (RGMC) held at ICRA 2024. Implementation details, discussion, and further quantitative experimental results are presented in this letter, which aims to comprehensively evaluate our approach and share our key takeaways from the competition. Supplementary materials including video and code are available at https://rgmc-xl-team.github.io/ingrasp_manipulation .
comment: Accepted by RA-L. Project website: https://rgmc-xl-team.github.io/ingrasp_manipulation
Artificial Intelligence 188
☆ Measuring the Robustness of Audio Deepfake Detectors
Deepfakes have become a universal and rapidly intensifying concern of generative AI across various media types such as images, audio, and videos. Among these, audio deepfakes have been of particular concern due to the ease of high-quality voice synthesis and distribution via platforms such as social media and robocalls. Consequently, detecting audio deepfakes plays a critical role in combating the growing misuse of AI-synthesized speech. However, real-world scenarios often introduce various audio corruptions, such as noise, modification, and compression, that may significantly impact detection performance. This work systematically evaluates the robustness of 10 audio deepfake detection models against 16 common corruptions, categorized into noise perturbation, audio modification, and compression. Using both traditional deep learning models and state-of-the-art foundation models, we make four unique observations. First, our findings show that while most models demonstrate strong robustness to noise, they are notably more vulnerable to modifications and compression, especially when neural codecs are applied. Second, speech foundation models generally outperform traditional models across most scenarios, likely due to their self-supervised learning paradigm and large-scale pre-training. Third, our results show that increasing model size improves robustness, albeit with diminishing returns. Fourth, we demonstrate how targeted data augmentation during training can enhance model resilience to unseen perturbations. A case study on political speech deepfakes highlights the effectiveness of foundation models in achieving high accuracy under real-world conditions. These findings emphasize the importance of developing more robust detection frameworks to ensure reliability in practical deployment settings.
☆ Fairness-Driven LLM-based Causal Discovery with Active Learning and Dynamic Scoring
Causal discovery (CD) plays a pivotal role in numerous scientific fields by clarifying the causal relationships that underlie phenomena observed in diverse disciplines. Despite significant advancements in CD algorithms that enhance bias and fairness analyses in machine learning, their application faces challenges due to the high computational demands and complexities of large-scale data. This paper introduces a framework that leverages Large Language Models (LLMs) for CD, utilizing a metadata-based approach akin to the reasoning processes of human experts. By shifting from pairwise queries to a more scalable breadth-first search (BFS) strategy, the number of required queries is reduced from quadratic to linear in terms of variable count, thereby addressing scalability concerns inherent in previous approaches. This method utilizes an Active Learning (AL) and a Dynamic Scoring Mechanism that prioritizes queries based on their potential information gain, combining mutual information, partial correlation, and LLM confidence scores to refine the causal graph more efficiently and accurately. This BFS query strategy reduces the required number of queries significantly, thereby addressing scalability concerns inherent in previous approaches. This study provides a more scalable and efficient solution for leveraging LLMs in fairness-driven CD, highlighting the effects of the different parameters on performance. We perform fairness analyses on the inferred causal graphs, identifying direct and indirect effects of sensitive attributes on outcomes. A comparison of these analyses against those from graphs produced by baseline methods highlights the importance of accurate causal graph construction in understanding bias and ensuring fairness in machine learning systems.
☆ Autonomous Radiotherapy Treatment Planning Using DOLA: A Privacy-Preserving, LLM-Based Optimization Agent
Radiotherapy treatment planning is a complex and time-intensive process, often impacted by inter-planner variability and subjective decision-making. To address these challenges, we introduce Dose Optimization Language Agent (DOLA), an autonomous large language model (LLM)-based agent designed for optimizing radiotherapy treatment plans while rigorously protecting patient privacy. DOLA integrates the LLaMa3.1 LLM directly with a commercial treatment planning system, utilizing chain-of-thought prompting, retrieval-augmented generation (RAG), and reinforcement learning (RL). Operating entirely within secure local infrastructure, this agent eliminates external data sharing. We evaluated DOLA using a retrospective cohort of 18 prostate cancer patients prescribed 60 Gy in 20 fractions, comparing model sizes (8 billion vs. 70 billion parameters) and optimization strategies (No-RAG, RAG, and RAG+RL) over 10 planning iterations. The 70B model demonstrated significantly improved performance, achieving approximately 16.4% higher final scores than the 8B model. The RAG approach outperformed the No-RAG baseline by 19.8%, and incorporating RL accelerated convergence, highlighting the synergy of retrieval-based memory and reinforcement learning. Optimal temperature hyperparameter analysis identified 0.4 as providing the best balance between exploration and exploitation. This proof of concept study represents the first successful deployment of locally hosted LLM agents for autonomous optimization of treatment plans within a commercial radiotherapy planning system. By extending human-machine interaction through interpretable natural language reasoning, DOLA offers a scalable and privacy-conscious framework, with significant potential for clinical implementation and workflow improvement.
comment: 19 pages, 5 figures, preprint
☆ Audio-Enhanced Vision-Language Modeling with Latent Space Broadening for High Quality Data Expansion
Transformer-based multimodal models are widely used in industrial-scale recommendation, search, and advertising systems for content understanding and relevance ranking. Enhancing labeled training data quality and cross-modal fusion significantly improves model performance, influencing key metrics such as quality view rates and ad revenue. High-quality annotations are crucial for advancing content modeling, yet traditional statistical-based active learning (AL) methods face limitations: they struggle to detect overconfident misclassifications and are less effective in distinguishing semantically similar items in deep neural networks. Additionally, audio information plays an increasing role, especially in short-video platforms, yet most pre-trained multimodal architectures primarily focus on text and images. While training from scratch across all three modalities is possible, it sacrifices the benefits of leveraging existing pre-trained visual-language (VL) and audio models. To address these challenges, we propose kNN-based Latent Space Broadening (LSB) to enhance AL efficiency and Vision-Language Modeling with Audio Enhancement (VLMAE), a mid-fusion approach integrating audio into VL models. This system deployed in production systems, leading to significant business gains.
☆ Learning Multi-Level Features with Matryoshka Sparse Autoencoders
Sparse autoencoders (SAEs) have emerged as a powerful tool for interpreting neural networks by extracting the concepts represented in their activations. However, choosing the size of the SAE dictionary (i.e. number of learned concepts) creates a tension: as dictionary size increases to capture more relevant concepts, sparsity incentivizes features to be split or absorbed into more specific features, leaving high-level features missing or warped. We introduce Matryoshka SAEs, a novel variant that addresses these issues by simultaneously training multiple nested dictionaries of increasing size, forcing the smaller dictionaries to independently reconstruct the inputs without using the larger dictionaries. This organizes features hierarchically - the smaller dictionaries learn general concepts, while the larger dictionaries learn more specific concepts, without incentive to absorb the high-level features. We train Matryoshka SAEs on Gemma-2-2B and TinyStories and find superior performance on sparse probing and targeted concept erasure tasks, more disentangled concept representations, and reduced feature absorption. While there is a minor tradeoff with reconstruction performance, we believe Matryoshka SAEs are a superior alternative for practical tasks, as they enable training arbitrarily large SAEs while retaining interpretable features at different levels of abstraction.
☆ PRIMAL: Physically Reactive and Interactive Motor Model for Avatar Learning
To build a motor system of the interactive avatar, it is essential to develop a generative motion model drives the body to move through 3D space in a perpetual, realistic, controllable, and responsive manner. Although motion generation has been extensively studied, most methods do not support ``embodied intelligence'' due to their offline setting, slow speed, limited motion lengths, or unnatural movements. To overcome these limitations, we propose PRIMAL, an autoregressive diffusion model that is learned with a two-stage paradigm, inspired by recent advances in foundation models. In the pretraining stage, the model learns motion dynamics from a large number of sub-second motion segments, providing ``motor primitives'' from which more complex motions are built. In the adaptation phase, we employ a ControlNet-like adaptor to fine-tune the motor control for semantic action generation and spatial target reaching. Experiments show that physics effects emerge from our training. Given a single-frame initial state, our model not only generates unbounded, realistic, and controllable motion, but also enables the avatar to be responsive to induced impulses in real time. In addition, we can effectively and efficiently adapt our base model to few-shot personalized actions and the task of spatial control. Evaluations show that our proposed method outperforms state-of-the-art baselines. We leverage the model to create a real-time character animation system in Unreal Engine that is highly responsive and natural. Code, models, and more results are available at: https://yz-cnsdqz.github.io/eigenmotion/PRIMAL
comment: 18 pages
☆ Bayesian Teaching Enables Probabilistic Reasoning in Large Language Models
Artificial intelligence systems based on large language models (LLMs) are increasingly used as agents that interact with users and with the world. To do so successfully, LLMs need to construct internal representations of the world and form probabilistic beliefs about those representations. To provide a user with personalized recommendations, for example, the LLM needs to gradually infer the user's preferences, over the course of multiple interactions. To evaluate whether contemporary LLMs are able to do so, we use the Bayesian inference framework from probability theory, which lays out the optimal way to update an agent's beliefs as it receives new information. We first show that the LLMs do not update their beliefs as expected from the Bayesian framework, and that consequently their predictions do not improve as expected as more information becomes available, even less so than we find is the case for humans. To address this issue, we teach the LLMs to reason in a Bayesian manner by training them to mimic the predictions of an optimal Bayesian model. We find that this approach not only significantly improves the LLM's performance on the particular recommendation task it is trained on, but also enables generalization to other tasks. This suggests that this method endows the LLM with broader Bayesian reasoning skills. More generally, our results indicate that LLMs can learn about reasoning strategies effectively and generalize those skills to new domains, which in part explains LLMs' empirical success.
☆ A Predictive Services Architecture for Efficient Airspace Operations
Predicting air traffic congestion and flow management is essential for airlines and Air Navigation Service Providers (ANSP) to enhance operational efficiency. Accurate estimates of future airport capacity and airspace density are vital for better airspace management, reducing air traffic controller workload and fuel consumption, ultimately promoting sustainable aviation. While existing literature has addressed these challenges, data management and query processing remain complex due to the vast volume of high-rate air traffic data. Many analytics use cases require a common pre-processing infrastructure, as ad-hoc approaches are insufficient. Additionally, linear prediction models often fall short, necessitating more advanced techniques. This paper presents a data processing and predictive services architecture that ingests large, uncorrelated, and noisy streaming data to forecast future airspace system states. The system continuously collects raw data, periodically compresses it, and stores it in NoSQL databases for efficient query processing. For prediction, the system learns from historical traffic by extracting key features such as airport arrival and departure events, sector boundary crossings, weather parameters, and other air traffic data. These features are input into various regression models, including linear, non-linear, and ensemble models, with the best-performing model selected for predictions. We evaluate this infrastructure across three prediction use cases in the US National Airspace System (NAS) and a segment of European airspace, using extensive real operations data, confirming that our system can predict future system states efficiently and accurately.
☆ Language Models May Verbatim Complete TextThey Were Not Explicitly Trained On
An important question today is whether a given text was used to train a large language model (LLM). A \emph{completion} test is often employed: check if the LLM completes a sufficiently complex text. This, however, requires a ground-truth definition of membership; most commonly, it is defined as a member based on the $n$-gram overlap between the target text and any text in the dataset. In this work, we demonstrate that this $n$-gram based membership definition can be effectively gamed. We study scenarios where sequences are \emph{non-members} for a given $n$ and we find that completion tests still succeed. We find many natural cases of this phenomenon by retraining LLMs from scratch after removing all training samples that were completed; these cases include exact duplicates, near-duplicates, and even short overlaps. They showcase that it is difficult to find a single viable choice of $n$ for membership definitions. Using these insights, we design adversarial datasets that can cause a given target sequence to be completed without containing it, for any reasonable choice of $n$. Our findings highlight the inadequacy of $n$-gram membership, suggesting membership definitions fail to account for auxiliary information available to the training algorithm.
comment: Main text: 9 pages, 7 figures, 1 table. Appendix: 29 pages, 20 tables, 15 figures
☆ Improving Quantization with Post-Training Model Expansion
The size of a model has been a strong predictor of its quality, as well as its cost. As such, the trade-off between model cost and quality has been well-studied. Post-training optimizations like quantization and pruning have typically focused on reducing the overall volume of pre-trained models to reduce inference costs while maintaining model quality. However, recent advancements have introduced optimization techniques that, interestingly, expand models post-training, increasing model size to improve quality when reducing volume. For instance, to enable 4-bit weight and activation quantization, incoherence processing often necessitates inserting online Hadamard rotations in the compute graph, and preserving highly sensitive weights often calls for additional higher precision computations. However, if application requirements cannot be met, the prevailing solution is to relax quantization constraints. In contrast, we demonstrate post-training model expansion is a viable strategy to improve model quality within a quantization co-design space, and provide theoretical justification. We show it is possible to progressively and selectively expand the size of a pre-trained large language model (LLM) to improve model quality without end-to-end retraining. In particular, when quantizing the weights and activations to 4 bits for Llama3 1B, we reduce the zero-shot accuracy gap to full precision by an average of 3% relative to both QuaRot and SpinQuant with only 5% more parameters, which is still a 3.8% reduction in volume relative to a BF16 reference model.
☆ Follow-up Question Generation For Enhanced Patient-Provider Conversations
Follow-up question generation is an essential feature of dialogue systems as it can reduce conversational ambiguity and enhance modeling complex interactions. Conversational contexts often pose core NLP challenges such as (i) extracting relevant information buried in fragmented data sources, and (ii) modeling parallel thought processes. These two challenges occur frequently in medical dialogue as a doctor asks questions based not only on patient utterances but also their prior EHR data and current diagnostic hypotheses. Asking medical questions in asynchronous conversations compounds these issues as doctors can only rely on static EHR information to motivate follow-up questions. To address these challenges, we introduce FollowupQ, a novel framework for enhancing asynchronous medical conversation. FollowupQ is a multi-agent framework that processes patient messages and EHR data to generate personalized follow-up questions, clarifying patient-reported medical conditions. FollowupQ reduces requisite provider follow-up communications by 34%. It also improves performance by 17% and 5% on real and synthetic data, respectively. We also release the first public dataset of asynchronous medical messages with linked EHR data alongside 2,300 follow-up questions written by clinical experts for the wider NLP research community.
comment: 17 Pages, 7 Figures, 6 Tables
☆ Large Language Models (LLMs) for Source Code Analysis: applications, models and datasets
Large language models (LLMs) and transformer-based architectures are increasingly utilized for source code analysis. As software systems grow in complexity, integrating LLMs into code analysis workflows becomes essential for enhancing efficiency, accuracy, and automation. This paper explores the role of LLMs for different code analysis tasks, focusing on three key aspects: 1) what they can analyze and their applications, 2) what models are used and 3) what datasets are used, and the challenges they face. Regarding the goal of this research, we investigate scholarly articles that explore the use of LLMs for source code analysis to uncover research developments, current trends, and the intellectual structure of this emerging field. Additionally, we summarize limitations and highlight essential tools, datasets, and key challenges, which could be valuable for future work.
☆ Efficient Knowledge Distillation via Curriculum Extraction
Knowledge distillation is a technique used to train a small student network using the output generated by a large teacher network, and has many empirical advantages~\citep{Hinton2015DistillingTK}. While the standard one-shot approach to distillation only uses the output of the final teacher network, recent work~\citep{panigrahi2024progressive} has shown that using intermediate checkpoints from the teacher's training process as an implicit ``curriculum'' for progressive distillation can significantly speed up training. However, such schemes require storing these checkpoints, and often require careful selection of the intermediate checkpoints to train on, which can be impractical for large-scale training. In this paper, we show that a curriculum can be \emph{extracted} from just the fully trained teacher network, and that this extracted curriculum can give similar efficiency benefits to those of progressive distillation. Our extraction scheme is natural; we use a random projection of the hidden representations of the teacher network to progressively train the student network, before training using the output of the full network. We show that our scheme significantly outperforms one-shot distillation and achieves a performance similar to that of progressive distillation for learning sparse parities with two-layer networks, and provide theoretical guarantees for this setting. Additionally, we show that our method outperforms one-shot distillation even when using transformer-based architectures, both for sparse-parity learning, and language modeling tasks.
☆ ProtoGS: Efficient and High-Quality Rendering with 3D Gaussian Prototypes
3D Gaussian Splatting (3DGS) has made significant strides in novel view synthesis but is limited by the substantial number of Gaussian primitives required, posing challenges for deployment on lightweight devices. Recent methods address this issue by compressing the storage size of densified Gaussians, yet fail to preserve rendering quality and efficiency. To overcome these limitations, we propose ProtoGS to learn Gaussian prototypes to represent Gaussian primitives, significantly reducing the total Gaussian amount without sacrificing visual quality. Our method directly uses Gaussian prototypes to enable efficient rendering and leverage the resulting reconstruction loss to guide prototype learning. To further optimize memory efficiency during training, we incorporate structure-from-motion (SfM) points as anchor points to group Gaussian primitives. Gaussian prototypes are derived within each group by clustering of K-means, and both the anchor points and the prototypes are optimized jointly. Our experiments on real-world and synthetic datasets prove that we outperform existing methods, achieving a substantial reduction in the number of Gaussians, and enabling high rendering speed while maintaining or even enhancing rendering fidelity.
☆ SaudiCulture: A Benchmark for Evaluating Large Language Models Cultural Competence within Saudi Arabia
Large Language Models (LLMs) have demonstrated remarkable capabilities in natural language processing; however, they often struggle to accurately capture and reflect cultural nuances. This research addresses this challenge by focusing on Saudi Arabia, a country characterized by diverse dialects and rich cultural traditions. We introduce SaudiCulture, a novel benchmark designed to evaluate the cultural competence of LLMs within the distinct geographical and cultural contexts of Saudi Arabia. SaudiCulture is a comprehensive dataset of questions covering five major geographical regions, such as West, East, South, North, and Center, along with general questions applicable across all regions. The dataset encompasses a broad spectrum of cultural domains, including food, clothing, entertainment, celebrations, and crafts. To ensure a rigorous evaluation, SaudiCulture includes questions of varying complexity, such as open-ended, single-choice, and multiple-choice formats, with some requiring multiple correct answers. Additionally, the dataset distinguishes between common cultural knowledge and specialized regional aspects. We conduct extensive evaluations on five LLMs, such as GPT-4, Llama 3.3, FANAR, Jais, and AceGPT, analyzing their performance across different question types and cultural contexts. Our findings reveal that all models experience significant performance declines when faced with highly specialized or region-specific questions, particularly those requiring multiple correct responses. Additionally, certain cultural categories are more easily identifiable than others, further highlighting inconsistencies in LLMs cultural understanding. These results emphasize the importance of incorporating region-specific knowledge into LLMs training to enhance their cultural competence.
comment: 34 pages, under-review
☆ What's Producible May Not Be Reachable: Measuring the Steerability of Generative Models
How should we evaluate the quality of generative models? Many existing metrics focus on a model's producibility, i.e. the quality and breadth of outputs it can generate. However, the actual value from using a generative model stems not just from what it can produce but whether a user with a specific goal can produce an output that satisfies that goal. We refer to this property as steerability. In this paper, we first introduce a mathematical framework for evaluating steerability independently from producibility. Steerability is more challenging to evaluate than producibility because it requires knowing a user's goals. We address this issue by creating a benchmark task that relies on one key idea: sample an output from a generative model and ask users to reproduce it. We implement this benchmark in a large-scale user study of text-to-image models and large language models. Despite the ability of these models to produce high-quality outputs, they all perform poorly on steerabilty. This suggests that we need to focus on improving the steerability of generative models. We show such improvements are indeed possible: through reinforcement learning techniques, we create an alternative steering mechanism for image models that achieves more than 2x improvement on this benchmark.
☆ Your voice is your voice: Supporting Self-expression through Speech Generation and LLMs in Augmented and Alternative Communication
In this paper, we present Speak Ease: an augmentative and alternative communication (AAC) system to support users' expressivity by integrating multimodal input, including text, voice, and contextual cues (conversational partner and emotional tone), with large language models (LLMs). Speak Ease combines automatic speech recognition (ASR), context-aware LLM-based outputs, and personalized text-to-speech technologies to enable more personalized, natural-sounding, and expressive communication. Through an exploratory feasibility study and focus group evaluation with speech and language pathologists (SLPs), we assessed Speak Ease's potential to enable expressivity in AAC. The findings highlight the priorities and needs of AAC users and the system's ability to enhance user expressivity by supporting more personalized and contextually relevant communication. This work provides insights into the use of multimodal inputs and LLM-driven features to improve AAC systems and support expressivity.
☆ Spatiotemporal Learning with Context-aware Video Tubelets for Ultrasound Video Analysis
Computer-aided pathology detection algorithms for video-based imaging modalities must accurately interpret complex spatiotemporal information by integrating findings across multiple frames. Current state-of-the-art methods operate by classifying on video sub-volumes (tubelets), but they often lose global spatial context by focusing only on local regions within detection ROIs. Here we propose a lightweight framework for tubelet-based object detection and video classification that preserves both global spatial context and fine spatiotemporal features. To address the loss of global context, we embed tubelet location, size, and confidence as inputs to the classifier. Additionally, we use ROI-aligned feature maps from a pre-trained detection model, leveraging learned feature representations to increase the receptive field and reduce computational complexity. Our method is efficient, with the spatiotemporal tubelet classifier comprising only 0.4M parameters. We apply our approach to detect and classify lung consolidation and pleural effusion in ultrasound videos. Five-fold cross-validation on 14,804 videos from 828 patients shows our method outperforms previous tubelet-based approaches and is suited for real-time workflows.
comment: ISBI Oral 2025
☆ Language-specific Neurons Do Not Facilitate Cross-Lingual Transfer NAACL 2025
Multilingual large language models (LLMs) aim towards robust natural language understanding across diverse languages, yet their performance significantly degrades on low-resource languages. This work explores whether existing techniques to identify language-specific neurons can be leveraged to enhance cross-lingual task performance of lowresource languages. We conduct detailed experiments covering existing language-specific neuron identification techniques (such as Language Activation Probability Entropy and activation probability-based thresholding) and neuron-specific LoRA fine-tuning with models like Llama 3.1 and Mistral Nemo. We find that such neuron-specific interventions are insufficient to yield cross-lingual improvements on downstream tasks (XNLI, XQuAD) in lowresource languages. This study highlights the challenges in achieving cross-lingual generalization and provides critical insights for multilingual LLMs.
comment: Accepted (oral) at NAACL 2025 (InsightsNLP)
☆ CausalRivers -- Scaling up benchmarking of causal discovery for real-world time-series ICLR2025
Causal discovery, or identifying causal relationships from observational data, is a notoriously challenging task, with numerous methods proposed to tackle it. Despite this, in-the-wild evaluation of these methods is still lacking, as works frequently rely on synthetic data evaluation and sparse real-world examples under critical theoretical assumptions. Real-world causal structures, however, are often complex, making it hard to decide on a proper causal discovery strategy. To bridge this gap, we introduce CausalRivers, the largest in-the-wild causal discovery benchmarking kit for time-series data to date. CausalRivers features an extensive dataset on river discharge that covers the eastern German territory (666 measurement stations) and the state of Bavaria (494 measurement stations). It spans the years 2019 to 2023 with a 15-minute temporal resolution. Further, we provide additional data from a flood around the Elbe River, as an event with a pronounced distributional shift. Leveraging multiple sources of information and time-series meta-data, we constructed two distinct causal ground truth graphs (Bavaria and eastern Germany). These graphs can be sampled to generate thousands of subgraphs to benchmark causal discovery across diverse and challenging settings. To demonstrate the utility of CausalRivers, we evaluate several causal discovery approaches through a set of experiments to identify areas for improvement. CausalRivers has the potential to facilitate robust evaluations and comparisons of causal discovery methods. Besides this primary purpose, we also expect that this dataset will be relevant for connected areas of research, such as time-series forecasting and anomaly detection. Based on this, we hope to push benchmark-driven method development that fosters advanced techniques for causal discovery, as is the case for many other areas of machine learning.
comment: 10 pages, 8 figures, ICLR2025 main track
☆ LEMMA: Learning from Errors for MatheMatical Advancement in LLMs
Large language models (LLMs) have demonstrated remarkable reasoning capability in solving mathematical problems. However, existing approaches primarily focus on improving the quality of correct training data, e.g., distilling high-quality correct solutions from advanced models, neglecting the value contained in error data, potentially hindering the model's reflective ability. Though some studies attempt to leverage error data, they often involve complex mechanisms, such as Monte Carlo Tree Search (MCTS) to explore error nodes. In this work, we propose to enhance LLMs' reasoning ability by Learning from Errors for Mathematical Advancement (LEMMA). LEMMA constructs data consisting of an incorrect solution with an erroneous step and a reflection connection to a correct solution for fine-tuning. Specifically, we systematically analyze the model-generated error types and introduce an error-type grounded mistake augmentation method to collect diverse and representative errors. Correct solutions are either from fixing the errors or generating a fresh start. Through a model-aware smooth reflection connection, the erroneous solution is transferred to the correct one. By fine-tuning on the constructed dataset, the model is able to self-correct errors autonomously within the generation process without relying on external critique models. Experimental results demonstrate that LEMMA achieves significant performance improvements over other strong baselines.
comment: 9 pages, 6 figures, 4 tables, under review
☆ HCAST: Human-Calibrated Autonomy Software Tasks
To understand and predict the societal impacts of highly autonomous AI systems, we need benchmarks with grounding, i.e., metrics that directly connect AI performance to real-world effects we care about. We present HCAST (Human-Calibrated Autonomy Software Tasks), a benchmark of 189 machine learning engineering, cybersecurity, software engineering, and general reasoning tasks. We collect 563 human baselines (totaling over 1500 hours) from people skilled in these domains, working under identical conditions as AI agents, which lets us estimate that HCAST tasks take humans between one minute and 8+ hours. Measuring the time tasks take for humans provides an intuitive metric for evaluating AI capabilities, helping answer the question "can an agent be trusted to complete a task that would take a human X hours?" We evaluate the success rates of AI agents built on frontier foundation models, and we find that current agents succeed 70-80% of the time on tasks that take humans less than one hour, and less than 20% of the time on tasks that take humans more than 4 hours.
comment: 32 pages, 10 figures, 5 tables
☆ NdLinear Is All You Need for Representation Learning
Many high-impact machine learning tasks involve multi-dimensional data (e.g., images, volumetric medical scans, multivariate time-series). Yet, most neural architectures flatten inputs, discarding critical cross-dimension information. We introduce NdLinear, a novel linear transformation that preserves these structures without extra overhead. By operating separately along each dimension, NdLinear captures dependencies that standard fully connected layers overlook. Extensive experiments across convolutional, recurrent, and transformer-based networks show significant improvements in representational power and parameter efficiency. Crucially, NdLinear serves as a foundational building block for large-scale foundation models by operating on any unimodal or multimodal data in its native form. This removes the need for flattening or modality-specific preprocessing. Ndlinear rethinks core architectural priorities beyond attention, enabling more expressive, context-aware models at scale. We propose NdLinear as a drop-in replacement for standard linear layers -- marking an important step toward next-generation neural architectures.
comment: Code is available at https://github.com/ensemble-core/NdLinear
☆ Align Your Rhythm: Generating Highly Aligned Dance Poses with Gating-Enhanced Rhythm-Aware Feature Representation
Automatically generating natural, diverse and rhythmic human dance movements driven by music is vital for virtual reality and film industries. However, generating dance that naturally follows music remains a challenge, as existing methods lack proper beat alignment and exhibit unnatural motion dynamics. In this paper, we propose Danceba, a novel framework that leverages gating mechanism to enhance rhythm-aware feature representation for music-driven dance generation, which achieves highly aligned dance poses with enhanced rhythmic sensitivity. Specifically, we introduce Phase-Based Rhythm Extraction (PRE) to precisely extract rhythmic information from musical phase data, capitalizing on the intrinsic periodicity and temporal structures of music. Additionally, we propose Temporal-Gated Causal Attention (TGCA) to focus on global rhythmic features, ensuring that dance movements closely follow the musical rhythm. We also introduce Parallel Mamba Motion Modeling (PMMM) architecture to separately model upper and lower body motions along with musical features, thereby improving the naturalness and diversity of generated dance movements. Extensive experiments confirm that Danceba outperforms state-of-the-art methods, achieving significantly better rhythmic alignment and motion diversity. Project page: https://danceba.github.io/ .
comment: 10 pages, 6 figures
☆ Can AI expose tax loopholes? Towards a new generation of legal policy assistants
The legislative process is the backbone of a state built on solid institutions. Yet, due to the complexity of laws -- particularly tax law -- policies may lead to inequality and social tensions. In this study, we introduce a novel prototype system designed to address the issues of tax loopholes and tax avoidance. Our hybrid solution integrates a natural language interface with a domain-specific language tailored for planning. We demonstrate on a case study how tax loopholes and avoidance schemes can be exposed. We conclude that our prototype can help enhance social welfare by systematically identifying and addressing tax gaps stemming from loopholes.
comment: 13 pages, 6 figures
☆ Capturing Individual Human Preferences with Reward Features
Reinforcement learning from human feedback usually models preferences using a reward model that does not distinguish between people. We argue that this is unlikely to be a good design choice in contexts with high potential for disagreement, like in the training of large language models. We propose a method to specialise a reward model to a person or group of people. Our approach builds on the observation that individual preferences can be captured as a linear combination of a set of general reward features. We show how to learn such features and subsequently use them to quickly adapt the reward model to a specific individual, even if their preferences are not reflected in the training data. We present experiments with large language models comparing the proposed architecture with a non-adaptive reward model and also adaptive counterparts, including models that do in-context personalisation. Depending on how much disagreement there is in the training data, our model either significantly outperforms the baselines or matches their performance with a simpler architecture and more stable training.
☆ Efficient Intent-Based Filtering for Multi-Party Conversations Using Knowledge Distillation from LLMs
Large language models (LLMs) have showcased remarkable capabilities in conversational AI, enabling open-domain responses in chat-bots, as well as advanced processing of conversations like summarization, intent classification, and insights generation. However, these models are resource-intensive, demanding substantial memory and computational power. To address this, we propose a cost-effective solution that filters conversational snippets of interest for LLM processing, tailored to the target downstream application, rather than processing every snippet. In this work, we introduce an innovative approach that leverages knowledge distillation from LLMs to develop an intent-based filter for multi-party conversations, optimized for compute power constrained environments. Our method combines different strategies to create a diverse multi-party conversational dataset, that is annotated with the target intents and is then used to fine-tune the MobileBERT model for multi-label intent classification. This model achieves a balance between efficiency and performance, effectively filtering conversation snippets based on their intents. By passing only the relevant snippets to the LLM for further processing, our approach significantly reduces overall operational costs depending on the intents and the data distribution as demonstrated in our experiments.
☆ CVE-Bench: A Benchmark for AI Agents' Ability to Exploit Real-World Web Application Vulnerabilities
Large language model (LLM) agents are increasingly capable of autonomously conducting cyberattacks, posing significant threats to existing applications. This growing risk highlights the urgent need for a real-world benchmark to evaluate the ability of LLM agents to exploit web application vulnerabilities. However, existing benchmarks fall short as they are limited to abstracted Capture the Flag competitions or lack comprehensive coverage. Building a benchmark for real-world vulnerabilities involves both specialized expertise to reproduce exploits and a systematic approach to evaluating unpredictable threats. To address this challenge, we introduce CVE-Bench, a real-world cybersecurity benchmark based on critical-severity Common Vulnerabilities and Exposures. In CVE-Bench, we design a sandbox framework that enables LLM agents to exploit vulnerable web applications in scenarios that mimic real-world conditions, while also providing effective evaluation of their exploits. Our evaluation shows that the state-of-the-art agent framework can resolve up to 13% of vulnerabilities.
comment: 15 pages, 4 figures, 5 tables
☆ LLM+MAP: Bimanual Robot Task Planning using Large Language Models and Planning Domain Definition Language
Bimanual robotic manipulation provides significant versatility, but also presents an inherent challenge due to the complexity involved in the spatial and temporal coordination between two hands. Existing works predominantly focus on attaining human-level manipulation skills for robotic hands, yet little attention has been paid to task planning on long-horizon timescales. With their outstanding in-context learning and zero-shot generation abilities, Large Language Models (LLMs) have been applied and grounded in diverse robotic embodiments to facilitate task planning. However, LLMs still suffer from errors in long-horizon reasoning and from hallucinations in complex robotic tasks, lacking a guarantee of logical correctness when generating the plan. Previous works, such as LLM+P, extended LLMs with symbolic planners. However, none have been successfully applied to bimanual robots. New challenges inevitably arise in bimanual manipulation, necessitating not only effective task decomposition but also efficient task allocation. To address these challenges, this paper introduces LLM+MAP, a bimanual planning framework that integrates LLM reasoning and multi-agent planning, automating effective and efficient bimanual task planning. We conduct simulated experiments on various long-horizon manipulation tasks of differing complexity. Our method is built using GPT-4o as the backend, and we compare its performance against plans generated directly by LLMs, including GPT-4o, V3 and also recent strong reasoning models o1 and R1. By analyzing metrics such as planning time, success rate, group debits, and planning-step reduction rate, we demonstrate the superior performance of LLM+MAP, while also providing insights into robotic reasoning. Code is available at https://github.com/Kchu/LLM-MAP.
comment: Code and video are available at https://github.com/Kchu/LLM-MAP
☆ Preference-Guided Diffusion for Multi-Objective Offline Optimization
Offline multi-objective optimization aims to identify Pareto-optimal solutions given a dataset of designs and their objective values. In this work, we propose a preference-guided diffusion model that generates Pareto-optimal designs by leveraging a classifier-based guidance mechanism. Our guidance classifier is a preference model trained to predict the probability that one design dominates another, directing the diffusion model toward optimal regions of the design space. Crucially, this preference model generalizes beyond the training distribution, enabling the discovery of Pareto-optimal solutions outside the observed dataset. We introduce a novel diversity-aware preference guidance, augmenting Pareto dominance preference with diversity criteria. This ensures that generated solutions are optimal and well-distributed across the objective space, a capability absent in prior generative methods for offline multi-objective optimization. We evaluate our approach on various continuous offline multi-objective optimization tasks and find that it consistently outperforms other inverse/generative approaches while remaining competitive with forward/surrogate-based optimization methods. Our results highlight the effectiveness of classifier-guided diffusion models in generating diverse and high-quality solutions that approximate the Pareto front well.
☆ Breaking the Symmetries of Indistinguishable Objects
Indistinguishable objects often occur when modelling problems in constraint programming, as well as in other related paradigms. They occur when objects can be viewed as being drawn from a set of unlabelled objects, and the only operation allowed on them is equality testing. For example, the golfers in the social golfer problem are indistinguishable. If we do label the golfers, then any relabelling of the golfers in one solution gives another valid solution. Therefore, we can regard the symmetric group of size $n$ as acting on a set of $n$ indistinguishable objects. In this paper, we show how we can break the symmetries resulting from indistinguishable objects. We show how symmetries on indistinguishable objects can be defined properly in complex types, for example in a matrix indexed by indistinguishable objects. We then show how the resulting symmetries can be broken correctly. In Essence, a high-level modelling language, indistinguishable objects are encapsulated in "unnamed types". We provide an implementation of complete symmetry breaking for unnamed types in Essence.
☆ KL3M Tokenizers: A Family of Domain-Specific and Character-Level Tokenizers for Legal, Financial, and Preprocessing Applications
We present the KL3M tokenizers, a family of specialized tokenizers for legal, financial, and governmental text. Despite established work on tokenization, specialized tokenizers for professional domains remain understudied. Our paper offers two main contributions to this area. First, we introduce domain-specific BPE tokenizers for legal, financial, and governmental text. Our kl3m-004-128k-cased tokenizer uses 9-17% fewer tokens than GPT-4o and Llama3 for domain-specific documents, despite having a smaller vocabulary. For specialized terminology, our cased tokenizer is even more efficient, using up to 83% fewer tokens for legal terms and 39% fewer tokens for financial terms. Second, we develop character-level BPE tokenizers (4K, 8K, and 16K vocabulary sizes) for text correction tasks like OCR post-processing. These tokenizers keep consistent token boundaries between error-containing and correct text, making it easier for models to learn correction patterns. These tokenizers help professional applications by fitting more text in context windows, reducing computational needs, and preserving the meaning of domain-specific terms. Our analysis shows these efficiency gains directly benefit the processing of long legal and financial documents. We release all tokenizers and code through GitHub and Hugging Face to support further research in specialized tokenization.
comment: 12 pages, 7 tables, 3 figures; Source code available at https://github.com/alea-institute/kl3m-tokenizer-paper
☆ SafeMERGE: Preserving Safety Alignment in Fine-Tuned Large Language Models via Selective Layer-Wise Model Merging
Fine-tuning large language models (LLMs) on downstream tasks can inadvertently erode their safety alignment, even for benign fine-tuning datasets. We address this challenge by proposing SafeMERGE, a post-fine-tuning framework that preserves safety while maintaining task utility. It achieves this by selectively merging fine-tuned and safety-aligned model layers only when those deviate from safe behavior, measured by a cosine similarity criterion. We evaluate SafeMERGE against other fine-tuning- and post-fine-tuning-stage approaches for Llama-2-7B-Chat and Qwen-2-7B-Instruct models on GSM8K and PubMedQA tasks while exploring different merging strategies. We find that SafeMERGE consistently reduces harmful outputs compared to other baselines without significantly sacrificing performance, sometimes even enhancing it. The results suggest that our selective, subspace-guided, and per-layer merging method provides an effective safeguard against the inadvertent loss of safety in fine-tuned LLMs while outperforming simpler post-fine-tuning-stage defenses.
☆ Strong Baseline: Multi-UAV Tracking via YOLOv12 with BoT-SORT-ReID
Detecting and tracking multiple unmanned aerial vehicles (UAVs) in thermal infrared video is inherently challenging due to low contrast, environmental noise, and small target sizes. This paper provides a straightforward approach to address multi-UAV tracking in thermal infrared video, leveraging recent advances in detection and tracking. Instead of relying on the YOLOv5 with the DeepSORT pipeline, we present a tracking framework built on YOLOv12 and BoT-SORT, enhanced with tailored training and inference strategies. We evaluate our approach following the metrics from the 4th Anti-UAV Challenge and demonstrate competitive performance. Notably, we achieve strong results without using contrast enhancement or temporal information fusion to enrich UAV features, highlighting our approach as a "Strong Baseline" for the multi-UAV tracking task. We provide implementation details, in-depth experimental analysis, and a discussion of potential improvements. The code is available at https://github.com/wish44165/YOLOv12-BoT-SORT-ReID .
comment: 10 pages, 5 figures, 5 tables
☆ FactSelfCheck: Fact-Level Black-Box Hallucination Detection for LLMs
Large Language Models (LLMs) frequently generate hallucinated content, posing significant challenges for applications where factuality is crucial. While existing hallucination detection methods typically operate at the sentence level or passage level, we propose FactSelfCheck, a novel black-box sampling-based method that enables fine-grained fact-level detection. Our approach represents text as knowledge graphs consisting of facts in the form of triples. Through analyzing factual consistency across multiple LLM responses, we compute fine-grained hallucination scores without requiring external resources or training data. Our evaluation demonstrates that FactSelfCheck performs competitively with leading sampling-based methods while providing more detailed insights. Most notably, our fact-level approach significantly improves hallucination correction, achieving a 35% increase in factual content compared to the baseline, while sentence-level SelfCheckGPT yields only an 8% improvement. The granular nature of our detection enables more precise identification and correction of hallucinated content.
comment: Preprint
☆ Neuro-Symbolic Scene Graph Conditioning for Synthetic Image Dataset Generation
As machine learning models increase in scale and complexity, obtaining sufficient training data has become a critical bottleneck due to acquisition costs, privacy constraints, and data scarcity in specialised domains. While synthetic data generation has emerged as a promising alternative, a notable performance gap remains compared to models trained on real data, particularly as task complexity grows. Concurrently, Neuro-Symbolic methods, which combine neural networks' learning strengths with symbolic reasoning's structured representations, have demonstrated significant potential across various cognitive tasks. This paper explores the utility of Neuro-Symbolic conditioning for synthetic image dataset generation, focusing specifically on improving the performance of Scene Graph Generation models. The research investigates whether structured symbolic representations in the form of scene graphs can enhance synthetic data quality through explicit encoding of relational constraints. The results demonstrate that Neuro-Symbolic conditioning yields significant improvements of up to +2.59% in standard Recall metrics and +2.83% in No Graph Constraint Recall metrics when used for dataset augmentation. These findings establish that merging Neuro-Symbolic and generative approaches produces synthetic data with complementary structural information that enhances model performance when combined with real data, providing a novel approach to overcome data scarcity limitations even for complex visual reasoning tasks.
☆ Automating Adjudication of Cardiovascular Events Using Large Language Models
Cardiovascular events, such as heart attacks and strokes, remain a leading cause of mortality globally, necessitating meticulous monitoring and adjudication in clinical trials. This process, traditionally performed manually by clinical experts, is time-consuming, resource-intensive, and prone to inter-reviewer variability, potentially introducing bias and hindering trial progress. This study addresses these critical limitations by presenting a novel framework for automating the adjudication of cardiovascular events in clinical trials using Large Language Models (LLMs). We developed a two-stage approach: first, employing an LLM-based pipeline for event information extraction from unstructured clinical data and second, using an LLM-based adjudication process guided by a Tree of Thoughts approach and clinical endpoint committee (CEC) guidelines. Using cardiovascular event-specific clinical trial data, the framework achieved an F1-score of 0.82 for event extraction and an accuracy of 0.68 for adjudication. Furthermore, we introduce the CLEART score, a novel, automated metric specifically designed for evaluating the quality of AI-generated clinical reasoning in adjudicating cardiovascular events. This approach demonstrates significant potential for substantially reducing adjudication time and costs while maintaining high-quality, consistent, and auditable outcomes in clinical trials. The reduced variability and enhanced standardization also allow for faster identification and mitigation of risks associated with cardiovascular therapies.
☆ PP-DocLayout: A Unified Document Layout Detection Model to Accelerate Large-Scale Data Construction
Document layout analysis is a critical preprocessing step in document intelligence, enabling the detection and localization of structural elements such as titles, text blocks, tables, and formulas. Despite its importance, existing layout detection models face significant challenges in generalizing across diverse document types, handling complex layouts, and achieving real-time performance for large-scale data processing. To address these limitations, we present PP-DocLayout, which achieves high precision and efficiency in recognizing 23 types of layout regions across diverse document formats. To meet different needs, we offer three models of varying scales. PP-DocLayout-L is a high-precision model based on the RT-DETR-L detector, achieving 90.4% mAP@0.5 and an end-to-end inference time of 13.4 ms per page on a T4 GPU. PP-DocLayout-M is a balanced model, offering 75.2% mAP@0.5 with an inference time of 12.7 ms per page on a T4 GPU. PP-DocLayout-S is a high-efficiency model designed for resource-constrained environments and real-time applications, with an inference time of 8.1 ms per page on a T4 GPU and 14.5 ms on a CPU. This work not only advances the state of the art in document layout analysis but also provides a robust solution for constructing high-quality training data, enabling advancements in document intelligence and multimodal AI systems. Code and models are available at https://github.com/PaddlePaddle/PaddleX .
comment: Github Repo: https://github.com/PaddlePaddle/PaddleX
☆ TreeSynth: Synthesizing Diverse Data from Scratch via Tree-Guided Subspace Partitioning
Model customization requires high-quality and diverse datasets, but acquiring such data remains challenging and costly. Although large language models (LLMs) can synthesize training data, current approaches are constrained by limited seed data, model bias and insufficient control over the generation process, resulting in limited diversity and biased distribution with the increase of data scales. To tackle this challenge, we present TreeSynth, a tree-guided subspace-based data synthesis framework that recursively partitions the entire data space into hierar-chical subspaces, enabling comprehensive and diverse scaling of data synthesis. Briefly, given a task-specific description, we construct a data space partitioning tree by iteratively executing criteria determination and subspace coverage steps. This hierarchically divides the whole space (i.e., root node) into mutually exclusive and complementary atomic subspaces (i.e., leaf nodes). By collecting synthesized data according to the attributes of each leaf node, we obtain a diverse dataset that fully covers the data space. Empirically, our extensive experiments demonstrate that TreeSynth surpasses both human-designed datasets and the state-of-the-art data synthesis baselines, achieving maximum improvements of 45.2% in data diversity and 17.6% in downstream task performance across various models and tasks. Hopefully, TreeSynth provides a scalable solution to synthesize diverse and comprehensive datasets from scratch without human intervention.
☆ D2Fusion: Dual-domain Fusion with Feature Superposition for Deepfake Detection
Deepfake detection is crucial for curbing the harm it causes to society. However, current Deepfake detection methods fail to thoroughly explore artifact information across different domains due to insufficient intrinsic interactions. These interactions refer to the fusion and coordination after feature extraction processes across different domains, which are crucial for recognizing complex forgery clues. Focusing on more generalized Deepfake detection, in this work, we introduce a novel bi-directional attention module to capture the local positional information of artifact clues from the spatial domain. This enables accurate artifact localization, thus addressing the coarse processing with artifact features. To further address the limitation that the proposed bi-directional attention module may not well capture global subtle forgery information in the artifact feature (e.g., textures or edges), we employ a fine-grained frequency attention module in the frequency domain. By doing so, we can obtain high-frequency information in the fine-grained features, which contains the global and subtle forgery information. Although these features from the diverse domains can be effectively and independently improved, fusing them directly does not effectively improve the detection performance. Therefore, we propose a feature superposition strategy that complements information from spatial and frequency domains. This strategy turns the feature components into the form of wave-like tokens, which are updated based on their phase, such that the distinctions between authentic and artifact features can be amplified. Our method demonstrates significant improvements over state-of-the-art (SOTA) methods on five public Deepfake datasets in capturing abnormalities across different manipulated operations and real-life.
☆ LLMs Love Python: A Study of LLMs' Bias for Programming Languages and Libraries
Programming language and library choices are crucial to software reliability and security. Poor or inconsistent choices can lead to increased technical debt, security vulnerabilities, and even catastrophic failures in safety-critical systems. As Large Language Models (LLMs) play an increasing role in code generation, it is essential to understand how they make these decisions. However, little is known about their preferences when selecting programming languages and libraries for different coding tasks. To fill this gap, this study provides the first in-depth investigation into LLM preferences for programming languages and libraries used when generating code. We assess the preferences of eight diverse LLMs by prompting them to complete various coding tasks, including widely-studied benchmarks and the more practical task of generating the initial structural code for new projects (a crucial step that often determines a project's language or library choices). Our findings reveal that LLMs heavily favour Python when solving language-agnostic problems, using it in 90%-97% of cases for benchmark tasks. Even when generating initial project code where Python is not a suitable language, it remains the most-used language in 58% of instances. Moreover, LLMs contradict their own language recommendations in 83% of project initialisation tasks, raising concerns about their reliability in guiding language selection. Similar biases toward well-established libraries further create serious discoverability challenges for newer open-source projects. These results highlight the need to improve LLMs' adaptability to diverse programming contexts and to develop mechanisms for mitigating programming language and library bias.
comment: 12 pages, 1 figure
☆ Temporal-Guided Spiking Neural Networks for Event-Based Human Action Recognition
This paper explores the promising interplay between spiking neural networks (SNNs) and event-based cameras for privacy-preserving human action recognition (HAR). The unique feature of event cameras in capturing only the outlines of motion, combined with SNNs' proficiency in processing spatiotemporal data through spikes, establishes a highly synergistic compatibility for event-based HAR. Previous studies, however, have been limited by SNNs' ability to process long-term temporal information, essential for precise HAR. In this paper, we introduce two novel frameworks to address this: temporal segment-based SNN (\textit{TS-SNN}) and 3D convolutional SNN (\textit{3D-SNN}). The \textit{TS-SNN} extracts long-term temporal information by dividing actions into shorter segments, while the \textit{3D-SNN} replaces 2D spatial elements with 3D components to facilitate the transmission of temporal information. To promote further research in event-based HAR, we create a dataset, \textit{FallingDetection-CeleX}, collected using the high-resolution CeleX-V event camera $(1280 \times 800)$, comprising 7 distinct actions. Extensive experimental results show that our proposed frameworks surpass state-of-the-art SNN methods on our newly collected dataset and three other neuromorphic datasets, showcasing their effectiveness in handling long-range temporal information for event-based HAR.
☆ The CASTLE 2024 Dataset: Advancing the Art of Multimodal Understanding
Egocentric video has seen increased interest in recent years, as it is used in a range of areas. However, most existing datasets are limited to a single perspective. In this paper, we present the CASTLE 2024 dataset, a multimodal collection containing ego- and exo-centric (i.e., first- and third-person perspective) video and audio from 15 time-aligned sources, as well as other sensor streams and auxiliary data. The dataset was recorded by volunteer participants over four days in a fixed location and includes the point of view of 10 participants, with an additional 5 fixed cameras providing an exocentric perspective. The entire dataset contains over 600 hours of UHD video recorded at 50 frames per second. In contrast to other datasets, CASTLE 2024 does not contain any partial censoring, such as blurred faces or distorted audio. The dataset is available via https://castle-dataset.github.io/.
comment: 7 pages, 6 figures, dataset available via https://castle-dataset.github.io/
☆ FFaceNeRF: Few-shot Face Editing in Neural Radiance Fields CVPR2025
Recent 3D face editing methods using masks have produced high-quality edited images by leveraging Neural Radiance Fields (NeRF). Despite their impressive performance, existing methods often provide limited user control due to the use of pre-trained segmentation masks. To utilize masks with a desired layout, an extensive training dataset is required, which is challenging to gather. We present FFaceNeRF, a NeRF-based face editing technique that can overcome the challenge of limited user control due to the use of fixed mask layouts. Our method employs a geometry adapter with feature injection, allowing for effective manipulation of geometry attributes. Additionally, we adopt latent mixing for tri-plane augmentation, which enables training with a few samples. This facilitates rapid model adaptation to desired mask layouts, crucial for applications in fields like personalized medical imaging or creative face editing. Our comparative evaluations demonstrate that FFaceNeRF surpasses existing mask based face editing methods in terms of flexibility, control, and generated image quality, paving the way for future advancements in customized and high-fidelity 3D face editing. The code is available on the {\href{https://kwanyun.github.io/FFaceNeRF_page/}{project-page}}.
comment: CVPR2025, 11 pages, 14 figures
☆ Does a Rising Tide Lift All Boats? Bias Mitigation for AI-based CMR Segmentation
Artificial intelligence (AI) is increasingly being used for medical imaging tasks. However, there can be biases in the resulting models, particularly when they were trained using imbalanced training datasets. One such example has been the strong race bias effect in cardiac magnetic resonance (CMR) image segmentation models. Although this phenomenon has been reported in a number of publications, little is known about the effectiveness of bias mitigation algorithms in this domain. We aim to investigate the impact of common bias mitigation methods to address bias between Black and White subjects in AI-based CMR segmentation models. Specifically, we use oversampling, importance reweighing and Group DRO as well as combinations of these techniques to mitigate the race bias. Furthermore, motivated by recent findings on the root causes of AI-based CMR segmentation bias, we evaluate the same methods using models trained and evaluated on cropped CMR images. We find that bias can be mitigated using oversampling, significantly improving performance for the underrepresented Black subjects whilst not significantly reducing the majority White subjects' performance. Group DRO also improves performance for Black subjects but not significantly, while reweighing decreases performance for Black subjects. Using a combination of oversampling and Group DRO also improves performance for Black subjects but not significantly. Using cropped images increases performance for both races and reduces the bias, whilst adding oversampling as a bias mitigation technique with cropped images reduces the bias further.
☆ Deterministic AI Agent Personality Expression through Standard Psychological Diagnostics
Artificial intelligence (AI) systems powered by large language models have become increasingly prevalent in modern society, enabling a wide range of applications through natural language interaction. As AI agents proliferate in our daily lives, their generic and uniform expressiveness presents a significant limitation to their appeal and adoption. Personality expression represents a key prerequisite for creating more human-like and distinctive AI systems. We show that AI models can express deterministic and consistent personalities when instructed using established psychological frameworks, with varying degrees of accuracy depending on model capabilities. We find that more advanced models like GPT-4o and o1 demonstrate the highest accuracy in expressing specified personalities across both Big Five and Myers-Briggs assessments, and further analysis suggests that personality expression emerges from a combination of intelligence and reasoning capabilities. Our results reveal that personality expression operates through holistic reasoning rather than question-by-question optimization, with response-scale metrics showing higher variance than test-scale metrics. Furthermore, we find that model fine-tuning affects communication style independently of personality expression accuracy. These findings establish a foundation for creating AI agents with diverse and consistent personalities, which could significantly enhance human-AI interaction across applications from education to healthcare, while additionally enabling a broader range of more unique AI agents. The ability to quantitatively assess and implement personality expression in AI systems opens new avenues for research into more relatable, trustworthy, and ethically designed AI.
comment: 25 pages, 8 figures, 4 tables; appeared in ADI (March 2025)
☆ A Thorough Assessment of the Non-IID Data Impact in Federated Learning
Federated learning (FL) allows collaborative machine learning (ML) model training among decentralized clients' information, ensuring data privacy. The decentralized nature of FL deals with non-independent and identically distributed (non-IID) data. This open problem has notable consequences, such as decreased model performance and more significant convergence times. Despite its importance, experimental studies systematically addressing all types of data heterogeneity (a.k.a. non-IIDness) remain scarce. We aim to fill this gap by assessing and quantifying the non-IID effect through a thorough empirical analysis. We use the Hellinger Distance (HD) to measure differences in distribution among clients. Our study benchmarks four state-of-the-art strategies for handling non-IID data, including label, feature, quantity, and spatiotemporal skewness, under realistic and controlled conditions. This is the first comprehensive analysis of the spatiotemporal skew effect in FL. Our findings highlight the significant impact of label and spatiotemporal skew non-IID types on FL model performance, with notable performance drops occurring at specific HD thresholds. Additionally, the FL performance is heavily affected mainly when the non-IIDness is extreme. Thus, we provide recommendations for FL research to tackle data heterogeneity effectively. Our work represents the most extensive examination of non-IIDness in FL, offering a robust foundation for future research.
☆ PVChat: Personalized Video Chat with One-Shot Learning
Video large language models (ViLLMs) excel in general video understanding, e.g., recognizing activities like talking and eating, but struggle with identity-aware comprehension, such as "Wilson is receiving chemotherapy" or "Tom is discussing with Sarah", limiting their applicability in smart healthcare and smart home environments. To address this limitation, we propose a one-shot learning framework PVChat, the first personalized ViLLM that enables subject-aware question answering (QA) from a single video for each subject. Our approach optimizes a Mixture-of-Heads (MoH) enhanced ViLLM on a synthetically augmented video-QA dataset, leveraging a progressive image-to-video learning strategy. Specifically, we introduce an automated augmentation pipeline that synthesizes identity-preserving positive samples and retrieves hard negatives from existing video corpora, generating a diverse training dataset with four QA types: existence, appearance, action, and location inquiries. To enhance subject-specific learning, we propose a ReLU Routing MoH attention mechanism, alongside two novel objectives: (1) Smooth Proximity Regularization for progressive learning through exponential distance scaling and (2) Head Activation Enhancement for balanced attention routing. Finally, we adopt a two-stage training strategy, transitioning from image pre-training to video fine-tuning, enabling a gradual learning process from static attributes to dynamic representations. We evaluate PVChat on diverse datasets covering medical scenarios, TV series, anime, and real-world footage, demonstrating its superiority in personalized feature understanding after learning from a single video, compared to state-of-the-art ViLLMs.
☆ Replay4NCL: An Efficient Memory Replay-based Methodology for Neuromorphic Continual Learning in Embedded AI Systems
Neuromorphic Continual Learning (NCL) paradigm leverages Spiking Neural Networks (SNNs) to enable continual learning (CL) capabilities for AI systems to adapt to dynamically changing environments. Currently, the state-of-the-art employ a memory replay-based method to maintain the old knowledge. However, this technique relies on long timesteps and compression-decompression steps, thereby incurring significant latency and energy overheads, which are not suitable for tightly-constrained embedded AI systems (e.g., mobile agents/robotics). To address this, we propose Replay4NCL, a novel efficient memory replay-based methodology for enabling NCL in embedded AI systems. Specifically, Replay4NCL compresses the latent data (old knowledge), then replays them during the NCL training phase with small timesteps, to minimize the processing latency and energy consumption. To compensate the information loss from reduced spikes, we adjust the neuron threshold potential and learning rate settings. Experimental results on the class-incremental scenario with the Spiking Heidelberg Digits (SHD) dataset show that Replay4NCL can preserve old knowledge with Top-1 accuracy of 90.43% compared to 86.22% from the state-of-the-art, while effectively learning new tasks, achieving 4.88x latency speed-up, 20% latent memory saving, and 36.43% energy saving. These results highlight the potential of our Replay4NCL methodology to further advances NCL capabilities for embedded AI systems.
comment: Accepted at the 62th Design Automation Conference (DAC) 2025, June 2025, San Francisco, CA, USA
☆ Data-Driven Optimization of EV Charging Station Placement Using Causal Discovery
This paper addresses the critical challenge of optimizing electric vehicle charging station placement through a novel data-driven methodology employing causal discovery techniques. While traditional approaches prioritize economic factors or power grid constraints, they often neglect empirical charging patterns that ultimately determine station utilization. We analyze extensive charging data from Palo Alto and Boulder (337,344 events across 100 stations) to uncover latent relationships between station characteristics and utilization. Applying structural learning algorithms (NOTEARS and DAGMA) to this data reveals that charging demand is primarily determined by three factors: proximity to amenities, EV registration density, and adjacency to high-traffic routes. These findings, consistent across multiple algorithms and urban contexts, challenge conventional infrastructure distribution strategies. We develop an optimization framework that translates these insights into actionable placement recommendations, identifying locations likely to experience high utilization based on the discovered dependency structures. The resulting site selection model prioritizes strategic clustering in high-amenity areas with substantial EV populations rather than uniform spatial distribution. Our approach contributes a framework that integrates empirical charging behavior into infrastructure planning, potentially enhancing both station utilization and user convenience. By focusing on data-driven insights instead of theoretical distribution models, we provide a more effective strategy for expanding charging networks that can adjust to various stages of EV market development.
comment: Under review of IEEE CASE 2025; This is also the master thesis project from Julius supervised by Dr. Ziyue Li
☆ HAPI: A Model for Learning Robot Facial Expressions from Human Preferences
Automatic robotic facial expression generation is crucial for human-robot interaction, as handcrafted methods based on fixed joint configurations often yield rigid and unnatural behaviors. Although recent automated techniques reduce the need for manual tuning, they tend to fall short by not adequately bridging the gap between human preferences and model predictions-resulting in a deficiency of nuanced and realistic expressions due to limited degrees of freedom and insufficient perceptual integration. In this work, we propose a novel learning-to-rank framework that leverages human feedback to address this discrepancy and enhanced the expressiveness of robotic faces. Specifically, we conduct pairwise comparison annotations to collect human preference data and develop the Human Affective Pairwise Impressions (HAPI) model, a Siamese RankNet-based approach that refines expression evaluation. Results obtained via Bayesian Optimization and online expression survey on a 35-DOF android platform demonstrate that our approach produces significantly more realistic and socially resonant expressions of Anger, Happiness, and Surprise than those generated by baseline and expert-designed methods. This confirms that our framework effectively bridges the gap between human preferences and model predictions while robustly aligning robotic expression generation with human affective responses.
☆ Summarization Metrics for Spanish and Basque: Do Automatic Scores and LLM-Judges Correlate with Humans?
Studies on evaluation metrics and LLM-as-a-Judge models for automatic text summarization have largely been focused on English, limiting our understanding of their effectiveness in other languages. Through our new dataset BASSE (BAsque and Spanish Summarization Evaluation), we address this situation by collecting human judgments on 2,040 abstractive summaries in Basque and Spanish, generated either manually or by five LLMs with four different prompts. For each summary, annotators evaluated five criteria on a 5-point Likert scale: coherence, consistency, fluency, relevance, and 5W1H. We use these data to reevaluate traditional automatic metrics used for evaluating summaries, as well as several LLM-as-a-Judge models that show strong performance on this task in English. Our results show that currently proprietary judge LLMs have the highest correlation with human judgments, followed by criteria-specific automatic metrics, while open-sourced judge LLMs perform poorly. We release BASSE and our code publicly, along with the first large-scale Basque summarization dataset containing 22,525 news articles with their subheads.
☆ Enhanced Smart Contract Reputability Analysis using Multimodal Data Fusion on Ethereum
The evaluation of smart contract reputability is essential to foster trust in decentralized ecosystems. However, existing methods that rely solely on static code analysis or transactional data, offer limited insight into evolving trustworthiness. We propose a multimodal data fusion framework that integrates static code features with transactional data to enhance reputability prediction. Our framework initially focuses on static code analysis, utilizing GAN-augmented opcode embeddings to address class imbalance, achieving 97.67% accuracy and a recall of 0.942 in detecting illicit contracts, surpassing traditional oversampling methods. This forms the crux of a reputability-centric fusion strategy, where combining static and transactional data improves recall by 7.25% over single-source models, demonstrating robust performance across validation sets. By providing a holistic view of smart contract behaviour, our approach enhances the model's ability to assess reputability, identify fraudulent activities, and predict anomalous patterns. These capabilities contribute to more accurate reputability assessments, proactive risk mitigation, and enhanced blockchain security.
☆ An Attentive Representative Sample Selection Strategy Combined with Balanced Batch Training for Skin Lesion Segmentation
An often overlooked problem in medical image segmentation research is the effective selection of training subsets to annotate from a complete set of unlabelled data. Many studies select their training sets at random, which may lead to suboptimal model performance, especially in the minimal supervision setting where each training image has a profound effect on performance outcomes. This work aims to address this issue. We use prototypical contrasting learning and clustering to extract representative and diverse samples for annotation. We improve upon prior works with a bespoke cluster-based image selection process. Additionally, we introduce the concept of unsupervised balanced batch dataloading to medical image segmentation, which aims to improve model learning with minimally annotated data. We evaluated our method on a public skin lesion dataset (ISIC 2018) and compared it to another state-of-the-art data sampling method. Our method achieved superior performance in a low annotation budget scenario.
comment: Accepted to ISBI 2025
☆ Exploring the Efficacy of Partial Denoising Using Bit Plane Slicing for Enhanced Fracture Identification: A Comparative Study of Deep Learning-Based Approaches and Handcrafted Feature Extraction Techniques
Computer vision has transformed medical diagnosis, treatment, and research through advanced image processing and machine learning techniques. Fracture classification, a critical area in healthcare, has greatly benefited from these advancements, yet accurate detection is challenged by complex patterns and image noise. Bit plane slicing enhances medical images by reducing noise interference and extracting informative features. This research explores partial denoising techniques to provide practical solutions for improved fracture analysis, ultimately enhancing patient care. The study explores deep learning model DenseNet and handcrafted feature extraction. Decision Tree and Random Forest, were employed to train and evaluate distinct image representations. These include the original image, the concatenation of the four bit planes from the LSB as well as MSB, the fully denoised image, and an image consisting of 6 bit planes from MSB and 2 denoised bit planes from LSB. The purpose of forming these diverse image representations is to analyze SNR as well as classification accuracy and identify the bit planes that contain the most informative features. Moreover, the study delves into the significance of partial denoising techniques in preserving crucial features, leading to improvements in classification results. Notably, this study shows that employing the Random Forest classifier, the partially denoised image representation exhibited a testing accuracy of 95.61% surpassing the performance of other image representations. The outcomes of this research provide valuable insights into the development of efficient preprocessing, feature extraction and classification approaches for fracture identification. By enhancing diagnostic accuracy, these advancements hold the potential to positively impact patient care and overall medical outcomes.
☆ A Guide to Bayesian Networks Software Packages for Structure and Parameter Learning -- 2025 Edition
A representation of the cause-effect mechanism is needed to enable artificial intelligence to represent how the world works. Bayesian Networks (BNs) have proven to be an effective and versatile tool for this task. BNs require constructing a structure of dependencies among variables and learning the parameters that govern these relationships. These tasks, referred to as structural learning and parameter learning, are actively investigated by the research community, with several algorithms proposed and no single method having established itself as standard. A wide range of software, tools, and packages have been developed for BNs analysis and made available to academic researchers and industry practitioners. As a consequence of having no one-size-fits-all solution, moving the first practical steps and getting oriented into this field is proving to be challenging to outsiders and beginners. In this paper, we review the most relevant tools and software for BNs structural and parameter learning to date, providing our subjective recommendations directed to an audience of beginners. In addition, we provide an extensive easy-to-consult overview table summarizing all software packages and their main features. By improving the reader understanding of which available software might best suit their needs, we improve accessibility to the field and make it easier for beginners to take their first step into it.
comment: 11 pages, 1 figure
☆ Symbolic Audio Classification via Modal Decision Tree Learning
The range of potential applications of acoustic analysis is wide. Classification of sounds, in particular, is a typical machine learning task that received a lot of attention in recent years. The most common approaches to sound classification are sub-symbolic, typically based on neural networks, and result in black-box models with high performances but very low transparency. In this work, we consider several audio tasks, namely, age and gender recognition, emotion classification, and respiratory disease diagnosis, and we approach them with a symbolic technique, that is, (modal) decision tree learning. We prove that such tasks can be solved using the same symbolic pipeline, that allows to extract simple rules with very high accuracy and low complexity. In principle, all such tasks could be associated to an autonomous conversation system, which could be useful in different contexts, such as an automatic reservation agent for an hospital or a clinic.
☆ Developing Critical Thinking in Second Language Learners: Exploring Generative AI like ChatGPT as a Tool for Argumentative Essay Writing
This study employs the Paul-Elder Critical Thinking Model and Tan's argumentative writing framework to create a structured methodology. This methodology, ChatGPT Guideline for Critical Argumentative Writing (CGCAW) framework, integrates the models with ChatGPT's capabilities to guide L2 learners in utilizing ChatGPT to enhance their critical thinking skills. A quantitative experiment was conducted with 10 participants from a state university, divided into experimental and control groups. The experimental group utilized the CGCAW framework, while the control group used ChatGPT without specific guidelines. Participants wrote an argumentative essay within a 40-minute timeframe, and essays were evaluated by three assessors: ChatGPT, Grammarly, and a course instructor. Results indicated that the experimental group showed improvements in clarity, logical coherence, and use of evidence, demonstrating ChatGPT's potential to enhance specific aspects of argumentative writing. However, the control group performed better in overall language mechanics and articulation of main arguments, indicating areas where the CGCAW framework could be further refined. This study highlights the need for further research to optimize the use of AI tools like ChatGPT in L2 learning environments to enhance critical thinking and writing skills.
comment: 12 pages, 3 figures. Uses Paul-Elder Critical Thinking Model and Tan's argumentative writing framework. Includes an experimental study with 10 participants
☆ Targetless 6DoF Calibration of LiDAR and 2D Scanning Radar Based on Cylindrical Occupancy
Owing to the capability for reliable and all-weather long-range sensing, the fusion of LiDAR and Radar has been widely applied to autonomous vehicles for robust perception. In practical operation, well manually calibrated extrinsic parameters, which are crucial for the fusion of multi-modal sensors, may drift due to the vibration. To address this issue, we present a novel targetless calibration approach, termed LiRaCo, for the extrinsic 6DoF calibration of LiDAR and Radar sensors. Although both types of sensors can obtain geometric information, bridging the geometric correspondences between multi-modal data without any clues of explicit artificial markers is nontrivial, mainly due to the low vertical resolution of scanning Radar. To achieve the targetless calibration, LiRaCo leverages a spatial occupancy consistency between LiDAR point clouds and Radar scans in a common cylindrical representation, considering the increasing data sparsity with distance for both sensors. Specifically, LiRaCo expands the valid Radar scanned pixels into 3D occupancy grids to constrain LiDAR point clouds based on spatial consistency. Consequently, a cost function involving extrinsic calibration parameters is formulated based on the spatial overlap of 3D grids and LiDAR points. Extrinsic parameters are finally estimated by optimizing the cost function. Comprehensive quantitative and qualitative experiments on two real outdoor datasets with different LiDAR sensors demonstrate the feasibility and accuracy of the proposed method. The source code will be publicly available.
☆ Data to Decisions: A Computational Framework to Identify skill requirements from Advertorial Data
Among the factors of production, human capital or skilled manpower is the one that keeps evolving and adapts to changing conditions and resources. This adaptability makes human capital the most crucial factor in ensuring a sustainable growth of industry/sector. As new technologies are developed and adopted, the new generations are required to acquire skills in newer technologies in order to be employable. At the same time professionals are required to upskill and reskill themselves to remain relevant in the industry. There is however no straightforward method to identify the skill needs of the industry at a given point of time. Therefore, this paper proposes a data to decision framework that can successfully identify the desired skill set in a given area by analysing the advertorial data collected from popular online job portals and supplied as input to the framework. The proposed framework uses techniques of statistical analysis, data mining and natural language processing for the purpose. The applicability of the framework is demonstrated on CS&IT job advertisement data from India. The analytical results not only provide useful insights about current state of skill needs in CS&IT industry but also provide practical implications to prospective job applicants, training agencies, and institutions of higher education & professional training.
☆ Enabling Versatile Controls for Video Diffusion Models
Despite substantial progress in text-to-video generation, achieving precise and flexible control over fine-grained spatiotemporal attributes remains a significant unresolved challenge in video generation research. To address these limitations, we introduce VCtrl (also termed PP-VCtrl), a novel framework designed to enable fine-grained control over pre-trained video diffusion models in a unified manner. VCtrl integrates diverse user-specified control signals-such as Canny edges, segmentation masks, and human keypoints-into pretrained video diffusion models via a generalizable conditional module capable of uniformly encoding multiple types of auxiliary signals without modifying the underlying generator. Additionally, we design a unified control signal encoding pipeline and a sparse residual connection mechanism to efficiently incorporate control representations. Comprehensive experiments and human evaluations demonstrate that VCtrl effectively enhances controllability and generation quality. The source code and pre-trained models are publicly available and implemented using the PaddlePaddle framework at http://github.com/PaddlePaddle/PaddleMIX/tree/develop/ppdiffusers/examples/ppvctrl.
comment: Codes and Supplementary Material: http://github.com/PaddlePaddle/PaddleMIX/tree/develop/ppdiffusers/examples/ppvctrl
☆ Token Dynamics: Towards Efficient and Dynamic Video Token Representation for Video Large Language Models
Token-based video representation has emerged as a promising approach for enabling large language models to interpret video content. However, existing token reduction techniques, such as token pruning and token merging, often disrupt essential spatial-temporal positional embeddings, failing to adequately balance computational efficiency with fewer tokens. Consequently, these methods result in relatively lengthy token sequences, limiting their applicability in scenarios requiring extreme token compression, such as video large language models. In this paper, we introduce the novel task of extreme short token reduction, aiming to represent extensive video sequences with a minimal number of tokens. To address this challenge, we propose Token Dynamics, a new video representation framework that dynamically reduces token count while preserving spatial-temporal coherence. Specifically, we disentangle video representations by separating visual embeddings from grid-level motion information, structuring them into: 1. a concise token base, created by clustering tokens that describe object-level content; 2. a token dynamics map, capturing detailed spatial-temporal motion patterns across grids. Furthermore, we introduce a cross-dynamics attention mechanism that integrates motion features into the token base without increasing token length, thereby maintaining compactness and spatial-temporal integrity. The experiments demonstrate a reduction of token count to merely 0.07% of the original tokens, with only a minor performance drop of 1.13%. Additionally, we propose two novel subtasks within extreme token reduction (fixed-length and adaptive-length compression), both effectively representing long token sequences for video-language tasks. Our method offers significantly lower theoretical complexity, fewer tokens, and enhanced throughput, thus providing an efficient solution for video LLMs.
☆ Real-Time Diffusion Policies for Games: Enhancing Consistency Policies with Q-Ensembles
Diffusion models have shown impressive performance in capturing complex and multi-modal action distributions for game agents, but their slow inference speed prevents practical deployment in real-time game environments. While consistency models offer a promising approach for one-step generation, they often suffer from training instability and performance degradation when applied to policy learning. In this paper, we present CPQE (Consistency Policy with Q-Ensembles), which combines consistency models with Q-ensembles to address these challenges.CPQE leverages uncertainty estimation through Q-ensembles to provide more reliable value function approximations, resulting in better training stability and improved performance compared to classic double Q-network methods. Our extensive experiments across multiple game scenarios demonstrate that CPQE achieves inference speeds of up to 60 Hz -- a significant improvement over state-of-the-art diffusion policies that operate at only 20 Hz -- while maintaining comparable performance to multi-step diffusion approaches. CPQE consistently outperforms state-of-the-art consistency model approaches, showing both higher rewards and enhanced training stability throughout the learning process. These results indicate that CPQE offers a practical solution for deploying diffusion-based policies in games and other real-time applications where both multi-modal behavior modeling and rapid inference are critical requirements.
☆ GeoT: Geometry-guided Instance-dependent Transition Matrix for Semi-supervised Tooth Point Cloud Segmentation
Achieving meticulous segmentation of tooth point clouds from intra-oral scans stands as an indispensable prerequisite for various orthodontic applications. Given the labor-intensive nature of dental annotation, a significant amount of data remains unlabeled, driving increasing interest in semi-supervised approaches. One primary challenge of existing semi-supervised medical segmentation methods lies in noisy pseudo labels generated for unlabeled data. To address this challenge, we propose GeoT, the first framework that employs instance-dependent transition matrix (IDTM) to explicitly model noise in pseudo labels for semi-supervised dental segmentation. Specifically, to handle the extensive solution space of IDTM arising from tens of thousands of dental points, we introduce tooth geometric priors through two key components: point-level geometric regularization (PLGR) to enhance consistency between point adjacency relationships in 3D and IDTM spaces, and class-level geometric smoothing (CLGS) to leverage the fixed spatial distribution of tooth categories for optimal IDTM estimation. Extensive experiments performed on the public Teeth3DS dataset and private dataset demonstrate that our method can make full utilization of unlabeled data to facilitate segmentation, achieving performance comparable to fully supervised methods with only $20\%$ of the labeled data.
comment: IPMI2025
☆ Assessing Consistency and Reproducibility in the Outputs of Large Language Models: Evidence Across Diverse Finance and Accounting Tasks
This study provides the first comprehensive assessment of consistency and reproducibility in Large Language Model (LLM) outputs in finance and accounting research. We evaluate how consistently LLMs produce outputs given identical inputs through extensive experimentation with 50 independent runs across five common tasks: classification, sentiment analysis, summarization, text generation, and prediction. Using three OpenAI models (GPT-3.5-turbo, GPT-4o-mini, and GPT-4o), we generate over 3.4 million outputs from diverse financial source texts and data, covering MD&As, FOMC statements, finance news articles, earnings call transcripts, and financial statements. Our findings reveal substantial but task-dependent consistency, with binary classification and sentiment analysis achieving near-perfect reproducibility, while complex tasks show greater variability. More advanced models do not consistently demonstrate better consistency and reproducibility, with task-specific patterns emerging. LLMs significantly outperform expert human annotators in consistency and maintain high agreement even where human experts significantly disagree. We further find that simple aggregation strategies across 3-5 runs dramatically improve consistency. Simulation analysis reveals that despite measurable inconsistency in LLM outputs, downstream statistical inferences remain remarkably robust. These findings address concerns about what we term "G-hacking," the selective reporting of favorable outcomes from multiple Generative AI runs, by demonstrating that such risks are relatively low for finance and accounting tasks.
comment: 96 pages
☆ ARFlow: Human Action-Reaction Flow Matching with Physical Guidance
Human action-reaction synthesis, a fundamental challenge in modeling causal human interactions, plays a critical role in applications ranging from virtual reality to social robotics. While diffusion-based models have demonstrated promising performance, they exhibit two key limitations for interaction synthesis: reliance on complex noise-to-reaction generators with intricate conditional mechanisms, and frequent physical violations in generated motions. To address these issues, we propose Action-Reaction Flow Matching (ARFlow), a novel framework that establishes direct action-to-reaction mappings, eliminating the need for complex conditional mechanisms. Our approach introduces two key innovations: an x1-prediction method that directly outputs human motions instead of velocity fields, enabling explicit constraint enforcement; and a training-free, gradient-based physical guidance mechanism that effectively prevents body penetration artifacts during sampling. Extensive experiments on NTU120 and Chi3D datasets demonstrate that ARFlow not only outperforms existing methods in terms of Fr\'echet Inception Distance and motion diversity but also significantly reduces body collisions, as measured by our new Intersection Volume and Intersection Frequency metrics.
comment: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
☆ From Faces to Voices: Learning Hierarchical Representations for High-quality Video-to-Speech CVPR 2025
The objective of this study is to generate high-quality speech from silent talking face videos, a task also known as video-to-speech synthesis. A significant challenge in video-to-speech synthesis lies in the substantial modality gap between silent video and multi-faceted speech. In this paper, we propose a novel video-to-speech system that effectively bridges this modality gap, significantly enhancing the quality of synthesized speech. This is achieved by learning of hierarchical representations from video to speech. Specifically, we gradually transform silent video into acoustic feature spaces through three sequential stages -- content, timbre, and prosody modeling. In each stage, we align visual factors -- lip movements, face identity, and facial expressions -- with corresponding acoustic counterparts to ensure the seamless transformation. Additionally, to generate realistic and coherent speech from the visual representations, we employ a flow matching model that estimates direct trajectories from a simple prior distribution to the target speech distribution. Extensive experiments demonstrate that our method achieves exceptional generation quality comparable to real utterances, outperforming existing methods by a significant margin.
comment: CVPR 2025, demo page: https://mm.kaist.ac.kr/projects/faces2voices/
☆ Neural-Guided Equation Discovery
Deep learning approaches are becoming increasingly attractive for equation discovery. We show the advantages and disadvantages of using neural-guided equation discovery by giving an overview of recent papers and the results of experiments using our modular equation discovery system MGMT ($\textbf{M}$ulti-Task $\textbf{G}$rammar-Guided $\textbf{M}$onte-Carlo $\textbf{T}$ree Search for Equation Discovery). The system uses neural-guided Monte-Carlo Tree Search (MCTS) and supports both supervised and reinforcement learning, with a search space defined by a context-free grammar. We summarize seven desirable properties of equation discovery systems, emphasizing the importance of embedding tabular data sets for such learning approaches. Using the modular structure of MGMT, we compare seven architectures (among them, RNNs, CNNs, and Transformers) for embedding tabular datasets on the auxiliary task of contrastive learning for tabular data sets on an equation discovery task. For almost all combinations of modules, supervised learning outperforms reinforcement learning. Moreover, our experiments indicate an advantage of using grammar rules as action space instead of tokens. Two adaptations of MCTS -- risk-seeking MCTS and AmEx-MCTS -- can improve equation discovery with that kind of search.
comment: 32 pages + 4 pages appendix, 9 figures, book chapter
☆ On-Sensor Convolutional Neural Networks with Early-Exits SC
Tiny Machine Learning (TinyML) is a novel research field aiming at integrating Machine Learning (ML) within embedded devices with limited memory, computation, and energy. Recently, a new branch of TinyML has emerged, focusing on integrating ML directly into the sensors to further reduce the power consumption of embedded devices. Interestingly, despite their state-of-the-art performance in many tasks, none of the current solutions in the literature aims to optimize the implementation of Convolutional Neural Networks (CNNs) operating directly into sensors. In this paper, we introduce for the first time in the literature the optimized design and implementation of Depth-First CNNs operating on the Intelligent Sensor Processing Unit (ISPU) within an Inertial Measurement Unit (IMU) by STMicroelectronics. Our approach partitions the CNN between the ISPU and the microcontroller (MCU) and employs an Early-Exit mechanism to stop the computations on the IMU when enough confidence about the results is achieved, hence significantly reducing power consumption. When using a NUCLEO-F411RE board, this solution achieved an average current consumption of 4.8 mA, marking an 11% reduction compared to the regular inference pipeline on the MCU, while having equal accuracy.
comment: Presented at IEEE SSCI
☆ Interpretable Machine Learning for Oral Lesion Diagnosis through Prototypical Instances Identification
Decision-making processes in healthcare can be highly complex and challenging. Machine Learning tools offer significant potential to assist in these processes. However, many current methodologies rely on complex models that are not easily interpretable by experts. This underscores the need to develop interpretable models that can provide meaningful support in clinical decision-making. When approaching such tasks, humans typically compare the situation at hand to a few key examples and representative cases imprinted in their memory. Using an approach which selects such exemplary cases and grounds its predictions on them could contribute to obtaining high-performing interpretable solutions to such problems. To this end, we evaluate PivotTree, an interpretable prototype selection model, on an oral lesion detection problem, specifically trying to detect the presence of neoplastic, aphthous and traumatic ulcerated lesions from oral cavity images. We demonstrate the efficacy of using such method in terms of performance and offer a qualitative and quantitative comparison between exemplary cases and ground-truth prototypes selected by experts.
☆ Rude Humans and Vengeful Robots: Examining Human Perceptions of Robot Retaliatory Intentions in Professional Settings
Humans and robots are increasingly working in personal and professional settings. In workplace settings, humans and robots may work together as colleagues, potentially leading to social expectations, or violation thereof. Extant research has primarily sought to understand social interactions and expectations in personal rather than professional settings, and none of these studies have examined negative outcomes arising from violations of social expectations. This paper reports the results of a 2x3 online experiment that used a unique first-person perspective video to immerse participants in a collaborative workplace setting. The results are nuanced and reveal that while robots are expected to act in accordance with social expectations despite human behavior, there are benefits for robots perceived as being the bigger person in the face of human rudeness. Theoretical and practical implications are provided which discuss the import of these findings for the design of social robots.
comment: This is the author version of the manuscript submitted to ACM Transactions on Human-Robot Interaction. The final version, if accepted, will be published by ACM and available via the ACM Digital Library. 12 pages, 1 figure, 2 tables
☆ TEMPO: Temporal Preference Optimization of Video LLMs via Difficulty Scheduling and Pre-SFT Alignment
Video Large Language Models (Video LLMs) have achieved significant success by leveraging a two-stage paradigm: pretraining on large-scale video-text data for vision-language alignment, followed by supervised fine-tuning (SFT) for task-specific capabilities. However, existing approaches struggle with temporal reasoning due to weak temporal correspondence in the data and reliance on the next-token prediction paradigm during training. To address these limitations, we propose TEMPO (TEMporal Preference Optimization), a systematic framework that enhances Video LLMs' temporal reasoning capabilities through Direct Preference Optimization (DPO). To facilitate this, we introduce an automated preference data generation pipeline that systematically constructs preference pairs by selecting videos that are rich in temporal information, designing video-specific perturbation strategies, and finally evaluating model responses on clean and perturbed video inputs. Our temporal alignment features two key innovations: curriculum learning which that progressively increases perturbation difficulty to improve model robustness and adaptability; and ``Pre-SFT Alignment'', applying preference optimization before instruction tuning to prioritize fine-grained temporal comprehension. Extensive experiments demonstrate that our approach consistently improves Video LLM performance across multiple benchmarks with a relatively small set of self-generated DPO data. We further analyze the transferability of DPO data across architectures and the role of difficulty scheduling in optimization. Our findings highlight our TEMPO as a scalable and efficient complement to SFT-based methods, paving the way for developing reliable Video LLMs.
RustEvo^2: An Evolving Benchmark for API Evolution in LLM-based Rust Code Generation
Large Language Models (LLMs) have become pivotal tools for automating code generation in software development. However, these models face significant challenges in producing version-aware code for rapidly evolving languages like Rust, where frequent Application Programming Interfaces (API) changes across versions lead to compatibility issues and correctness errors. Existing benchmarks lack systematic evaluation of how models navigate API transitions, relying on labor-intensive manual curation and offering limited version-specific insights. To address this gap, we present RustEvo, a novel framework for constructing dynamic benchmarks that evaluate the ability of LLMs to adapt to evolving Rust APIs. RustEvo automates dataset creation by synthesizing 588 API changes (380 from Rust standard libraries, 208 from 15 third-party crates) into programming tasks mirroring real-world challenges. These tasks cover four API evolution categories: Stabilizations, Signature Changes, Behavioral Changes, and Deprecations, reflecting their actual distribution in the Rust ecosystem. Experiments on state-of-the-art (SOTA) LLMs reveal significant performance variations: models achieve a 65.8% average success rate on stabilized APIs but only 38.0% on behavioral changes, highlighting difficulties in detecting semantic shifts without signature alterations. Knowledge cutoff dates strongly influence performance, with models scoring 56.1% on before-cutoff APIs versus 32.5% on after-cutoff tasks. Retrieval-Augmented Generation (RAG) mitigates this gap, improving success rates by 13.5% on average for APIs released after model training. Our findings underscore the necessity of our evolution-aware benchmarks to advance the adaptability of LLMs in fast-paced software ecosystems. The framework and the benchmarks are publicly released at https://github.com/SYSUSELab/RustEvo.
☆ When Preferences Diverge: Aligning Diffusion Models with Minority-Aware Adaptive DPO
In recent years, the field of image generation has witnessed significant advancements, particularly in fine-tuning methods that align models with universal human preferences. This paper explores the critical role of preference data in the training process of diffusion models, particularly in the context of Diffusion-DPO and its subsequent adaptations. We investigate the complexities surrounding universal human preferences in image generation, highlighting the subjective nature of these preferences and the challenges posed by minority samples in preference datasets. Through pilot experiments, we demonstrate the existence of minority samples and their detrimental effects on model performance. We propose Adaptive-DPO -- a novel approach that incorporates a minority-instance-aware metric into the DPO objective. This metric, which includes intra-annotator confidence and inter-annotator stability, distinguishes between majority and minority samples. We introduce an Adaptive-DPO loss function which improves the DPO loss in two ways: enhancing the model's learning of majority labels while mitigating the negative impact of minority samples. Our experiments demonstrate that this method effectively handles both synthetic minority data and real-world preference data, paving the way for more effective training methodologies in image generation tasks.
☆ Understanding Social Support Needs in Questions: A Hybrid Approach Integrating Semi-Supervised Learning and LLM-based Data Augmentation
Patients are increasingly turning to online health Q&A communities for social support to improve their well-being. However, when this support received does not align with their specific needs, it may prove ineffective or even detrimental. This necessitates a model capable of identifying the social support needs in questions. However, training such a model is challenging due to the scarcity and class imbalance issues of labeled data. To overcome these challenges, we follow the computational design science paradigm to develop a novel framework, Hybrid Approach for SOcial Support need classification (HA-SOS). HA-SOS integrates an answer-enhanced semi-supervised learning approach, a text data augmentation technique leveraging large language models (LLMs) with reliability- and diversity-aware sample selection mechanism, and a unified training process to automatically label social support needs in questions. Extensive empirical evaluations demonstrate that HA-SOS significantly outperforms existing question classification models and alternative semi-supervised learning approaches. This research contributes to the literature on social support, question classification, semi-supervised learning, and text data augmentation. In practice, our HA-SOS framework facilitates online Q&A platform managers and answerers to better understand users' social support needs, enabling them to provide timely, personalized answers and interventions.
comment: 55 pages
☆ A New Segment Routing method with Swap Node Selection Strategy Based on Deep Reinforcement Learning for Software Defined Network
The existing segment routing (SR) methods need to determine the routing first and then use path segmentation approaches to select swap nodes to form a segment routing path (SRP). They require re-segmentation of the path when the routing changes. Furthermore, they do not consider the flow table issuance time, which cannot maximize the speed of issuance flow table. To address these issues, this paper establishes an optimization model that can simultaneously form routing strategies and path segmentation strategies for selecting the appropriate swap nodes to reduce flow table issuance time. It also designs an intelligent segment routing algorithm based on deep reinforcement learning (DRL-SR) to solve the proposed model. First, a traffic matrix is designed as the state space for the deep reinforcement learning agent; this matrix includes multiple QoS performance indicators, flow table issuance time overhead and SR label stack depth. Second, the action selection strategy and corresponding reward function are designed, where the agent selects the next node considering the routing; in addition, the action selection strategy whether the newly added node is selected as the swap node and the corresponding reward function are designed considering the time cost factor for the controller to issue the flow table to the swap node. Finally, a series of experiments and their results show that, compared with the existing methods, the designed segmented route optimization model and the intelligent solution algorithm (DRL-SR) can reduce the time overhead required to complete the segmented route establishment task while optimizing performance metrics such as throughput, delays and packet losses.
☆ MAPS: A Multi-Agent Framework Based on Big Seven Personality and Socratic Guidance for Multimodal Scientific Problem Solving
Multimodal scientific problems (MSPs) involve complex issues that require the integration of multiple modalities, such as text and diagrams, presenting a significant challenge in artificial intelligence. While progress has been made in addressing traditional scientific problems, MSPs still face two primary issues: the challenge of multi-modal comprehensive reasoning in scientific problem-solving and the lack of reflective and rethinking capabilities. To address these issues, we introduce a Multi-Agent framework based on the Big Seven Personality and Socratic guidance (MAPS). This framework employs seven distinct agents that leverage feedback mechanisms and the Socratic method to guide the resolution of MSPs. To tackle the first issue, we propose a progressive four-agent solving strategy, where each agent focuses on a specific stage of the problem-solving process. For the second issue, we introduce a Critic agent, inspired by Socratic questioning, which prompts critical thinking and stimulates autonomous learning. We conduct extensive experiments on the EMMA, Olympiad, and MathVista datasets, achieving promising results that outperform the current SOTA model by 15.84% across all tasks. Meanwhile, the additional analytical experiments also verify the model's progress as well as generalization ability.
☆ Deep Learning for Human Locomotion Analysis in Lower-Limb Exoskeletons: A Comparative Study
Wearable robotics for lower-limb assistance have become a pivotal area of research, aiming to enhance mobility for individuals with physical impairments or augment the performance of able-bodied users. Accurate and adaptive control systems are essential to ensure seamless interaction between the wearer and the robotic device, particularly when navigating diverse and dynamic terrains. Despite the recent advances in neural networks for time series analysis, no attempts have been directed towards the classification of ground conditions, categorized into five classes and subsequently determining the ramp's slope and stair's height. In this respect, this paper presents an experimental comparison between eight deep neural network backbones to predict high-level locomotion parameters across diverse terrains. All the models are trained on the publicly available CAMARGO 2021 dataset. IMU-only data equally or outperformed IMU+EMG inputs, promoting a cost-effective and efficient design. Indeeds, using three IMU sensors, the LSTM achieved high terrain classification accuracy (0.94 +- 0.04) and precise ramp slope (1.95 +- 0.58{\deg}) and the CNN-LSTM a stair height (15.65 +- 7.40 mm) estimations. As a further contribution, SHAP analysis justified sensor reduction without performance loss, ensuring a lightweight setup. The system operates with ~2 ms inference time, supporting real-time applications. The code is code available at https://github.com/cosbidev/Human-Locomotion-Identification.
comment: 26 pages, 6 figures
☆ MARS: A Multi-Agent Framework Incorporating Socratic Guidance for Automated Prompt Optimization
The basic question-answering format of large language models involves inputting a prompt and receiving a response, and the quality of the prompt directly impacts the effectiveness of the response. Automated Prompt Optimization (APO) aims to break free from the cognitive biases of manually designed prompts and explores a broader design space for prompts. However, existing APO methods suffer from limited flexibility of fixed templates and inefficient search in prompt spaces as key issues. To this end, we propose a Multi-Agent framework Incorporating Socratic guidance (MARS), which utilizes multi-agent fusion technology for automatic planning, with gradual continuous optimization and evaluation. Specifically, MARS comprises seven agents, each with distinct functionalities, which autonomously use the Planner to devise an optimization path that ensures flexibility. Additionally, it employs a Teacher-Critic-Student Socratic dialogue pattern to iteratively optimize the prompts while conducting effective search. We conduct extensive experiments on various datasets to validate the effectiveness of our method, and perform additional analytical experiments to assess the model's advancement as well as the interpretability.
☆ Classifier-guided CLIP Distillation for Unsupervised Multi-label Classification CVPR 2025
Multi-label classification is crucial for comprehensive image understanding, yet acquiring accurate annotations is challenging and costly. To address this, a recent study suggests exploiting unsupervised multi-label classification leveraging CLIP, a powerful vision-language model. Despite CLIP's proficiency, it suffers from view-dependent predictions and inherent bias, limiting its effectiveness. We propose a novel method that addresses these issues by leveraging multiple views near target objects, guided by Class Activation Mapping (CAM) of the classifier, and debiasing pseudo-labels derived from CLIP predictions. Our Classifier-guided CLIP Distillation (CCD) enables selecting multiple local views without extra labels and debiasing predictions to enhance classification performance. Experimental results validate our method's superiority over existing techniques across diverse datasets. The code is available at https://github.com/k0u-id/CCD.
comment: CVPR 2025 Accepted
☆ Sparse Logit Sampling: Accelerating Knowledge Distillation in LLMs
Knowledge distillation can be a cost-effective technique to distill knowledge in Large Language Models, if the teacher output logits can be pre-computed and cached. However, successfully applying this to pre-training remains largely unexplored. In this work, we prove that naive approaches for sparse knowledge distillation such as caching Top-K probabilities, while intuitive, provide biased estimates of teacher probability distribution to the student, resulting in suboptimal performance and calibration. We propose an importance-sampling-based method `Random Sampling Knowledge Distillation', which provides unbiased estimates, preserves the gradient in expectation, and requires storing significantly sparser logits. Our method enables faster training of student models with marginal overhead (<10%) compared to cross-entropy based training, while maintaining competitive performance compared to full distillation, across a range of model sizes from 300M to 3B.
comment: Anshumann, Mohd Abbas Zaidi and Akhil Kedia have Equal Contribution
☆ In-House Evaluation Is Not Enough: Towards Robust Third-Party Flaw Disclosure for General-Purpose AI
The widespread deployment of general-purpose AI (GPAI) systems introduces significant new risks. Yet the infrastructure, practices, and norms for reporting flaws in GPAI systems remain seriously underdeveloped, lagging far behind more established fields like software security. Based on a collaboration between experts from the fields of software security, machine learning, law, social science, and policy, we identify key gaps in the evaluation and reporting of flaws in GPAI systems. We call for three interventions to advance system safety. First, we propose using standardized AI flaw reports and rules of engagement for researchers in order to ease the process of submitting, reproducing, and triaging flaws in GPAI systems. Second, we propose GPAI system providers adopt broadly-scoped flaw disclosure programs, borrowing from bug bounties, with legal safe harbors to protect researchers. Third, we advocate for the development of improved infrastructure to coordinate distribution of flaw reports across the many stakeholders who may be impacted. These interventions are increasingly urgent, as evidenced by the prevalence of jailbreaks and other flaws that can transfer across different providers' GPAI systems. By promoting robust reporting and coordination in the AI ecosystem, these proposals could significantly improve the safety, security, and accountability of GPAI systems.
☆ MTBench: A Multimodal Time Series Benchmark for Temporal Reasoning and Question Answering
Understanding the relationship between textual news and time-series evolution is a critical yet under-explored challenge in applied data science. While multimodal learning has gained traction, existing multimodal time-series datasets fall short in evaluating cross-modal reasoning and complex question answering, which are essential for capturing complex interactions between narrative information and temporal patterns. To bridge this gap, we introduce Multimodal Time Series Benchmark (MTBench), a large-scale benchmark designed to evaluate large language models (LLMs) on time series and text understanding across financial and weather domains. MTbench comprises paired time series and textual data, including financial news with corresponding stock price movements and weather reports aligned with historical temperature records. Unlike existing benchmarks that focus on isolated modalities, MTbench provides a comprehensive testbed for models to jointly reason over structured numerical trends and unstructured textual narratives. The richness of MTbench enables formulation of diverse tasks that require a deep understanding of both text and time-series data, including time-series forecasting, semantic and technical trend analysis, and news-driven question answering (QA). These tasks target the model's ability to capture temporal dependencies, extract key insights from textual context, and integrate cross-modal information. We evaluate state-of-the-art LLMs on MTbench, analyzing their effectiveness in modeling the complex relationships between news narratives and temporal patterns. Our findings reveal significant challenges in current models, including difficulties in capturing long-term dependencies, interpreting causality in financial and weather trends, and effectively fusing multimodal information.
comment: 14 pages
☆ Imagine to Hear: Auditory Knowledge Generation can be an Effective Assistant for Language Models
Language models pretrained on text-only corpora often struggle with tasks that require auditory commonsense knowledge. Previous work addresses this problem by augmenting the language model to retrieve knowledge from external audio databases. This approach has several limitations, such as the potential lack of relevant audio in databases and the high costs associated with constructing and querying the databases. To address these issues, we propose Imagine to Hear, a novel approach that dynamically generates auditory knowledge using generative models. Our framework detects multiple audio-related textual spans from the given prompt and generates corresponding auditory knowledge. We develop several mechanisms to efficiently process multiple auditory knowledge, including a CLAP-based rejection sampler and a language-audio fusion module. Our experiments show that our method achieves state-of-the-art performance on AuditoryBench without relying on external databases, highlighting the effectiveness of our generation-based approach.
comment: Preprint
☆ Casual Inference via Style Bias Deconfounding for Domain Generalization
Deep neural networks (DNNs) often struggle with out-of-distribution data, limiting their reliability in diverse realworld applications. To address this issue, domain generalization methods have been developed to learn domain-invariant features from single or multiple training domains, enabling generalization to unseen testing domains. However, existing approaches usually overlook the impact of style frequency within the training set. This oversight predisposes models to capture spurious visual correlations caused by style confounding factors, rather than learning truly causal representations, thereby undermining inference reliability. In this work, we introduce Style Deconfounding Causal Learning (SDCL), a novel causal inference-based framework designed to explicitly address style as a confounding factor. Our approaches begins with constructing a structural causal model (SCM) tailored to the domain generalization problem and applies a backdoor adjustment strategy to account for style influence. Building on this foundation, we design a style-guided expert module (SGEM) to adaptively clusters style distributions during training, capturing the global confounding style. Additionally, a back-door causal learning module (BDCL) performs causal interventions during feature extraction, ensuring fair integration of global confounding styles into sample predictions, effectively reducing style bias. The SDCL framework is highly versatile and can be seamlessly integrated with state-of-the-art data augmentation techniques. Extensive experiments across diverse natural and medical image recognition tasks validate its efficacy, demonstrating superior performance in both multi-domain and the more challenging single-domain generalization scenarios.
comment: under review
☆ Physics-Informed Neural Network Surrogate Models for River Stage Prediction
This work investigates the feasibility of using Physics-Informed Neural Networks (PINNs) as surrogate models for river stage prediction, aiming to reduce computational cost while maintaining predictive accuracy. Our primary contribution demonstrates that PINNs can successfully approximate HEC-RAS numerical solutions when trained on a single river, achieving strong predictive accuracy with generally low relative errors, though some river segments exhibit higher deviations. By integrating the governing Saint-Venant equations into the learning process, the proposed PINN-based surrogate model enforces physical consistency and significantly improves computational efficiency compared to HEC-RAS. We evaluate the model's performance in terms of accuracy and computational speed, demonstrating that it closely approximates HEC-RAS predictions while enabling real-time inference. These results highlight the potential of PINNs as effective surrogate models for single-river hydrodynamics, offering a promising alternative for computationally efficient river stage forecasting. Future work will explore techniques to enhance PINN training stability and robustness across a more generalized multi-river model.
comment: 10 pages, 5 figures
☆ The Deployment of End-to-End Audio Language Models Should Take into Account the Principle of Least Privilege
We are at a turning point for language models that accept audio input. The latest end-to-end audio language models (Audio LMs) process speech directly instead of relying on a separate transcription step. This shift preserves detailed information, such as intonation or the presence of multiple speakers, that would otherwise be lost in transcription. However, it also introduces new safety risks, including the potential misuse of speaker identity cues and other sensitive vocal attributes, which could have legal implications. In this position paper, we urge a closer examination of how these models are built and deployed. We argue that the principle of least privilege should guide decisions on whether to deploy cascaded or end-to-end models. Specifically, evaluations should assess (1) whether end-to-end modeling is necessary for a given application; and (2), the appropriate scope of information access. Finally, We highlight related gaps in current audio LM benchmarks and identify key open research questions, both technical and policy-related, that must be addressed to enable the responsible deployment of end-to-end Audio LMs.
☆ DyWA: Dynamics-adaptive World Action Model for Generalizable Non-prehensile Manipulation
Nonprehensile manipulation is crucial for handling objects that are too thin, large, or otherwise ungraspable in unstructured environments. While conventional planning-based approaches struggle with complex contact modeling, learning-based methods have recently emerged as a promising alternative. However, existing learning-based approaches face two major limitations: they heavily rely on multi-view cameras and precise pose tracking, and they fail to generalize across varying physical conditions, such as changes in object mass and table friction. To address these challenges, we propose the Dynamics-Adaptive World Action Model (DyWA), a novel framework that enhances action learning by jointly predicting future states while adapting to dynamics variations based on historical trajectories. By unifying the modeling of geometry, state, physics, and robot actions, DyWA enables more robust policy learning under partial observability. Compared to baselines, our method improves the success rate by 31.5% using only single-view point cloud observations in the simulation. Furthermore, DyWA achieves an average success rate of 68% in real-world experiments, demonstrating its ability to generalize across diverse object geometries, adapt to varying table friction, and robustness in challenging scenarios such as half-filled water bottles and slippery surfaces.
comment: Project Page:https://pku-epic.github.io/DyWA/
☆ Auto-Regressive Diffusion for Generating 3D Human-Object Interactions
Text-driven Human-Object Interaction (Text-to-HOI) generation is an emerging field with applications in animation, video games, virtual reality, and robotics. A key challenge in HOI generation is maintaining interaction consistency in long sequences. Existing Text-to-Motion-based approaches, such as discrete motion tokenization, cannot be directly applied to HOI generation due to limited data in this domain and the complexity of the modality. To address the problem of interaction consistency in long sequences, we propose an autoregressive diffusion model (ARDHOI) that predicts the next continuous token. Specifically, we introduce a Contrastive Variational Autoencoder (cVAE) to learn a physically plausible space of continuous HOI tokens, thereby ensuring that generated human-object motions are realistic and natural. For generating sequences autoregressively, we develop a Mamba-based context encoder to capture and maintain consistent sequential actions. Additionally, we implement an MLP-based denoiser to generate the subsequent token conditioned on the encoded context. Our model has been evaluated on the OMOMO and BEHAVE datasets, where it outperforms existing state-of-the-art methods in terms of both performance and inference speed. This makes ARDHOI a robust and efficient solution for text-driven HOI tasks
☆ Causally Aligned Curriculum Learning ICLR 2024
A pervasive challenge in Reinforcement Learning (RL) is the "curse of dimensionality" which is the exponential growth in the state-action space when optimizing a high-dimensional target task. The framework of curriculum learning trains the agent in a curriculum composed of a sequence of related and more manageable source tasks. The expectation is that when some optimal decision rules are shared across source tasks and the target task, the agent could more quickly pick up the necessary skills to behave optimally in the environment, thus accelerating the learning process. However, this critical assumption of invariant optimal decision rules does not necessarily hold in many practical applications, specifically when the underlying environment contains unobserved confounders. This paper studies the problem of curriculum RL through causal lenses. We derive a sufficient graphical condition characterizing causally aligned source tasks, i.e., the invariance of optimal decision rules holds. We further develop an efficient algorithm to generate a causally aligned curriculum, provided with qualitative causal knowledge of the target task. Finally, we validate our proposed methodology through experiments in discrete and continuous confounded tasks with pixel observations.
comment: Accepted as Posters in ICLR 2024
☆ A Learnability Analysis on Neuro-Symbolic Learning
This paper analyzes the learnability of neuro-symbolic (NeSy) tasks within hybrid systems. We show that the learnability of NeSy tasks can be characterized by their derived constraint satisfaction problems (DCSPs). Specifically, a task is learnable if the corresponding DCSP has a unique solution; otherwise, it is unlearnable. For learnable tasks, we establish error bounds by exploiting the clustering property of the hypothesis space. Additionally, we analyze the asymptotic error for general NeSy tasks, showing that the expected error scales with the disagreement among solutions. Our results offer a principled approach to determining learnability and provide insights into the design of new algorithms.
☆ "The Diagram is like Guardrails": Structuring GenAI-assisted Hypotheses Exploration with an Interactive Shared Representation
Data analysis encompasses a spectrum of tasks, from high-level conceptual reasoning to lower-level execution. While AI-powered tools increasingly support execution tasks, there remains a need for intelligent assistance in conceptual tasks. This paper investigates the design of an ordered node-link tree interface augmented with AI-generated information hints and visualizations, as a potential shared representation for hypothesis exploration. Through a design probe (n=22), participants generated diagrams averaging 21.82 hypotheses. Our findings showed that the node-link diagram acts as "guardrails" for hypothesis exploration, facilitating structured workflows, providing comprehensive overviews, and enabling efficient backtracking. The AI-generated information hints, particularly visualizations, aided users in transforming abstract ideas into data-backed concepts while reducing cognitive load. We further discuss how node-link diagrams can support both parallel exploration and iterative refinement in hypothesis formulation, potentially enhancing the breadth and depth of human-AI collaborative data analysis.
☆ Does Chain-of-Thought Reasoning Help Mobile GUI Agent? An Empirical Study
Reasoning capabilities have significantly improved the performance of vision-language models (VLMs) in domains such as mathematical problem-solving, coding, and visual question-answering. However, their impact on real-world applications remains unclear. This paper presents the first empirical study on the effectiveness of reasoning-enabled VLMs in mobile GUI agents, a domain that requires interpreting complex screen layouts, understanding user instructions, and executing multi-turn interactions. We evaluate two pairs of commercial models--Gemini 2.0 Flash and Claude 3.7 Sonnet--comparing their base and reasoning-enhanced versions across two static benchmarks (ScreenSpot and AndroidControl) and one interactive environment (AndroidWorld). We surprisingly find the Claude 3.7 Sonnet reasoning model achieves state-of-the-art performance on AndroidWorld. However, reasoning VLMs generally offer marginal improvements over non-reasoning models on static benchmarks and even degrade performance in some agent setups. Notably, reasoning and non-reasoning VLMs fail on different sets of tasks, suggesting that reasoning does have an impact, but its benefits and drawbacks counterbalance each other. We attribute these inconsistencies to the limitations of benchmarks and VLMs. Based on the findings, we provide insights for further enhancing mobile GUI agents in terms of benchmarks, VLMs, and their adaptability in dynamically invoking reasoning VLMs. The experimental data are publicly available at https://github.com/LlamaTouch/VLM-Reasoning-Traces.
☆ Learning Part Knowledge to Facilitate Category Understanding for Fine-Grained Generalized Category Discovery
Generalized Category Discovery (GCD) aims to classify unlabeled data containing both seen and novel categories. Although existing methods perform well on generic datasets, they struggle in fine-grained scenarios. We attribute this difficulty to their reliance on contrastive learning over global image features to automatically capture discriminative cues, which fails to capture the subtle local differences essential for distinguishing fine-grained categories. Therefore, in this paper, we propose incorporating part knowledge to address fine-grained GCD, which introduces two key challenges: the absence of annotations for novel classes complicates the extraction of the part features, and global contrastive learning prioritizes holistic feature invariance, inadvertently suppressing discriminative local part patterns. To address these challenges, we propose PartGCD, including 1) Adaptive Part Decomposition, which automatically extracts class-specific semantic parts via Gaussian Mixture Models, and 2) Part Discrepancy Regularization, enforcing explicit separation between part features to amplify fine-grained local part distinctions. Experiments demonstrate state-of-the-art performance across multiple fine-grained benchmarks while maintaining competitiveness on generic datasets, validating the effectiveness and robustness of our approach.
☆ Chain-of-Tools: Utilizing Massive Unseen Tools in the CoT Reasoning of Frozen Language Models
Tool learning can further broaden the usage scenarios of large language models (LLMs). However most of the existing methods either need to finetune that the model can only use tools seen in the training data, or add tool demonstrations into the prompt with lower efficiency. In this paper, we present a new Tool Learning method Chain-of-Tools. It makes full use of the powerful semantic representation capability of frozen LLMs to finish tool calling in CoT reasoning with a huge and flexible tool pool which may contain unseen tools. Especially, to validate the effectiveness of our approach in the massive unseen tool scenario, we construct a new dataset SimpleToolQuestions. We conduct experiments on two numerical reasoning benchmarks (GSM8K-XL and FuncQA) and two knowledge-based question answering benchmarks (KAMEL and SimpleToolQuestions). Experimental results show that our approach performs better than the baseline. We also identify dimensions of the model output that are critical in tool selection, enhancing the model interpretability. Our code and data are available at: https://github.com/fairyshine/Chain-of-Tools .
comment: 11 pages, 10 figures
☆ Generative Modeling of Class Probability for Multi-Modal Representation Learning CVPR2025
Multi-modal understanding plays a crucial role in artificial intelligence by enabling models to jointly interpret inputs from different modalities. However, conventional approaches such as contrastive learning often struggle with modality discrepancies, leading to potential misalignments. In this paper, we propose a novel class anchor alignment approach that leverages class probability distributions for multi-modal representation learning. Our method, Class-anchor-ALigned generative Modeling (CALM), encodes class anchors as prompts to generate and align class probability distributions for each modality, enabling more effective alignment. Furthermore, we introduce a cross-modal probabilistic variational autoencoder to model uncertainty in the alignment, enhancing the ability to capture deeper relationships between modalities and data variations. Extensive experiments on four benchmark datasets demonstrate that our approach significantly outperforms state-of-the-art methods, especially in out-of-domain evaluations. This highlights its superior generalization capabilities in multi-modal representation learning.
comment: Accepted to CVPR2025
☆ Debugging and Runtime Analysis of Neural Networks with VLMs (A Case Study)
Debugging of Deep Neural Networks (DNNs), particularly vision models, is very challenging due to the complex and opaque decision-making processes in these networks. In this paper, we explore multi-modal Vision-Language Models (VLMs), such as CLIP, to automatically interpret the opaque representation space of vision models using natural language. This in turn, enables a semantic analysis of model behavior using human-understandable concepts, without requiring costly human annotations. Key to our approach is the notion of semantic heatmap, that succinctly captures the statistical properties of DNNs in terms of the concepts discovered with the VLM and that are computed off-line using a held-out data set. We show the utility of semantic heatmaps for fault localization -- an essential step in debugging -- in vision models. Our proposed technique helps localize the fault in the network (encoder vs head) and also highlights the responsible high-level concepts, by leveraging novel differential heatmaps, which summarize the semantic differences between the correct and incorrect behaviour of the analyzed DNN. We further propose a lightweight runtime analysis to detect and filter-out defects at runtime, thus improving the reliability of the analyzed DNNs. The runtime analysis works by measuring and comparing the similarity between the heatmap computed for a new (unseen) input and the heatmaps computed a-priori for correct vs incorrect DNN behavior. We consider two types of defects: misclassifications and vulnerabilities to adversarial attacks. We demonstrate the debugging and runtime analysis on a case study involving a complex ResNet-based classifier trained on the RIVAL10 dataset.
comment: CAIN 2025 (4th International Conference on AI Engineering -- Software Engineering for AI)
☆ Enhancing Subsequent Video Retrieval via Vision-Language Models (VLMs)
The rapid growth of video content demands efficient and precise retrieval systems. While vision-language models (VLMs) excel in representation learning, they often struggle with adaptive, time-sensitive video retrieval. This paper introduces a novel framework that combines vector similarity search with graph-based data structures. By leveraging VLM embeddings for initial retrieval and modeling contextual relationships among video segments, our approach enables adaptive query refinement and improves retrieval accuracy. Experiments demonstrate its precision, scalability, and robustness, offering an effective solution for interactive video retrieval in dynamic environments.
♻ ☆ Deep Model Merging: The Sister of Neural Network Interpretability -- A Survey
We survey the model merging literature through the lens of loss landscape geometry to connect observations from empirical studies on model merging and loss landscape analysis to phenomena that govern neural network training and the emergence of their inner representations. We distill repeated empirical observations from the literature in these fields into descriptions of four major characteristics of loss landscape geometry: mode convexity, determinism, directedness, and connectivity. We argue that insights into the structure of learned representations from model merging have applications to model interpretability and robustness, subsequently we propose promising new research directions at the intersection of these fields.
♻ ☆ ResBench: Benchmarking LLM-Generated FPGA Designs with Resource Awareness
Field-Programmable Gate Arrays (FPGAs) are widely used in modern hardware design, yet writing Hardware Description Language (HDL) code for FPGA implementation remains a complex and time-consuming task. Large Language Models (LLMs) have emerged as a promising tool for HDL generation, but existing benchmarks for LLM-based code generation primarily focus on functional correctness while overlooking hardware resource usage. Furthermore, current benchmarks offer limited diversity and do not fully represent the wide range of real-world FPGA applications. To address these shortcomings, we introduce ResBench, the first resource-focused benchmark explicitly designed to distinguish between resource-optimized and inefficient LLM-generated HDL code. ResBench consists of 56 problems across 12 categories, covering applications from finite state machines to financial computing. Our open-source evaluation framework automatically tests LLMs by generating Verilog code, verifying correctness, and measuring resource usage. The experiments, which primarily analyze Lookup Table (LUT) usage, reveal significant differences among LLMs, demonstrating ResBench's capability to identify models that generate more resource-optimized FPGA designs.
comment: to be published in International Symposium on Highly Efficient Accelerators and Reconfigurable Technologies 2025
♻ ☆ Constrained multi-fidelity Bayesian optimization with automatic stop condition
Bayesian optimization (BO) is increasingly employed in critical applications to find the optimal design with minimal cost. While BO is known for its sample efficiency, relying solely on costly high-fidelity data can still result in high costs. This is especially the case in constrained search spaces where BO must not only optimize but also ensure feasibility. A related issue in the BO literature is the lack of a systematic stopping criterion. To solve these challenges, we develop a constrained cost-aware multi-fidelity BO (CMFBO) framework whose goal is to minimize overall sampling costs by utilizing inexpensive low-fidelity sources while ensuring feasibility. In our case, the constraints can change across the data sources and may be even black-box functions. We also introduce a systematic stopping criterion that addresses the long-lasting issue associated with BO's convergence assessment. Our framework is publicly available on GitHub through the GP+ Python package and herein we validate it's efficacy on multiple benchmark problems.
♻ ☆ Humans and Large Language Models in Clinical Decision Support: A Study with Medical Calculators
Although large language models (LLMs) have been assessed for general medical knowledge using licensing exams, their ability to support clinical decision-making, such as selecting medical calculators, remains uncertain. We assessed nine LLMs, including open-source, proprietary, and domain-specific models, with 1,009 multiple-choice question-answer pairs across 35 clinical calculators and compared LLMs to humans on a subset of questions. While the highest-performing LLM, OpenAI o1, provided an answer accuracy of 66.0% (CI: 56.7-75.3%) on the subset of 100 questions, two human annotators nominally outperformed LLMs with an average answer accuracy of 79.5% (CI: 73.5-85.0%). Ultimately, we evaluated medical trainees and LLMs in recommending medical calculators across clinical scenarios like risk stratification and diagnosis. With error analysis showing that the highest-performing LLMs continue to make mistakes in comprehension (49.3% of errors) and calculator knowledge (7.1% of errors), our findings highlight that LLMs are not superior to humans in calculator recommendation.
comment: 10 pages, 3 figures, 2 tables
♻ ☆ Few-Shot Recognition via Stage-Wise Retrieval-Augmented Finetuning CVPR 2025
Few-shot recognition (FSR) aims to train a classification model with only a few labeled examples of each concept concerned by a downstream task, where data annotation cost can be prohibitively high. We develop methods to solve FSR by leveraging a pretrained Vision-Language Model (VLM). We particularly explore retrieval-augmented learning (RAL), which retrieves open data, e.g., the VLM's pretraining dataset, to learn models for better serving downstream tasks. RAL has been studied in zero-shot recognition but remains under-explored in FSR. Although applying RAL to FSR may seem straightforward, we observe interesting and novel challenges and opportunities. First, somewhat surprisingly, finetuning a VLM on a large amount of retrieved data underperforms state-of-the-art zero-shot methods. This is due to the imbalanced distribution of retrieved data and its domain gaps with the few-shot examples in the downstream task. Second, more surprisingly, we find that simply finetuning a VLM solely on few-shot examples significantly outperforms previous FSR methods, and finetuning on the mix of retrieved and few-shot data yields even better results. Third, to mitigate the imbalanced distribution and domain gap issues, we propose Stage-Wise retrieval-Augmented fineTuning (SWAT), which involves end-to-end finetuning on mixed data in the first stage and retraining the classifier on the few-shot data in the second stage. Extensive experiments on nine popular benchmarks demonstrate that SWAT significantly outperforms previous methods by >6% accuracy.
comment: Accepted to CVPR 2025. Website and code: https://tian1327.github.io/SWAT/
♻ ☆ Passive Heart Rate Monitoring During Smartphone Use in Everyday Life
Resting heart rate (RHR) is an important biomarker of cardiovascular health and mortality, but tracking it longitudinally generally requires a wearable device, limiting its availability. We present PHRM, a deep learning system for passive heart rate (HR) and RHR measurements during everyday smartphone use, using facial video-based photoplethysmography. Our system was developed using 225,773 videos from 495 participants and validated on 185,970 videos from 205 participants in laboratory and free-living conditions, representing the largest validation study of its kind. Compared to reference electrocardiogram, PHRM achieved a mean absolute percentage error (MAPE) < 10% for HR measurements across three skin tone groups of light, medium and dark pigmentation; MAPE for each skin tone group was non-inferior versus the others. Daily RHR measured by PHRM had a mean absolute error < 5 bpm compared to a wearable HR tracker, and was associated with known risk factors. These results highlight the potential of smartphones to enable passive and equitable heart health monitoring.
comment: Updated author list
♻ ☆ TAIJI: Textual Anchoring for Immunizing Jailbreak Images in Vision Language Models
Vision Language Models (VLMs) have demonstrated impressive inference capabilities, but remain vulnerable to jailbreak attacks that can induce harmful or unethical responses. Existing defence methods are predominantly white-box approaches that require access to model parameters and extensive modifications, making them costly and impractical for many real-world scenarios. Although some black-box defences have been proposed, they often impose input constraints or require multiple queries, limiting their effectiveness in safety-critical tasks such as autonomous driving. To address these challenges, we propose a novel black-box defence framework called \textbf{T}extual \textbf{A}nchoring for \textbf{I}mmunizing \textbf{J}ailbreak \textbf{I}mages (\textbf{TAIJI}). TAIJI leverages key phrase-based textual anchoring to enhance the model's ability to assess and mitigate the harmful content embedded within both visual and textual prompts. Unlike existing methods, TAIJI operates effectively with a single query during inference, while preserving the VLM's performance on benign tasks. Extensive experiments demonstrate that TAIJI significantly enhances the safety and reliability of VLMs, providing a practical and efficient solution for real-world deployment.
♻ ☆ RoSTE: An Efficient Quantization-Aware Supervised Fine-Tuning Approach for Large Language Models
Supervised fine-tuning is a standard method for adapting pre-trained large language models (LLMs) to downstream tasks. Quantization has been recently studied as a post-training technique for efficient LLM deployment. To obtain quantized fine-tuned LLMs, conventional pipelines would first fine-tune the pre-trained models, followed by post-training quantization. This often yields suboptimal performance as it fails to leverage the synergy between fine-tuning and quantization. To effectively realize low-bit quantization of weights, activations, and KV caches in LLMs, we propose an algorithm named Rotated Straight-Through-Estimator (RoSTE), which combines quantization-aware supervised fine-tuning (QA-SFT) with an adaptive rotation strategy that identifies an effective rotation configuration to reduce activation outliers. We provide theoretical insights on RoSTE by analyzing its prediction error when applied to an overparameterized least square quantized training problem. Our findings reveal that the prediction error is directly proportional to the quantization error of the converged weights, which can be effectively managed through an optimized rotation configuration. Experiments on Pythia, Qwen and Llama models of different sizes demonstrate the effectiveness of RoSTE. Compared to existing post-SFT quantization baselines, our method consistently achieves superior performances across various tasks and different LLM architectures.
comment: 20 pages, 7 figures
♻ ☆ RAGO: Systematic Performance Optimization for Retrieval-Augmented Generation Serving
Retrieval-augmented generation (RAG), which combines large language models (LLMs) with retrievals from external knowledge databases, is emerging as a popular approach for reliable LLM serving. However, efficient RAG serving remains an open challenge due to the rapid emergence of many RAG variants and the substantial differences in workload characteristics across them. In this paper, we make three fundamental contributions to advancing RAG serving. First, we introduce RAGSchema, a structured abstraction that captures the wide range of RAG algorithms, serving as a foundation for performance optimization. Second, we analyze several representative RAG workloads with distinct RAGSchema, revealing significant performance variability across these workloads. Third, to address this variability and meet diverse performance requirements, we propose RAGO (Retrieval-Augmented Generation Optimizer), a system optimization framework for efficient RAG serving. Our evaluation shows that RAGO achieves up to a 2x increase in QPS per chip and a 55% reduction in time-to-first-token latency compared to RAG systems built on LLM-system extensions.
comment: 16 pages, 19 figures, 4 tables
♻ ☆ Temporal-Spatial Attention Network (TSAN) for DoS Attack Detection in Network Traffic
Denial-of-Service (DoS) attacks remain a critical threat to network security, disrupting services and causing significant economic losses. Traditional detection methods, including statistical and rule-based models, struggle to adapt to evolving attack patterns. To address this challenge, we propose a novel Temporal-Spatial Attention Network (TSAN) architecture for detecting Denial of Service (DoS) attacks in network traffic. By leveraging both temporal and spatial features of network traffic, our approach captures complex traffic patterns and anomalies that traditional methods might miss. The TSAN model incorporates transformer-based temporal encoding, convolutional spatial encoding, and a cross-attention mechanism to fuse these complementary feature spaces. Additionally, we employ multi-task learning with auxiliary tasks to enhance the model's robustness. Experimental results on the NSL-KDD dataset demonstrate that TSAN outperforms state-of-the-art models, achieving superior accuracy, precision, recall, and F1-score while maintaining computational efficiency for real-time deployment. The proposed architecture offers an optimal balance between detection accuracy and computational overhead, making it highly suitable for real-world network security applications.
comment: 19 Pages, 5 figures
♻ ☆ GreenIQ: A Deep Search Platform for Comprehensive Carbon Market Analysis and Automated Report Generation
This study introduces GreenIQ, an AI-powered deep search platform designed to revolutionise carbon market intelligence through autonomous analysis and automated report generation. Carbon markets operate across diverse regulatory landscapes, generating vast amounts of heterogeneous data from policy documents, industry reports, academic literature, and real-time trading platforms. Traditional research approaches remain labour-intensive, slow, and difficult to scale. GreenIQ addresses these limitations through a multi-agent architecture powered by Large Language Models (LLMs), integrating five specialised AI agents: a Main Researcher Agent for intelligent information retrieval, a Report Writing Agent for structured synthesis, a Final Reviewer Agent for accuracy verification, a Data Visualisation Agent for enhanced interpretability, and a Translator Agent for multilingual adaptation. The system achieves seamless integration of structured and unstructured information with AI-driven citation verification, ensuring high transparency and reliability. GreenIQ delivers a 99.2\% reduction in processing time and a 99.7\% cost reduction compared to traditional research methodologies. A novel AI persona-based evaluation framework involving 16 domain-specific AI personas highlights its superior cross-jurisdictional analytical capabilities and regulatory insight generation. GreenIQ sets new standards in AI-driven research synthesis, policy analysis, and sustainability finance by streamlining carbon market research. It offers an efficient and scalable framework for environmental and financial intelligence, enabling more accurate, timely, and cost-effective decision-making in complex regulatory landscapes
comment: 12 Pages, 1 figure
♻ ☆ DINO-LG: A Task-Specific DINO Model for Coronary Calcium Scoring
Coronary artery disease (CAD), one of the leading causes of mortality worldwide, necessitates effective risk assessment strategies, with coronary artery calcium (CAC) scoring via computed tomography (CT) being a key method for prevention. Traditional methods, primarily based on UNET architectures implemented on pre-built models, face challenges like the scarcity of annotated CT scans containing CAC and imbalanced datasets, leading to reduced performance in segmentation and scoring tasks. In this study, we address these limitations by incorporating the self-supervised learning (SSL) technique of DINO (self-distillation with no labels), which trains without requiring CAC-specific annotations, enhancing its robustness in generating distinct features. The DINO-LG model, which leverages label guidance to focus on calcified areas, achieves significant improvements, with a sensitivity of 89% and specificity of 90% for detecting CAC-containing CT slices, compared to the standard DINO model's sensitivity of 79% and specificity of 77%. Additionally, false-negative and false-positive rates are reduced by 49% and 59%, respectively, instilling greater confidence in clinicians when ruling out calcification in low-risk patients and minimizing unnecessary imaging reviews by radiologists. Further, CAC scoring and segmentation tasks are conducted using a basic UNET architecture, applied specifically to CT slices identified by the DINO-LG model as containing calcified areas. This targeted approach enhances CAC scoring accuracy by feeding the UNET model with relevant slices, significantly improving diagnostic precision, reducing both false positives and false negatives, and ultimately lowering overall healthcare costs by minimizing unnecessary tests and treatments, presenting a valuable advancement in CAD risk assessment.
comment: Developed by Center for Applied Artificial Intelligence (CAAI), University of Kentucky
♻ ☆ ModServe: Scalable and Resource-Efficient Large Multimodal Model Serving
Large multimodal models (LMMs) demonstrate impressive capabilities in understanding images, videos, and audio beyond text. However, efficiently serving LMMs in production environments poses significant challenges due to their complex architectures and heterogeneous characteristics across their multi-stage inference pipelines. We present the first comprehensive systems analysis of two prominent LMM architectures, decoder-only and cross-attention, across six representative open-source models, revealing key systems design implications. We also present an in-depth analysis of production LMM inference traces, uncovering unique workload characteristics, including variable, heavy-tailed request distributions and bursty traffic patterns. Based on these insights, we propose ModServe, a modular LMM serving system that decouples stages for independent optimization and adaptive scaling. ModServe dynamically reconfigures stages and handles bursty traffic with modality-aware scheduling and autoscaling to meet tail latency SLOs while minimizing costs. ModServe achieves 3.3-5.5x higher throughput (leading to 25-41.3% cost saving) while meeting SLOs on a 128-GPU cluster with production traces.
♻ ☆ Graph Masked Language Models
Language Models (LMs) and Graph Neural Networks (GNNs) have shown great promise in their respective areas, yet integrating structured graph data with rich textual information remains challenging. In this work, we propose \emph{Graph Masked Language Models} (GMLM), a novel dual-branch architecture that combines the structural learning of GNNs with the contextual power of pretrained language models. Our approach introduces two key innovations: (i) a \emph{semantic masking strategy} that leverages graph topology to selectively mask nodes based on their structural importance, and (ii) a \emph{soft masking mechanism} that interpolates between original node features and a learnable mask token, ensuring smoother information flow during training. Extensive experiments on multiple node classification and language understanding benchmarks demonstrate that GMLM not only achieves state-of-the-art performance but also exhibits enhanced robustness and stability. This work underscores the benefits of integrating structured and unstructured data representations for improved graph learning.
♻ ☆ Bridging Technology and Humanities: Evaluating the Impact of Large Language Models on Social Sciences Research with DeepSeek-R1
In recent years, the development of Large Language Models (LLMs) has made significant breakthroughs in the field of natural language processing and has gradually been applied to the field of humanities and social sciences research. LLMs have a wide range of application value in the field of humanities and social sciences because of its strong text understanding, generation and reasoning capabilities. In humanities and social sciences research, LLMs can analyze large-scale text data and make inferences. This article analyzes the large language model DeepSeek-R1 from seven aspects: low-resource language translation, educational question-answering, student writing improvement in higher education, logical reasoning, educational measurement and psychometrics, public health policy analysis, and art education.Then we compare the answers given by DeepSeek-R1 in the seven aspects with the answers given by o1-preview. DeepSeek-R1 performs well in the humanities and social sciences, answering most questions correctly and logically, and can give reasonable analysis processes and explanations. Compared with o1-preview, it can automatically generate reasoning processes and provide more detailed explanations, which is suitable for beginners or people who need to have a detailed understanding of this knowledge, while o1-preview is more suitable for quick reading. Through analysis, it is found that LLM has broad application potential in the field of humanities and social sciences, and shows great advantages in improving text analysis efficiency, language communication and other fields. LLM's powerful language understanding and generation capabilities enable it to deeply explore complex problems in the field of humanities and social sciences, and provide innovative tools for academic research and practical applications.
comment: 52 pages, 19 figures
♻ ☆ Multi-Aggregator Time-Warping Heterogeneous Graph Neural Network for Personalized Micro-Video Recommendation
Micro-video recommendation is attracting global attention and becoming a popular daily service for people of all ages. Recently, Graph Neural Networks-based micro-video recommendation has displayed performance improvement for many kinds of recommendation tasks. However, the existing works fail to fully consider the characteristics of micro-videos, such as the high timeliness of news nature micro-video recommendation and sequential interactions of frequently changed interests. In this paper, a novel Multi-aggregator Time-warping Heterogeneous Graph Neural Network (MTHGNN) is proposed for personalized news nature micro-video recommendation based on sequential sessions, where characteristics of micro-videos are comprehensively studied, users' preference is mined via multi-aggregator, the temporal and dynamic changes of users' preference are captured, and timeliness is considered. Through the comparison with the state-of-the-arts, the experimental results validate the superiority of our MTHGNN model.
♻ ☆ TruthPrInt: Mitigating LVLM Object Hallucination Via Latent Truthful-Guided Pre-Intervention
Object Hallucination (OH) has been acknowledged as one of the major trustworthy challenges in Large Vision-Language Models (LVLMs). Recent advancements in Large Language Models (LLMs) indicate that internal states, such as hidden states, encode the "overall truthfulness" of generated responses. However, it remains under-explored how internal states in LVLMs function and whether they could serve as "per-token" hallucination indicators, which is essential for mitigating OH. In this paper, we first conduct an in-depth exploration of LVLM internal states in relation to OH issues and discover that (1) LVLM internal states are high-specificity per-token indicators of hallucination behaviors. Moreover, (2) different LVLMs encode universal patterns of hallucinations in common latent subspaces, indicating that there exist "generic truthful directions" shared by various LVLMs. Based on these discoveries, we propose Truthful-Guided Pre-Intervention (TruthPrInt) that first learns the truthful direction of LVLM decoding and then applies truthful-guided inference-time intervention during LVLM decoding. We further propose ComnHallu to enhance both cross-LVLM and cross-data hallucination detection transferability by constructing and aligning hallucination latent subspaces. We evaluate TruthPrInt in extensive experimental settings, including in-domain and out-of-domain scenarios, over popular LVLMs and OH benchmarks. Experimental results indicate that TruthPrInt significantly outperforms state-of-the-art methods. Codes will be available at https://github.com/jinhaoduan/TruthPrInt.
comment: 15 pages, 9 figures, the first two authors contributed equally
♻ ☆ LazyDiT: Lazy Learning for the Acceleration of Diffusion Transformers AAAI 2025
Diffusion Transformers have emerged as the preeminent models for a wide array of generative tasks, demonstrating superior performance and efficacy across various applications. The promising results come at the cost of slow inference, as each denoising step requires running the whole transformer model with a large amount of parameters. In this paper, we show that performing the full computation of the model at each diffusion step is unnecessary, as some computations can be skipped by lazily reusing the results of previous steps. Furthermore, we show that the lower bound of similarity between outputs at consecutive steps is notably high, and this similarity can be linearly approximated using the inputs. To verify our demonstrations, we propose the \textbf{LazyDiT}, a lazy learning framework that efficiently leverages cached results from earlier steps to skip redundant computations. Specifically, we incorporate lazy learning layers into the model, effectively trained to maximize laziness, enabling dynamic skipping of redundant computations. Experimental results show that LazyDiT outperforms the DDIM sampler across multiple diffusion transformer models at various resolutions. Furthermore, we implement our method on mobile devices, achieving better performance than DDIM with similar latency. Code: https://github.com/shawnricecake/lazydit
comment: Accepted by AAAI 2025
♻ ☆ RadioActive: 3D Radiological Interactive Segmentation Benchmark
Effortless and precise segmentation with minimal clinician effort could greatly streamline clinical workflows. Recent interactive segmentation models, inspired by METAs Segment Anything, have made significant progress but face critical limitations in 3D radiology. These include impractical human interaction requirements such as slice-by-slice operations for 2D models on 3D data and a lack of iterative refinement. Prior studies have been hindered by inadequate evaluation protocols, resulting in unreliable performance assessments and inconsistent findings across studies. The RadioActive benchmark addresses these challenges by providing a rigorous and reproducible evaluation framework for interactive segmentation methods in clinically relevant scenarios. It features diverse datasets, a wide range of target structures, and the most impactful 2D and 3D interactive segmentation methods, all within a flexible and extensible codebase. We also introduce advanced prompting techniques that reduce interaction steps, enabling fair comparisons between 2D and 3D models. Surprisingly, SAM2 outperforms all specialized medical 2D and 3D models in a setting requiring only a few interactions to generate prompts for a 3D volume. This challenges prevailing assumptions and demonstrates that general-purpose models surpass specialized medical approaches. By open-sourcing RadioActive, we invite researchers to integrate their models and prompting techniques, ensuring continuous and transparent evaluation of 3D medical interactive models.
comment: Undergoing Peer-Review
♻ ☆ GREEN-CODE: Learning to Optimize Energy Efficiency in LLM-based Code Generation
Large Language Models (LLMs) are becoming integral to daily life, showcasing their vast potential across various Natural Language Processing (NLP) tasks. Beyond NLP, LLMs are increasingly used in software development tasks, such as code completion, modification, bug fixing, and code translation. Software engineers widely use tools like GitHub Copilot and Amazon Q, streamlining workflows and automating tasks with high accuracy. While the resource and energy intensity of LLM training is often highlighted, inference can be even more resource-intensive over time, as it's a continuous process with a high number of invocations. Therefore, developing resource-efficient alternatives for LLM inference is crucial for sustainability. This work proposes GREEN-CODE, a framework for energy-aware code generation in LLMs. GREEN-CODE performs dynamic early exit during LLM inference. We train a Reinforcement Learning (RL) agent that learns to balance the trade-offs between accuracy, latency, and energy consumption. Our approach is evaluated on two open-source LLMs, Llama 3.2 3B and OPT 2.7B, using the JavaCorpus and PY150 datasets. Results show that our method reduces the energy consumption between 23-50 % on average for code generation tasks without significantly affecting accuracy.
comment: Under submission in ACM/IEEE conference, 11 pages
♻ ☆ Uncertainty modeling for fine-tuned implicit functions
Implicit functions such as Neural Radiance Fields (NeRFs), occupancy networks, and signed distance functions (SDFs) have become pivotal in computer vision for reconstructing detailed object shapes from sparse views. Achieving optimal performance with these models can be challenging due to the extreme sparsity of inputs and distribution shifts induced by data corruptions. To this end, large, noise-free synthetic datasets can serve as shape priors to help models fill in gaps, but the resulting reconstructions must be approached with caution. Uncertainty estimation is crucial for assessing the quality of these reconstructions, particularly in identifying areas where the model is uncertain about the parts it has inferred from the prior. In this paper, we introduce Dropsembles, a novel method for uncertainty estimation in tuned implicit functions. We demonstrate the efficacy of our approach through a series of experiments, starting with toy examples and progressing to a real-world scenario. Specifically, we train a Convolutional Occupancy Network on synthetic anatomical data and test it on low-resolution MRI segmentations of the lumbar spine. Our results show that Dropsembles achieve the accuracy and calibration levels of deep ensembles but with significantly less computational cost.
♻ ☆ LitLLMs, LLMs for Literature Review: Are we there yet?
Literature reviews are an essential component of scientific research, but they remain time-intensive and challenging to write, especially due to the recent influx of research papers. This paper explores the zero-shot abilities of recent Large Language Models (LLMs) in assisting with the writing of literature reviews based on an abstract. We decompose the task into two components: 1. Retrieving related works given a query abstract, and 2. Writing a literature review based on the retrieved results. We analyze how effective LLMs are for both components. For retrieval, we introduce a novel two-step search strategy that first uses an LLM to extract meaningful keywords from the abstract of a paper and then retrieves potentially relevant papers by querying an external knowledge base. Additionally, we study a prompting-based re-ranking mechanism with attribution and show that re-ranking doubles the normalized recall compared to naive search methods, while providing insights into the LLM's decision-making process. In the generation phase, we propose a two-step approach that first outlines a plan for the review and then executes steps in the plan to generate the actual review. To evaluate different LLM-based literature review methods, we create test sets from arXiv papers using a protocol designed for rolling use with newly released LLMs to avoid test set contamination in zero-shot evaluations. We release this evaluation protocol to promote additional research and development in this regard. Our empirical results suggest that LLMs show promising potential for writing literature reviews when the task is decomposed into smaller components of retrieval and planning. Our project page including a demonstration system and toolkit can be accessed here: https://litllm.github.io.
♻ ☆ Advanced Deep Learning Methods for Protein Structure Prediction and Design
After AlphaFold won the Nobel Prize, protein prediction with deep learning once again became a hot topic. We comprehensively explore advanced deep learning methods applied to protein structure prediction and design. It begins by examining recent innovations in prediction architectures, with detailed discussions on improvements such as diffusion based frameworks and novel pairwise attention modules. The text analyses key components including structure generation, evaluation metrics, multiple sequence alignment processing, and network architecture, thereby illustrating the current state of the art in computational protein modelling. Subsequent chapters focus on practical applications, presenting case studies that range from individual protein predictions to complex biomolecular interactions. Strategies for enhancing prediction accuracy and integrating deep learning techniques with experimental validation are thoroughly explored. The later sections review the industry landscape of protein design, highlighting the transformative role of artificial intelligence in biotechnology and discussing emerging market trends and future challenges. Supplementary appendices provide essential resources such as databases and open source tools, making this volume a valuable reference for researchers and students.
♻ ☆ Code-as-Monitor: Constraint-aware Visual Programming for Reactive and Proactive Robotic Failure Detection CVPR 2025
Automatic detection and prevention of open-set failures are crucial in closed-loop robotic systems. Recent studies often struggle to simultaneously identify unexpected failures reactively after they occur and prevent foreseeable ones proactively. To this end, we propose Code-as-Monitor (CaM), a novel paradigm leveraging the vision-language model (VLM) for both open-set reactive and proactive failure detection. The core of our method is to formulate both tasks as a unified set of spatio-temporal constraint satisfaction problems and use VLM-generated code to evaluate them for real-time monitoring. To enhance the accuracy and efficiency of monitoring, we further introduce constraint elements that abstract constraint-related entities or their parts into compact geometric elements. This approach offers greater generality, simplifies tracking, and facilitates constraint-aware visual programming by leveraging these elements as visual prompts. Experiments show that CaM achieves a 28.7% higher success rate and reduces execution time by 31.8% under severe disturbances compared to baselines across three simulators and a real-world setting. Moreover, CaM can be integrated with open-loop control policies to form closed-loop systems, enabling long-horizon tasks in cluttered scenes with dynamic environments.
comment: Accepted by CVPR 2025. Project page: https://zhoues.github.io/Code-as-Monitor/
♻ ☆ Algorithmic causal structure emerging through compression
We explore the relationship between causality, symmetry, and compression. We build on and generalize the known connection between learning and compression to a setting where causal models are not identifiable. We propose a framework where causality emerges as a consequence of compressing data across multiple environments. We define algorithmic causality as an alternative definition of causality when traditional assumptions for causal identifiability do not hold. We demonstrate how algorithmic causal and symmetric structures can emerge from minimizing upper bounds on Kolmogorov complexity, without knowledge of intervention targets. We hypothesize that these insights may also provide a novel perspective on the emergence of causality in machine learning models, such as large language models, where causal relationships may not be explicitly identifiable.
comment: Extended version of the camera-ready paper accepted at CLeaR 2025
♻ ☆ LitLLM: A Toolkit for Scientific Literature Review
Conducting literature reviews for scientific papers is essential for understanding research, its limitations, and building on existing work. It is a tedious task which makes an automatic literature review generator appealing. Unfortunately, many existing works that generate such reviews using Large Language Models (LLMs) have significant limitations. They tend to hallucinate-generate non-factual information-and ignore the latest research they have not been trained on. To address these limitations, we propose a toolkit that operates on Retrieval Augmented Generation (RAG) principles, specialized prompting and instructing techniques with the help of LLMs. Our system first initiates a web search to retrieve relevant papers by summarizing user-provided abstracts into keywords using an off-the-shelf LLM. Authors can enhance the search by supplementing it with relevant papers or keywords, contributing to a tailored retrieval process. Second, the system re-ranks the retrieved papers based on the user-provided abstract. Finally, the related work section is generated based on the re-ranked results and the abstract. There is a substantial reduction in time and effort for literature review compared to traditional methods, establishing our toolkit as an efficient alternative. Our project page including the demo and toolkit can be accessed here: https://litllm.github.io
♻ ☆ Generation of Geodesics with Actor-Critic Reinforcement Learning to Predict Midpoints
To find the shortest paths for all pairs on manifolds with infinitesimally defined metrics, we introduce a framework to generate them by predicting midpoints recursively. To learn midpoint prediction, we propose an actor-critic approach. We prove the soundness of our approach and show experimentally that the proposed method outperforms existing methods on several planning tasks, including path planning for agents with complex kinematics and motion planning for multi-degree-of-freedom robot arms.
comment: 17 pages with 8 pages of appendices and references, 9 figures
♻ ☆ Zero-Shot Reinforcement Learning via Function Encoders
Although reinforcement learning (RL) can solve many challenging sequential decision making problems, achieving zero-shot transfer across related tasks remains a challenge. The difficulty lies in finding a good representation for the current task so that the agent understands how it relates to previously seen tasks. To achieve zero-shot transfer, we introduce the function encoder, a representation learning algorithm which represents a function as a weighted combination of learned, non-linear basis functions. By using a function encoder to represent the reward function or the transition function, the agent has information on how the current task relates to previously seen tasks via a coherent vector representation. Thus, the agent is able to achieve transfer between related tasks at run time with no additional training. We demonstrate state-of-the-art data efficiency, asymptotic performance, and training stability in three RL fields by augmenting basic RL algorithms with a function encoder task representation.
comment: A critical issue was found in the multi-agent experiments published in version 2. We rerun the multi-agent experiments on a more challenging, partially observable Markov game
♻ ☆ GiVE: Guiding Visual Encoder to Perceive Overlooked Information ICME 2025
Multimodal Large Language Models have advanced AI in applications like text-to-video generation and visual question answering. These models rely on visual encoders to convert non-text data into vectors, but current encoders either lack semantic alignment or overlook non-salient objects. We propose the Guiding Visual Encoder to Perceive Overlooked Information (GiVE) approach. GiVE enhances visual representation with an Attention-Guided Adapter (AG-Adapter) module and an Object-focused Visual Semantic Learning module. These incorporate three novel loss terms: Object-focused Image-Text Contrast (OITC) loss, Object-focused Image-Image Contrast (OIIC) loss, and Object-focused Image Discrimination (OID) loss, improving object consideration, retrieval accuracy, and comprehensiveness. Our contributions include dynamic visual focus adjustment, novel loss functions to enhance object retrieval, and the Multi-Object Instruction (MOInst) dataset. Experiments show our approach achieves state-of-the-art performance.
comment: This paper was accepted by ICME 2025
♻ ☆ Learning Robust Reward Machines from Noisy Labels KR 2024
This paper presents PROB-IRM, an approach that learns robust reward machines (RMs) for reinforcement learning (RL) agents from noisy execution traces. The key aspect of RM-driven RL is the exploitation of a finite-state machine that decomposes the agent's task into different subtasks. PROB-IRM uses a state-of-the-art inductive logic programming framework robust to noisy examples to learn RMs from noisy traces using the Bayesian posterior degree of beliefs, thus ensuring robustness against inconsistencies. Pivotal for the results is the interleaving between RM learning and policy learning: a new RM is learned whenever the RL agent generates a trace that is believed not to be accepted by the current RM. To speed up the training of the RL agent, PROB-IRM employs a probabilistic formulation of reward shaping that uses the posterior Bayesian beliefs derived from the traces. Our experimental analysis shows that PROB-IRM can learn (potentially imperfect) RMs from noisy traces and exploit them to train an RL agent to solve its tasks successfully. Despite the complexity of learning the RM from noisy traces, agents trained with PROB-IRM perform comparably to agents provided with handcrafted RMs.
comment: Accepted at the 21st International Conference on Principles of Knowledge Representation and Reasoning (KR 2024)
♻ ☆ Causality Is Key to Understand and Balance Multiple Goals in Trustworthy ML and Foundation Models
Ensuring trustworthiness in machine learning (ML) systems is crucial as they become increasingly embedded in high-stakes domains. This paper advocates for integrating causal methods into machine learning to navigate the trade-offs among key principles of trustworthy ML, including fairness, privacy, robustness, accuracy, and explainability. While these objectives should ideally be satisfied simultaneously, they are often addressed in isolation, leading to conflicts and suboptimal solutions. Drawing on existing applications of causality in ML that successfully align goals such as fairness and accuracy or privacy and robustness, this paper argues that a causal approach is essential for balancing multiple competing objectives in both trustworthy ML and foundation models. Beyond highlighting these trade-offs, we examine how causality can be practically integrated into ML and foundation models, offering solutions to enhance their reliability and interpretability. Finally, we discuss the challenges, limitations, and opportunities in adopting causal frameworks, paving the way for more accountable and ethically sound AI systems.
♻ ☆ Instant Adversarial Purification with Adversarial Consistency Distillation CVPR2025
Neural networks have revolutionized numerous fields with their exceptional performance, yet they remain susceptible to adversarial attacks through subtle perturbations. While diffusion-based purification methods like DiffPure offer promising defense mechanisms, their computational overhead presents a significant practical limitation. In this paper, we introduce One Step Control Purification (OSCP), a novel defense framework that achieves robust adversarial purification in a single Neural Function Evaluation (NFE) within diffusion models. We propose Gaussian Adversarial Noise Distillation (GAND) as the distillation objective and Controlled Adversarial Purification (CAP) as the inference pipeline, which makes OSCP demonstrate remarkable efficiency while maintaining defense efficacy. Our proposed GAND addresses a fundamental tension between consistency distillation and adversarial perturbation, bridging the gap between natural and adversarial manifolds in the latent space, while remaining computationally efficient through Parameter-Efficient Fine-Tuning (PEFT) methods such as LoRA, eliminating the high computational budget request from full parameter fine-tuning. The CAP guides the purification process through the unlearnable edge detection operator calculated by the input image as an extra prompt, effectively preventing the purified images from deviating from their original appearance when large purification steps are used. Our experimental results on ImageNet showcase OSCP's superior performance, achieving a 74.19% defense success rate with merely 0.1s per purification -- a 100-fold speedup compared to conventional approaches.
comment: Accepted by CVPR2025
♻ ☆ Embedded Visual Prompt Tuning
Foundation models pre-trained on large-scale data have been widely witnessed to achieve success in various natural imaging downstream tasks. Parameter-efficient fine-tuning (PEFT) methods aim to adapt foundation models to new domains by updating only a small portion of parameters in order to reduce computational overhead. However, the effectiveness of these PEFT methods, especially in cross-domain few-shot scenarios, e.g., medical image analysis, has not been fully explored. In this work, we facilitate the study of the performance of PEFT when adapting foundation models to medical image classification tasks. Furthermore, to alleviate the limitations of prompt introducing ways and approximation capabilities on Transformer architectures of mainstream prompt tuning methods, we propose the Embedded Prompt Tuning (EPT) method by embedding prompt tokens into the expanded channels. We also find that there are anomalies in the feature space distribution of foundation models during pre-training process, and prompt tuning can help mitigate this negative impact. To explain this phenomenon, we also introduce a novel perspective to understand prompt tuning: Prompt tuning is a distribution calibrator. And we support it by analyzing patch-wise scaling and feature separation operations contained in EPT. Our experiments show that EPT outperforms several state-of-the-art fine-tuning methods by a significant margin on few-shot medical image classification tasks, and completes the fine-tuning process within highly competitive time, indicating EPT is an effective PEFT method. The source code is available at github.com/zuwenqiang/EPT.
♻ ☆ Autonomous AI imitators increase diversity in homogeneous information ecosystems
Recent breakthroughs in large language models (LLMs) have facilitated autonomous AI agents capable of imitating human-generated content. This technological advancement raises fundamental questions about AI's impact on the diversity and democratic value of information ecosystems. We introduce a large-scale simulation framework to examine AI-based imitation within news, a context crucial for public discourse. By systematically testing two distinct imitation strategies across a range of information environments varying in initial diversity, we demonstrate that AI-generated articles do not uniformly homogenize content. Instead, AI's influence is strongly context-dependent: AI-generated content can introduce valuable diversity in originally homogeneous news environments but diminish diversity in initially heterogeneous contexts. These results illustrate that the initial diversity of an information environment critically shapes AI's impact, challenging assumptions that AI-driven imitation uniformly threatens diversity. Instead, when information is initially homogeneous, AI-driven imitation can expand perspectives, styles, and topics. This is especially important in news contexts, where information diversity fosters richer public debate by exposing citizens to alternative viewpoints, challenging biases, and preventing narrative monopolies, which is essential for a resilient democracy.
comment: 35 pages, 10 figures, 4 tables; v2: corrected typographical errors, streamlined language, updated abstract, added supplementary information
♻ ☆ OptionZero: Planning with Learned Options ICLR 2025
Planning with options -- a sequence of primitive actions -- has been shown effective in reinforcement learning within complex environments. Previous studies have focused on planning with predefined options or learned options through expert demonstration data. Inspired by MuZero, which learns superhuman heuristics without any human knowledge, we propose a novel approach, named OptionZero. OptionZero incorporates an option network into MuZero, providing autonomous discovery of options through self-play games. Furthermore, we modify the dynamics network to provide environment transitions when using options, allowing searching deeper under the same simulation constraints. Empirical experiments conducted in 26 Atari games demonstrate that OptionZero outperforms MuZero, achieving a 131.58% improvement in mean human-normalized score. Our behavior analysis shows that OptionZero not only learns options but also acquires strategic skills tailored to different game characteristics. Our findings show promising directions for discovering and using options in planning. Our code is available at https://rlg.iis.sinica.edu.tw/papers/optionzero.
comment: Accepted by the Thirteenth International Conference on Learning Representations (ICLR 2025) as oral presentation
♻ ☆ Dynamic-LLaVA: Efficient Multimodal Large Language Models via Dynamic Vision-language Context Sparsification ICLR 2025
Multimodal Large Language Models (MLLMs) have achieved remarkable success in vision understanding, reasoning, and interaction. However, the inference computation and memory increase progressively with the generation of output tokens during decoding, directly affecting the efficacy of MLLMs. Existing methods attempt to reduce the vision context redundancy to achieve efficient MLLMs. Unfortunately, the efficiency benefits of the vision context reduction in the prefill stage gradually diminish during the decoding stage. To address this problem, we proposed a dynamic vision-language context sparsification framework Dynamic-LLaVA, which dynamically reduces the redundancy of vision context in the prefill stage and decreases the memory and computation overhead of the generated language context during decoding. Dynamic-LLaVA designs a tailored sparsification inference scheme for different inference modes, i.e., prefill, decoding with and without KV cache, to achieve efficient inference of MLLMs. In practice, Dynamic-LLaVA can reduce computation consumption by $\sim$75\% in the prefill stage. Meanwhile, throughout the entire generation process of MLLMs, Dynamic-LLaVA reduces the $\sim$50\% computation consumption under decoding without KV cache, while saving $\sim$50\% GPU memory overhead when decoding with KV cache, due to the vision-language context sparsification. Extensive experiments also demonstrate that Dynamic-LLaVA achieves efficient inference for MLLMs with negligible understanding and generation ability degradation or even performance gains compared to the full-context inference baselines. Code is available at https://github.com/Osilly/dynamic_llava .
comment: Accepted to ICLR 2025. Code is available at https://github.com/Osilly/dynamic_llava
♻ ☆ Optimizing Attention with Mirror Descent: Generalized Max-Margin Token Selection
Attention mechanisms have revolutionized several domains of artificial intelligence, such as natural language processing and computer vision, by enabling models to selectively focus on relevant parts of the input data. While recent work has characterized the optimization dynamics of gradient descent (GD) in attention-based models and the structural properties of its preferred solutions, less is known about more general optimization algorithms such as mirror descent (MD). In this paper, we investigate the convergence properties and implicit biases of a family of MD algorithms tailored for softmax attention mechanisms, with the potential function chosen as the $p$-th power of the $\ell_p$-norm. Specifically, we show that these algorithms converge in direction to a generalized hard-margin SVM with an $\ell_p$-norm objective when applied to a classification problem using a softmax attention model. Notably, our theoretical results reveal that the convergence rate is comparable to that of traditional GD in simpler models, despite the highly nonlinear and nonconvex nature of the present problem. Additionally, we delve into the joint optimization dynamics of the key-query matrix and the decoder, establishing conditions under which this complex joint optimization converges to their respective hard-margin SVM solutions. Lastly, our numerical experiments on real data demonstrate that MD algorithms improve generalization over standard GD and excel in optimal token selection.
♻ ☆ SATA: A Paradigm for LLM Jailbreak via Simple Assistive Task Linkage
Large language models (LLMs) have made significant advancements across various tasks, but their safety alignment remain a major concern. Exploring jailbreak prompts can expose LLMs' vulnerabilities and guide efforts to secure them. Existing methods primarily design sophisticated instructions for the LLM to follow, or rely on multiple iterations, which could hinder the performance and efficiency of jailbreaks. In this work, we propose a novel jailbreak paradigm, Simple Assistive Task Linkage (SATA), which can effectively circumvent LLM safeguards and elicit harmful responses. Specifically, SATA first masks harmful keywords within a malicious query to generate a relatively benign query containing one or multiple [MASK] special tokens. It then employs a simple assistive task such as a masked language model task or an element lookup by position task to encode the semantics of the masked keywords. Finally, SATA links the assistive task with the masked query to jointly perform the jailbreak. Extensive experiments show that SATA achieves state-of-the-art performance and outperforms baselines by a large margin. Specifically, on AdvBench dataset, with mask language model (MLM) assistive task, SATA achieves an overall attack success rate (ASR) of 85% and harmful score (HS) of 4.57, and with element lookup by position (ELP) assistive task, SATA attains an overall ASR of 76% and HS of 4.43.
♻ ☆ A General Adaptive Dual-level Weighting Mechanism for Remote Sensing Pansharpening CVPR
Currently, deep learning-based methods for remote sensing pansharpening have advanced rapidly. However, many existing methods struggle to fully leverage feature heterogeneity and redundancy, thereby limiting their effectiveness. We use the covariance matrix to model the feature heterogeneity and redundancy and propose Correlation-Aware Covariance Weighting (CACW) to adjust them. CACW captures these correlations through the covariance matrix, which is then processed by a nonlinear function to generate weights for adjustment. Building upon CACW, we introduce a general adaptive dual-level weighting mechanism (ADWM) to address these challenges from two key perspectives, enhancing a wide range of existing deep-learning methods. First, Intra-Feature Weighting (IFW) evaluates correlations among channels within each feature to reduce redundancy and enhance unique information. Second, Cross-Feature Weighting (CFW) adjusts contributions across layers based on inter-layer correlations, refining the final output. Extensive experiments demonstrate the superior performance of ADWM compared to recent state-of-the-art (SOTA) methods. Furthermore, we validate the effectiveness of our approach through generality experiments, redundancy visualization, comparison experiments, key variables and complexity analysis, and ablation studies. Our code is available at https://github.com/Jie-1203/ADWM.
comment: This paper is accepted at the CVPR Conference on Computer Vision and Pattern Recognition 2025
♻ ☆ NotaGen: Advancing Musicality in Symbolic Music Generation with Large Language Model Training Paradigms
We introduce NotaGen, a symbolic music generation model aiming to explore the potential of producing high-quality classical sheet music. Inspired by the success of Large Language Models (LLMs), NotaGen adopts pre-training, fine-tuning, and reinforcement learning paradigms (henceforth referred to as the LLM training paradigms). It is pre-trained on 1.6M pieces of music in ABC notation, and then fine-tuned on approximately 9K high-quality classical compositions conditioned on "period-composer-instrumentation" prompts. For reinforcement learning, we propose the CLaMP-DPO method, which further enhances generation quality and controllability without requiring human annotations or predefined rewards. Our experiments demonstrate the efficacy of CLaMP-DPO in symbolic music generation models with different architectures and encoding schemes. Furthermore, subjective A/B tests show that NotaGen outperforms baseline models against human compositions, greatly advancing musical aesthetics in symbolic music generation.
♻ ☆ Controlled Low-Rank Adaptation with Subspace Regularization for Continued Training on Large Language Models
Large language models (LLMs) exhibit remarkable capabilities in natural language processing but face catastrophic forgetting when learning new tasks, where adaptation to a new domain leads to a substantial decline in performance on previous tasks. In this paper, we propose Controlled LoRA (CLoRA), a sub-space regularization method on LoRA structure. Aiming to reduce the scale of output change while introduce minimal constraint on model capacity, CLoRA imposes constraint on the direction of updating matrix's null space. Experimental results on one-stage LLM finetuning tasks and continual learning settings highlight the superority of CLoRA as a effective parameter efficient finetuning method with catastrophic forgetting mitigating.Further investigation for model parameters indicates that CLoRA effectively balances the trade-off between model capacity and degree of forgetting.
♻ ☆ DITTO: Offline Imitation Learning with World Models
For imitation learning algorithms to scale to real-world challenges, they must handle high-dimensional observations, offline learning, and policy-induced covariate-shift. We propose DITTO, an offline imitation learning algorithm which addresses all three of these problems. DITTO optimizes a novel distance metric in the latent space of a learned world model: First, we train a world model on all available trajectory data, then, the imitation agent is unrolled from expert start states in the learned model, and penalized for its latent divergence from the expert dataset over multiple time steps. We optimize this multi-step latent divergence using standard reinforcement learning algorithms, which provably induces imitation learning, and empirically achieves state-of-the art performance and sample efficiency on a range of Atari environments from pixels, without any online environment access. We also adapt other standard imitation learning algorithms to the world model setting, and show that this considerably improves their performance. Our results show how creative use of world models can lead to a simple, robust, and highly-performant policy-learning framework.
♻ ☆ Knowledge Transfer based Evolutionary Deep Neural Network for Intelligent Fault Diagnosis
A faster response with commendable accuracy in intelligent systems is essential for the reliability and smooth operations of industrial machines. Two main challenges affect the design of such intelligent systems: (i) the selection of a suitable model and (ii) domain adaptation if there is a continuous change in operating conditions. Therefore, we propose an evolutionary Net2Net transformation (EvoN2N) that finds the best suitable DNN architecture with limited availability of labeled data samples. Net2Net transformation-based quick learning algorithm has been used in the evolutionary framework of Non-dominated sorting genetic algorithm II to obtain the best DNN architecture. Net2Net transformation-based quick learning algorithm uses the concept of knowledge transfer from one generation to the next for faster fitness evaluation. The proposed framework can obtain the best model for intelligent fault diagnosis without a long and time-consuming search process. The proposed framework has been validated on the Case Western Reserve University dataset, the Paderborn University dataset, and the gearbox fault detection dataset under different operating conditions. The best models obtained are capable of demonstrating an excellent diagnostic performance and classification accuracy of almost up to 100% for most of the operating conditions.
comment: Submitted to IEEE Transactions on Sustainable Computing
♻ ☆ Inverting Transformer-based Vision Models
Understanding the mechanisms underlying deep neural networks in computer vision remains a fundamental challenge. While many previous approaches have focused on visualizing intermediate representations within deep neural networks, particularly convolutional neural networks, these techniques have yet to be thoroughly explored in transformer-based vision models. In this study, we apply a modular approach of training inverse models to reconstruct input images from intermediate layers within a Detection Transformer and a Vision Transformer, showing that this approach is efficient and feasible. Through qualitative and quantitative evaluations of reconstructed images, we generate insights into the underlying mechanisms of these architectures, highlighting their similarities and differences in terms of contextual shape and preservation of image details, inter-layer correlation, and robustness to color perturbations. Our analysis illustrates how these properties emerge within the models, contributing to a deeper understanding of transformer-based vision models. The code for reproducing our experiments is available at github.com/wiskott-lab/inverse-detection-transformer.
♻ ☆ Application of linear regression method to the deep reinforcement learning in continuous action cases
The linear regression (LR) method offers the advantage that optimal parameters can be calculated relatively easily, although its representation capability is limited than that of the deep learning technique. To improve deep reinforcement learning, the Least Squares Deep Q Network (LS-DQN) method was proposed by Levine et al., which combines Deep Q Network (DQN) with LR method. However, the LS-DQN method assumes that the actions are discrete. In this study, we propose the Double Least Squares Deep Deterministic Policy Gradient (DLS-DDPG) method to address this limitation. This method combines the LR method with the Deep Deterministic Policy Gradient (DDPG) technique, one of the representative deep reinforcement learning algorithms for continuous action cases. Numerical experiments conducted in MuJoCo environments showed that the LR update improved performance at least in some tasks, although there are difficulties such as the inability to make the regularization terms small.
comment: 21 pages, 6 figures
♻ ☆ Offload Rethinking by Cloud Assistance for Efficient Environmental Sound Recognition on LPWANs
Learning-based environmental sound recognition has emerged as a crucial method for ultra-low-power environmental monitoring in biological research and city-scale sensing systems. These systems usually operate under limited resources and are often powered by harvested energy in remote areas. Recent efforts in on-device sound recognition suffer from low accuracy due to resource constraints, whereas cloud offloading strategies are hindered by high communication costs. In this work, we introduce ORCA, a novel resource-efficient cloud-assisted environmental sound recognition system on batteryless devices operating over the Low-Power Wide-Area Networks (LPWANs), targeting wide-area audio sensing applications. We propose a cloud assistance strategy that remedies the low accuracy of on-device inference while minimizing the communication costs for cloud offloading. By leveraging a self-attention-based cloud sub-spectral feature selection method to facilitate efficient on-device inference, ORCA resolves three key challenges for resource-constrained cloud offloading over LPWANs: 1) high communication costs and low data rates, 2) dynamic wireless channel conditions, and 3) unreliable offloading. We implement ORCA on an energy-harvesting batteryless microcontroller and evaluate it in a real world urban sound testbed. Our results show that ORCA outperforms state-of-the-art methods by up to $80 \times$ in energy savings and $220 \times$ in latency reduction while maintaining comparable accuracy.
comment: Accepted by The 23rd ACM Conference on Embedded Networked Sensor Systems (SenSys '25)
♻ ☆ Generating Likely Counterfactuals Using Sum-Product Networks
The need to explain decisions made by AI systems is driven by both recent regulation and user demand. The decisions are often explainable only post hoc. In counterfactual explanations, one may ask what constitutes the best counterfactual explanation. Clearly, multiple criteria must be taken into account, although "distance from the sample" is a key criterion. Recent methods that consider the plausibility of a counterfactual seem to sacrifice this original objective. Here, we present a system that provides high-likelihood explanations that are, at the same time, close and sparse. We show that the search for the most likely explanations satisfying many common desiderata for counterfactual explanations can be modeled using Mixed-Integer Optimization (MIO). We use a Sum-Product Network (SPN) to estimate the likelihood of a counterfactual. To achieve that, we propose an MIO formulation of an SPN, which can be of independent interest. The source code with examples is available at https://github.com/Epanemu/LiCE.
♻ ☆ Babel: A Scalable Pre-trained Model for Multi-Modal Sensing via Expandable Modality Alignment
This paper presents Babel, the expandable modality alignment model, specially designed for multi-modal sensing. While there has been considerable work on multi-modality alignment, they all struggle to effectively incorporate multiple sensing modalities due to the data scarcity constraints. How to utilize multi-modal data with partial pairings in sensing remains an unresolved challenge. Babel tackles this challenge by introducing the concept of expandable modality alignment. The key idea involves transforming the N-modality alignment into a series of binary-modality alignments. Novel techniques are also proposed to further mitigate data scarcity issue and balance the contribution of the newly incorporated modality with the previously established modality alignment during the expandable alignment process. We provide the comprehensive implementation. In the pre-training phase, Babel currently aligns 6 sensing modalities, namely Wi-Fi, mmWave, IMU, LiDAR, video, and depth. For the deployment phase, as a foundation model, any single or combination of aligned modalities could be selected from Babel and applied to downstream tasks. Evaluation demonstrates Babel's outstanding performance on eight human activity recognition datasets, compared to a broad range of baselines e.g., the SOTA single-modal sensing networks, multi-modal sensing framework, and multi-modal large language models. Babel not only improves the performance of individual modality sensing (12% averaged accuracy improvement), but also effectively fuses multiple available modalities (up to 22% accuracy increase). Case studies also highlight emerging application scenarios empowered by Babel, including cross-modality retrieval (i.e., sensing imaging), and bridging LLM for sensing comprehension.
comment: Accepted by SenSys'25
♻ ☆ SALOVA: Segment-Augmented Long Video Assistant for Targeted Retrieval and Routing in Long-Form Video Analysis
Despite advances in Large Multi-modal Models, applying them to long and untrimmed video content remains challenging due to limitations in context length and substantial memory overhead. These constraints often lead to significant information loss and reduced relevance in the model responses. With the exponential growth of video data across web platforms, understanding long-form video is crucial for advancing generalized intelligence. In this paper, we introduce SALOVA: Segment-Augmented LOng Video Assistant, a novel video-LLM framework designed to enhance the comprehension of lengthy video content through targeted retrieval process. We address two main challenges to achieve it: (i) We present the SceneWalk dataset, a high-quality collection of 87.8K long videos, each densely captioned at the segment level to enable models to capture scene continuity and maintain rich descriptive context. (ii) We develop robust architectural designs integrating dynamic routing mechanism and spatio-temporal projector to efficiently retrieve and process relevant video segments based on user queries. Our framework mitigates the limitations of current video-LMMs by allowing for precise identification and retrieval of relevant video segments in response to queries, thereby improving the contextual relevance of the generated responses. Through extensive experiments, SALOVA demonstrates enhanced capability in processing complex long-form videos, showing significant capability to maintain contextual integrity across extended sequences.
comment: Project page: https://ivy-lvlm.github.io/SALOVA/
♻ ☆ From 1,000,000 Users to Every User: Scaling Up Personalized Preference for User-level Alignment
Large language models (LLMs) have traditionally been aligned through one-size-fits-all approaches that assume uniform human preferences, fundamentally overlooking the diversity in user values and needs. This paper introduces a comprehensive framework for scalable personalized alignment of LLMs. We establish a systematic preference space characterizing psychological and behavioral dimensions, alongside diverse persona representations for robust preference inference in real-world scenarios. Building upon this foundation, we introduce \textsc{AlignX}, a large-scale dataset of over 1.3 million personalized preference examples, and develop two complementary alignment approaches: \textit{in-context alignment} directly conditioning on persona representations and \textit{preference-bridged alignment} modeling intermediate preference distributions. Extensive experiments demonstrate substantial improvements over existing methods, with an average 17.06\% accuracy gain across four benchmarks while exhibiting a strong adaptation capability to novel preferences, robustness to limited user data, and precise preference controllability. These results validate our framework's effectiveness, advancing toward truly user-adaptive AI systems.
♻ ☆ HAL 9000: a Risk Manager for ITSs
HAL 9000 is an Intrusion Tolerant Systems (ITSs) Risk Manager, which assesses configuration risks against potential intrusions. It utilizes gathered threat knowledge and remains operational, even in the absence of updated information. Based on its advice, the ITSs can dynamically and proactively adapt to recent threats to minimize and mitigate future intrusions from malicious adversaries. Our goal is to reduce the risk linked to the exploitation of recently uncovered vulnerabilities that have not been classified and/or do not have a script to reproduce the exploit, considering the potential that they may have already been exploited as zero-day exploits. Our experiments demonstrate that the proposed solution can effectively learn and replicate National Vulnerability Database's evaluation process with 99% accuracy.
comment: 10 pages, 4 figures
♻ ☆ SPHINX-X: Scaling Data and Parameters for a Family of Multi-modal Large Language Models ICML 2024
We propose SPHINX-X, an extensive Multimodality Large Language Model (MLLM) series developed upon SPHINX. To improve the architecture and training efficiency, we modify the SPHINX framework by removing redundant visual encoders, bypassing fully-padded sub-images with skip tokens, and simplifying multi-stage training into a one-stage all-in-one paradigm. To fully unleash the potential of MLLMs, we assemble a comprehensive multi-domain and multimodal dataset covering publicly available resources in language, vision, and vision-language tasks. We further enrich this collection with our curated OCR intensive and Set-of-Mark datasets, extending the diversity and generality. By training over different base LLMs including TinyLlama1.1B, InternLM2-7B, LLaMA2-13B, and Mixtral8x7B, we obtain a spectrum of MLLMs that vary in parameter size and multilingual capabilities. Comprehensive benchmarking reveals a strong correlation between the multi-modal performance with the data and parameter scales. Code and models are released at https://github.com/Alpha-VLLM/LLaMA2-Accessory
comment: Accepted by ICML 2024. Code and models are released at https://github.com/Alpha-VLLM/LLaMA2-Accessory
♻ ☆ Strength Estimation and Human-Like Strength Adjustment in Games ICLR 2025
Strength estimation and adjustment are crucial in designing human-AI interactions, particularly in games where AI surpasses human players. This paper introduces a novel strength system, including a strength estimator (SE) and an SE-based Monte Carlo tree search, denoted as SE-MCTS, which predicts strengths from games and offers different playing strengths with human styles. The strength estimator calculates strength scores and predicts ranks from games without direct human interaction. SE-MCTS utilizes the strength scores in a Monte Carlo tree search to adjust playing strength and style. We first conduct experiments in Go, a challenging board game with a wide range of ranks. Our strength estimator significantly achieves over 80% accuracy in predicting ranks by observing 15 games only, whereas the previous method reached 49% accuracy for 100 games. For strength adjustment, SE-MCTS successfully adjusts to designated ranks while achieving a 51.33% accuracy in aligning to human actions, outperforming a previous state-of-the-art, with only 42.56% accuracy. To demonstrate the generality of our strength system, we further apply SE and SE-MCTS to chess and obtain consistent results. These results show a promising approach to strength estimation and adjustment, enhancing human-AI interactions in games. Our code is available at https://rlg.iis.sinica.edu.tw/papers/strength-estimator.
comment: Accepted by the Thirteenth International Conference on Learning Representations (ICLR 2025)
♻ ☆ TAET: Two-Stage Adversarial Equalization Training on Long-Tailed Distributions CVPR2025
Adversarial robustness is a critical challenge in deploying deep neural networks for real-world applications. While adversarial training is a widely recognized defense strategy, most existing studies focus on balanced datasets, overlooking the prevalence of long-tailed distributions in real-world data, which significantly complicates robustness. This paper provides a comprehensive analysis of adversarial training under long-tailed distributions and identifies limitations in the current state-of-the-art method, AT-BSL, in achieving robust performance under such conditions. To address these challenges, we propose a novel training framework, TAET, which integrates an initial stabilization phase followed by a stratified equalization adversarial training phase. Additionally, prior work on long-tailed robustness has largely ignored the crucial evaluation metric of balanced accuracy. To bridge this gap, we introduce the concept of balanced robustness, a comprehensive metric tailored for assessing robustness under long-tailed distributions. Extensive experiments demonstrate that our method surpasses existing advanced defenses, achieving significant improvements in both memory and computational efficiency. This work represents a substantial advancement in addressing robustness challenges in real-world applications. Our code is available at: https://github.com/BuhuiOK/TAET-Two-Stage-Adversarial-Equalization-Training-on-Long-Tailed-Distributions.
comment: Text: 8 pages of main content, 5 pages of appendices have been accepted by CVPR2025
♻ ☆ Solving Drone Routing Problems with Quantum Computing: A Hybrid Approach Combining Quantum Annealing and Gate-Based Paradigms CEC 2025
This paper presents a novel hybrid approach to solving real-world drone routing problems by leveraging the capabilities of quantum computing. The proposed method, coined Quantum for Drone Routing (Q4DR), integrates the two most prominent paradigms in the field: quantum gate-based computing, through the Eclipse Qrisp programming language; and quantum annealers, by means of D-Wave System's devices. The algorithm is divided into two different phases: an initial clustering phase executed using a Quantum Approximate Optimization Algorithm (QAOA), and a routing phase employing quantum annealers. The efficacy of Q4DR is demonstrated through three use cases of increasing complexity, each incorporating real-world constraints such as asymmetric costs, forbidden paths, and itinerant charging points. This research contributes to the growing body of work in quantum optimization, showcasing the practical applications of quantum computing in logistics and route planning.
comment: 8 pages, 5 figures. Paper accepted for being presented in the IEEE Congress on Evolutionary Computation (IEEE CEC 2025)
♻ ☆ Advancing Tool-Augmented Large Language Models: Integrating Insights from Errors in Inference Trees NeurIPS 2024
Tool-augmented large language models (LLMs) leverage tools, often in the form of APIs, to improve their reasoning capabilities on complex tasks. This enables them to act as intelligent agents interacting with the real world. The recently introduced ToolLLaMA model by Qin et al. [2023] utilizes the depth-first search-based decision tree (DFSDT) mechanism for multi-step reasoning with $16000+$ real-world APIs, effectively enhancing the performance of tool-augmented LLMs compared to traditional chain reasoning mechanisms. However, their approach only employs successful paths from decision trees (also called inference trees) for supervised fine-tuning (SFT), missing out on the potential learning opportunities from failed paths. Inspired by this, we propose an inference trajectory optimization framework based on preference learning to address this limitation. We first introduce a novel method for constructing step-wise preference data from tree-like expert trajectories, which leverages the previously ignored failed explorations in the decision trees. In the subsequent training phase, we first fine-tune the LLM with successful tool-usage expert trajectories and then apply direct preference optimization (DPO) with the preference data to update the LLM's policy, resulting in our ToolPrefer-LLaMA (TP-LLaMA) model. This approach not only enhances the utilization of original expert data but also broadens the learning space of the model. Our experiments demonstrate that by obtaining insights from errors in inference trees, TP-LLaMA significantly outperforms the baselines across almost all test scenarios by a large margin and exhibits better generalization capabilities with unseen APIs. At the same time, TP-LLaMA has also demonstrated superior reasoning efficiency compared to the baselines, making it more suitable for complex tool-usage reasoning tasks.
comment: Accepted by NeurIPS 2024
♻ ☆ A Unified Framework for Real-Time Failure Handling in Robotics Using Vision-Language Models, Reactive Planner and Behavior Trees
Robotic systems often face execution failures due to unexpected obstacles, sensor errors, or environmental changes. Traditional failure recovery methods rely on predefined strategies or human intervention, making them less adaptable. This paper presents a unified failure recovery framework that combines Vision-Language Models (VLMs), a reactive planner, and Behavior Trees (BTs) to enable real-time failure handling. Our approach includes pre-execution verification, which checks for potential failures before execution, and reactive failure handling, which detects and corrects failures during execution by verifying existing BT conditions, adding missing preconditions and, when necessary, generating new skills. The framework uses a scene graph for structured environmental perception and an execution history for continuous monitoring, enabling context-aware and adaptive failure handling. We evaluate our framework through real-world experiments with an ABB YuMi robot on tasks like peg insertion, object sorting, and drawer placement, as well as in AI2-THOR simulator. Compared to using pre-execution and reactive methods separately, our approach achieves higher task success rates and greater adaptability. Ablation studies highlight the importance of VLM-based reasoning, structured scene representation, and execution history tracking for effective failure recovery in robotics.
♻ ☆ GAPartManip: A Large-scale Part-centric Dataset for Material-Agnostic Articulated Object Manipulation ICRA 2025
Effectively manipulating articulated objects in household scenarios is a crucial step toward achieving general embodied artificial intelligence. Mainstream research in 3D vision has primarily focused on manipulation through depth perception and pose detection. However, in real-world environments, these methods often face challenges due to imperfect depth perception, such as with transparent lids and reflective handles. Moreover, they generally lack the diversity in part-based interactions required for flexible and adaptable manipulation. To address these challenges, we introduced a large-scale part-centric dataset for articulated object manipulation that features both photo-realistic material randomization and detailed annotations of part-oriented, scene-level actionable interaction poses. We evaluated the effectiveness of our dataset by integrating it with several state-of-the-art methods for depth estimation and interaction pose prediction. Additionally, we proposed a novel modular framework that delivers superior and robust performance for generalizable articulated object manipulation. Our extensive experiments demonstrate that our dataset significantly improves the performance of depth perception and actionable interaction pose prediction in both simulation and real-world scenarios. More information and demos can be found at: https://pku-epic.github.io/GAPartManip/.
comment: Accepted by ICRA 2025. Project page: https://pku-epic.github.io/GAPartManip/
♻ ☆ Bridging Social Psychology and LLM Reasoning: Conflict-Aware Meta-Review Generation via Cognitive Alignment
The rapid growth of scholarly submissions has overwhelmed traditional peer review systems, driving the need for intelligent automation to preserve scientific rigor. While large language models (LLMs) show promise in automating manuscript critiques, their ability to synthesize high-stakes meta-reviews, which require conflict-aware reasoning and consensus derivation, remains underdeveloped. Existing methods fail to effectively handle conflicting viewpoints within differing opinions, and often introduce additional cognitive biases, such as anchoring effects and conformity bias.To overcome these limitations, we propose the Cognitive Alignment Framework (CAF), a dual-process architecture that transforms LLMs into adaptive scientific arbitrators. By operationalizing Kahneman's dual-process theory, CAF introduces a three-step cognitive pipeline: review initialization, incremental integration, and cognitive alignment.Empirical validation shows that CAF outperforms existing LLM-based methods, with sentiment consistency gains reaching up to 19.47\% and content consistency improving by as much as 12.95\%.
comment: 23 pages
♻ ☆ Aligning Text to Image in Diffusion Models is Easier Than You Think
While recent advancements in generative modeling have significantly improved text-image alignment, some residual misalignment between text and image representations still remains. Although many approaches have attempted to address this issue by fine-tuning models using various reward models, etc., we revisit the challenge from the perspective of representation alignment-an approach that has gained popularity with the success of REPresentation Alignment (REPA). We first argue that conventional text-to-image (T2I) diffusion models, typically trained on paired image and text data (i.e., positive pairs) by minimizing score matching or flow matching losses, is suboptimal from the standpoint of representation alignment. Instead, a better alignment can be achieved through contrastive learning that leverages both positive and negative pairs. To achieve this efficiently even with pretrained models, we introduce a lightweight contrastive fine tuning strategy called SoftREPA that uses soft text tokens. This approach improves alignment with minimal computational overhead by adding fewer than 1M trainable parameters to the pretrained model. Our theoretical analysis demonstrates that our method explicitly increases the mutual information between text and image representations, leading to enhanced semantic consistency. Experimental results across text-to-image generation and text-guided image editing tasks validate the effectiveness of our approach in improving the semantic consistency of T2I generative models.
♻ ☆ Analysis and Fully Memristor-based Reservoir Computing for Temporal Data Classification
Reservoir computing (RC) offers a neuromorphic framework that is particularly effective for processing spatiotemporal signals. Known for its temporal processing prowess, RC significantly lowers training costs compared to conventional recurrent neural networks. A key component in its hardware deployment is the ability to generate dynamic reservoir states. Our research introduces a novel dual-memory RC system, integrating a short-term memory via a WOx-based memristor, capable of achieving 16 distinct states encoded over 4 bits, and a long-term memory component using a TiOx-based memristor within the readout layer. We thoroughly examine both memristor types and leverage the RC system to process temporal data sets. The performance of the proposed RC system is validated through two benchmark tasks: isolated spoken digit recognition with incomplete inputs and Mackey-Glass time series prediction. The system delivered an impressive 98.84% accuracy in digit recognition and sustained a low normalized root mean square error (NRMSE) of 0.036 in the time series prediction task, underscoring its capability. This study illuminates the adeptness of memristor-based RC systems in managing intricate temporal challenges, laying the groundwork for further innovations in neuromorphic computing.
comment: 22 pages, 20 figures, Journal, Typo corrected and updated reference
♻ ☆ Catastrophic Failure of LLM Unlearning via Quantization ICLR 2025
Large language models (LLMs) have shown remarkable proficiency in generating text, benefiting from extensive training on vast textual corpora. However, LLMs may also acquire unwanted behaviors from the diverse and sensitive nature of their training data, which can include copyrighted and private content. Machine unlearning has been introduced as a viable solution to remove the influence of such problematic content without the need for costly and time-consuming retraining. This process aims to erase specific knowledge from LLMs while preserving as much model utility as possible. Despite the effectiveness of current unlearning methods, little attention has been given to whether existing unlearning methods for LLMs truly achieve forgetting or merely hide the knowledge, which current unlearning benchmarks fail to detect. This paper reveals that applying quantization to models that have undergone unlearning can restore the "forgotten" information. To thoroughly evaluate this phenomenon, we conduct comprehensive experiments using various quantization techniques across multiple precision levels. We find that for unlearning methods with utility constraints, the unlearned model retains an average of 21\% of the intended forgotten knowledge in full precision, which significantly increases to 83\% after 4-bit quantization. ... Our code is available at: \href{https://github.com/zzwjames/FailureLLMUnlearning}{https://github.com/zzwjames/FailureLLMUnlearning}.
comment: ICLR 2025
♻ ☆ Bias Testing and Mitigation in LLM-based Code Generation
As the adoption of LLMs becomes more widespread in software coding ecosystems, a pressing issue has emerged: does the generated code contain social bias and unfairness, such as those related to age, gender, and race? This issue concerns the integrity, fairness, and ethical foundation of software applications that depend on the code generated by these models but are underexplored in the literature. This paper presents a novel bias testing framework that is specifically designed for code generation tasks. Based on this framework, we conduct an extensive empirical study on the biases in code generated by five widely studied LLMs (i.e., PALM-2-CodeChat-bison, Claude-instant-1, GPT-3.5-turbo, GPT-4-turbo, and GPT-4). Our findings reveal that biases are prevalent. For example, 13.47% to 49.10% of the codes generated by these LLMs have biased behaviors towards gender. Moreover, we study five bias mitigation prompt strategies that are commonly used in current code generation scenarios, i.e., zero-shot, one-shot, few-shot, and two Chain-of-Thought (CoT) prompts, with and without provided feedback-driven refinement. Our evaluation results illustrate that using direct prompt engineering strategies has limited effectiveness in mitigating bias, but our test execution feedback can help to reduce the ratio of code biases to a large extent (e.g., from 59.88% to 4.79% for GPT-4).
comment: Accepted by TOSEM
♻ ☆ AVA: Attentive VLM Agent for Mastering StarCraft II
We introduce Attentive VLM Agent (AVA), a multimodal StarCraft II agent that aligns artificial agent perception with the human gameplay experience. Traditional frameworks such as SMAC rely on abstract state representations that diverge significantly from human perception, limiting the ecological validity of agent behavior. Our agent addresses this limitation by incorporating RGB visual inputs and natural language observations that more closely simulate human cognitive processes during gameplay. The AVA architecture consists of three integrated components: (1) a vision-language model enhanced with specialized self-attention mechanisms for strategic unit targeting and battlefield assessment, (2) a retrieval-augmented generation system that leverages domain-specific StarCraft II knowledge to inform tactical decisions, and (3) a dynamic role-based task distribution system that enables coordinated multi-agent behavior. The experimental evaluation in our proposed AVACraft environment, which contains 21 multimodal StarCraft II scenarios, demonstrates that AVA powered by foundation models (specifically Qwen-VL and GPT-4o) can execute complex tactical maneuvers without explicit training, achieving comparable performance to traditional MARL methods that require substantial training iterations. This work establishes a foundation for developing human-aligned StarCraft II agents and advances the broader research agenda of multimodal game AI. Our implementation is available at https://github.com/camel-ai/VLM-Play-StarCraft2.
comment: Under Review
♻ ☆ MBQ: Modality-Balanced Quantization for Large Vision-Language Models
Vision-Language Models (VLMs) have enabled a variety of real-world applications. The large parameter size of VLMs brings large memory and computation overhead which poses significant challenges for deployment. Post-Training Quantization (PTQ) is an effective technique to reduce the memory and computation overhead. Existing PTQ methods mainly focus on large language models (LLMs), without considering the differences across other modalities. In this paper, we discover that there is a significant difference in sensitivity between language and vision tokens in large VLMs. Therefore, treating tokens from different modalities equally, as in existing PTQ methods, may over-emphasize the insensitive modalities, leading to significant accuracy loss. To deal with the above issue, we propose a simple yet effective method, Modality-Balanced Quantization (MBQ), for large VLMs. Specifically, MBQ incorporates the different sensitivities across modalities during the calibration process to minimize the reconstruction loss for better quantization parameters. Extensive experiments show that MBQ can significantly improve task accuracy by up to 4.4% and 11.6% under W3 and W4A8 quantization for 7B to 70B VLMs, compared to SOTA baselines. Additionally, we implement a W3 GPU kernel that fuses the dequantization and GEMV operators, achieving a 1.4x speedup on LLaVA-onevision-7B on the RTX 4090. The code is available at https://github.com/thu-nics/MBQ.
♻ ☆ Diffusion Attribution Score: Evaluating Training Data Influence in Diffusion Models
As diffusion models become increasingly popular, the misuse of copyrighted and private images has emerged as a major concern. One promising solution to mitigate this issue is identifying the contribution of specific training samples in generative models, a process known as data attribution. Existing data attribution methods for diffusion models typically quantify the contribution of a training sample by evaluating the change in diffusion loss when the sample is included or excluded from the training process. However, we argue that the direct usage of diffusion loss cannot represent such a contribution accurately due to the calculation of diffusion loss. Specifically, these approaches measure the divergence between predicted and ground truth distributions, which leads to an indirect comparison between the predicted distributions and cannot represent the variances between model behaviors. To address these issues, we aim to measure the direct comparison between predicted distributions with an attribution score to analyse the training sample importance, which is achieved by Diffusion Attribution Score (\textit{DAS}). Underpinned by rigorous theoretical analysis, we elucidate the effectiveness of DAS. Additionally, we explore strategies to accelerate DAS calculations, facilitating its application to large-scale diffusion models. Our extensive experiments across various datasets and diffusion models demonstrate that DAS significantly surpasses previous benchmarks in terms of the linear data-modelling score, establishing new state-of-the-art performance. Code is available at \hyperlink{here}{https://github.com/Jinxu-Lin/DAS}.
♻ ☆ Large Language Models and Causal Inference in Collaboration: A Survey NAACL 2025
Causal inference has shown potential in enhancing the predictive accuracy, fairness, robustness, and explainability of Natural Language Processing (NLP) models by capturing causal relationships among variables. The emergence of generative Large Language Models (LLMs) has significantly impacted various NLP domains, particularly through their advanced reasoning capabilities. This survey focuses on evaluating and improving LLMs from a causal view in the following areas: understanding and improving the LLMs' reasoning capacity, addressing fairness and safety issues in LLMs, complementing LLMs with explanations, and handling multimodality. Meanwhile, LLMs' strong reasoning capacities can in turn contribute to the field of causal inference by aiding causal relationship discovery and causal effect estimations. This review explores the interplay between causal inference frameworks and LLMs from both perspectives, emphasizing their collective potential to further the development of more advanced and equitable artificial intelligence systems.
comment: Findings of the Association for Computational Linguistics: NAACL 2025
♻ ☆ Omni6D: Large-Vocabulary 3D Object Dataset for Category-Level 6D Object Pose Estimation ECCV 2024
6D object pose estimation aims at determining an object's translation, rotation, and scale, typically from a single RGBD image. Recent advancements have expanded this estimation from instance-level to category-level, allowing models to generalize across unseen instances within the same category. However, this generalization is limited by the narrow range of categories covered by existing datasets, such as NOCS, which also tend to overlook common real-world challenges like occlusion. To tackle these challenges, we introduce Omni6D, a comprehensive RGBD dataset featuring a wide range of categories and varied backgrounds, elevating the task to a more realistic context. 1) The dataset comprises an extensive spectrum of 166 categories, 4688 instances adjusted to the canonical pose, and over 0.8 million captures, significantly broadening the scope for evaluation. 2) We introduce a symmetry-aware metric and conduct systematic benchmarks of existing algorithms on Omni6D, offering a thorough exploration of new challenges and insights. 3) Additionally, we propose an effective fine-tuning approach that adapts models from previous datasets to our extensive vocabulary setting. We believe this initiative will pave the way for new insights and substantial progress in both the industrial and academic fields, pushing forward the boundaries of general 6D pose estimation.
comment: ECCV 2024 (poster). Github page: https://github.com/3DTopia/Omni6D
♻ ☆ AlignBot: Aligning VLM-powered Customized Task Planning with User Reminders Through Fine-Tuning for Household Robots
This paper presents AlignBot, a novel framework designed to optimize VLM-powered customized task planning for household robots by effectively aligning with user reminders. In domestic settings, aligning task planning with user reminders poses significant challenges due to the limited quantity, diversity, and multimodal nature of the reminders. To address these challenges, AlignBot employs a fine-tuned LLaVA-7B model, functioning as an adapter for GPT-4o. This adapter model internalizes diverse forms of user reminders-such as personalized preferences, corrective guidance, and contextual assistance-into structured instruction-formatted cues that prompt GPT-4o in generating customized task plans. Additionally, AlignBot integrates a dynamic retrieval mechanism that selects task-relevant historical successes as prompts for GPT-4o, further enhancing task planning accuracy. To validate the effectiveness of AlignBot, experiments are conducted in real-world household environments, which are constructed within the laboratory to replicate typical household settings. A multimodal dataset with over 1,500 entries derived from volunteer reminders is used for training and evaluation. The results demonstrate that AlignBot significantly improves customized task planning, outperforming existing LLM- and VLM-powered planners by interpreting and aligning with user reminders, achieving 86.8% success rate compared to the vanilla GPT-4o baseline at 21.6%, reflecting a 65% improvement and over four times greater effectiveness. Supplementary materials are available at: https://yding25.com/AlignBot/
♻ ☆ SeqAfford: Sequential 3D Affordance Reasoning via Multimodal Large Language Model
3D affordance segmentation aims to link human instructions to touchable regions of 3D objects for embodied manipulations. Existing efforts typically adhere to single-object, single-affordance paradigms, where each affordance type or explicit instruction strictly corresponds to a specific affordance region and are unable to handle long-horizon tasks. Such a paradigm cannot actively reason about complex user intentions that often imply sequential affordances. In this paper, we introduce the Sequential 3D Affordance Reasoning task, which extends the traditional paradigm by reasoning from cumbersome user intentions and then decomposing them into a series of segmentation maps. Toward this, we construct the first instruction-based affordance segmentation benchmark that includes reasoning over both single and sequential affordances, comprising 180K instruction-point cloud pairs. Based on the benchmark, we propose our model, SeqAfford, to unlock the 3D multi-modal large language model with additional affordance segmentation abilities, which ensures reasoning with world knowledge and fine-grained affordance grounding in a cohesive framework. We further introduce a multi-granular language-point integration module to endow 3D dense prediction. Extensive experimental evaluations show that our model excels over well-established methods and exhibits open-world generalization with sequential reasoning abilities.
♻ ☆ Specialized Foundation Models Struggle to Beat Supervised Baselines
Following its success for vision and text, the "foundation model" (FM) paradigm -- pretraining large models on massive data, then fine-tuning on target tasks -- has rapidly expanded to domains in the sciences, engineering, healthcare, and beyond. Has this achieved what the original FMs accomplished, i.e. the supplanting of traditional supervised learning in their domains? To answer we look at three modalities -- genomics, satellite imaging, and time series -- with multiple recent FMs and compare them to a standard supervised learning workflow: model development, hyperparameter tuning, and training, all using only data from the target task. Across these three specialized domains, we find that it is consistently possible to train simple supervised models -- no more complicated than a lightly modified wide ResNet or UNet -- that match or even outperform the latest foundation models. Our work demonstrates that the benefits of large-scale pretraining have yet to be realized in many specialized areas, reinforces the need to compare new FMs to strong, well-tuned baselines, and introduces two new, easy-to-use, open-source, and automated workflows for doing so.
comment: The first two authors contributed equally. The order was determined by coin flip
♻ ☆ Examining Two Hop Reasoning Through Information Content Scaling
Prior work has found that transformers have an inconsistent ability to learn to answer latent two-hop questions -- questions of the form "Who is Bob's mother's boss?" We study why this is the case by examining how transformers' capacity to learn datasets of two-hop questions and answers (two-hop QA) scales with their size, motivated by prior work on transformer knowledge capacity for simple factual memorization. We find that capacity scaling and generalization both support the hypothesis that latent two-hop QA requires transformers to learn each fact twice, while two-hop QA with chain of thought does not. We also show that with appropriate dataset parameters, it is possible to "trap" very small models in a regime where they memorize answers to two-hop questions independently, even though they would perform better if they could learn to answer them with function composition. Our findings show that measurement of capacity scaling can complement existing interpretability methods, though there are challenges in using it for this purpose.
♻ ☆ Mitigating Hallucinations in Multimodal Spatial Relations through Constraint-Aware Prompting
Spatial relation hallucinations pose a persistent challenge in large vision-language models (LVLMs), leading to generate incorrect predictions about object positions and spatial configurations within an image. To address this issue, we propose a constraint-aware prompting framework designed to reduce spatial relation hallucinations. Specifically, we introduce two types of constraints: (1) bidirectional constraint, which ensures consistency in pairwise object relations, and (2) transitivity constraint, which enforces relational dependence across multiple objects. By incorporating these constraints, LVLMs can produce more spatially coherent and consistent outputs. We evaluate our method on three widely-used spatial relation datasets, demonstrating performance improvements over existing approaches. Additionally, a systematic analysis of various bidirectional relation analysis choices and transitivity reference selections highlights greater possibilities of our methods in incorporating constraints to mitigate spatial relation hallucinations.
comment: 19 pages
♻ ☆ STP: Self-play LLM Theorem Provers with Iterative Conjecturing and Proving
A fundamental challenge in formal theorem proving by LLMs is the lack of high-quality training data. Although reinforcement learning or expert iteration partially mitigates this issue by alternating between LLM generating proofs and finetuning them on correctly generated ones, performance quickly plateaus due to the scarcity of correct proofs (sparse rewards). To keep improving the models with limited data, we draw inspiration from mathematicians, who continuously develop new results, partly by proposing novel conjectures or exercises (which are often variants of known results) and attempting to solve them. We design the Self-play Theorem Prover (STP) that simultaneously takes on two roles, conjecturer and prover, each providing training signals to the other. The conjecturer is trained iteratively on previously generated conjectures that are barely provable by the current prover, which incentivizes it to generate increasingly challenging conjectures over time. The prover attempts to prove the conjectures with standard expert iteration. We evaluate STP with both Lean and Isabelle formal versifiers. With 51.3 billion tokens generated during the training in Lean, STP proves 28.5% of the statements in the LeanWorkbook dataset, doubling the previous best result of 13.2% achieved through expert iteration. The final model achieves state-of-the-art performance among whole-proof generation methods on miniF2F-test (65.0%, pass@3200), Proofnet-test (23.9%, pass@3200) and PutnamBench (8/644, pass@3200). We release our code, model, and dataset in this URL: https://github.com/kfdong/STP.
comment: 25 pages, 5 figures
♻ ☆ MODL: Multilearner Online Deep Learning
Online deep learning tackles the challenge of learning from data streams by balancing two competing goals: fast learning and deep learning. However, existing research primarily emphasizes deep learning solutions, which are more adept at handling the ``deep'' aspect than the ``fast'' aspect of online learning. In this work, we introduce an alternative paradigm through a hybrid multilearner approach. We begin by developing a fast online logistic regression learner, which operates without relying on backpropagation. It leverages closed-form recursive updates of model parameters, efficiently addressing the fast learning component of the online learning challenge. This approach is further integrated with a cascaded multilearner design, where shallow and deep learners are co-trained in a cooperative, synergistic manner to solve the online learning problem. We demonstrate that this approach achieves state-of-the-art performance on standard online learning datasets. We make our code available: https://github.com/AntonValk/MODL
♻ ☆ Advancing Mobile GUI Agents: A Verifier-Driven Approach to Practical Deployment
We propose V-Droid, a mobile GUI task automation agent. Unlike previous mobile agents that utilize Large Language Models (LLMs) as generators to directly generate actions at each step, V-Droid employs LLMs as verifiers to evaluate candidate actions before making final decisions. To realize this novel paradigm, we introduce a comprehensive framework for constructing verifier-driven mobile agents: the discretized action space construction coupled with the prefilling-only workflow to accelerate the verification process, the pair-wise progress preference training to significantly enhance the verifier's decision-making capabilities, and the scalable human-agent joint annotation scheme to efficiently collect the necessary data at scale. V-Droid sets a new state-of-the-art task success rate across several public mobile task automation benchmarks: 59.5% on AndroidWorld, 38.3% on AndroidLab, and 49% on MobileAgentBench, surpassing existing agents by 9.5%, 2.1%, and 9%, respectively. Furthermore, V-Droid achieves an impressively low latency of 0.7 seconds per step, making it the first mobile agent capable of delivering near-real-time, effective decision-making capabilities.
comment: 14 pages, 4 iterations, refine figs
♻ ☆ Label Unbalance in High-frequency Trading
In financial trading, return prediction is one of the foundation for a successful trading system. By the fast development of the deep learning in various areas such as graphical processing, natural language, it has also demonstrate significant edge in handling with financial data. While the success of the deep learning relies on huge amount of labeled sample, labeling each time/event as profitable or unprofitable, under the transaction cost, especially in the high-frequency trading world, suffers from serious label imbalance issue.In this paper, we adopts rigurious end-to-end deep learning framework with comprehensive label imbalance adjustment methods and succeed in predicting in high-frequency return in the Chinese future market. The code for our method is publicly available at https://github.com/RS2002/Label-Unbalance-in-High-Frequency-Trading .
comment: Technical Report
♻ ☆ From Open Vocabulary to Open World: Teaching Vision Language Models to Detect Novel Objects
Traditional object detection methods operate under the closed-set assumption, where models can only detect a fixed number of objects predefined in the training set. Recent works on open vocabulary object detection (OVD) enable the detection of objects defined by an in-principle unbounded vocabulary, which reduces the cost of training models for specific tasks. However, OVD heavily relies on accurate prompts provided by an ``oracle'', which limits their use in critical applications such as driving scene perception. OVD models tend to misclassify near-out-of-distribution (NOOD) objects that have similar features to known classes, and ignore far-out-of-distribution (FOOD) objects. To address these limitations, we propose a framework that enables OVD models to operate in open world settings, by identifying and incrementally learning previously unseen objects. To detect FOOD objects, we propose Open World Embedding Learning (OWEL) and introduce the concept of Pseudo Unknown Embedding which infers the location of unknown classes in a continuous semantic space based on the information of known classes. We also propose Multi-Scale Contrastive Anchor Learning (MSCAL), which enables the identification of misclassified unknown objects by promoting the intra-class consistency of object embeddings at different scales. The proposed method achieves state-of-the-art performance on standard open world object detection and autonomous driving benchmarks while maintaining its open vocabulary object detection capability.
♻ ☆ Knowledge-aware contrastive heterogeneous molecular graph learning
Molecular representation learning is pivotal in predicting molecular properties and advancing drug design. Traditional methodologies, which predominantly rely on homogeneous graph encoding, are limited by their inability to integrate external knowledge and represent molecular structures across different levels of granularity. To address these limitations, we propose a paradigm shift by encoding molecular graphs into heterogeneous structures, introducing a novel framework: Knowledge-aware Contrastive Heterogeneous Molecular Graph Learning (KCHML). This approach leverages contrastive learning to enrich molecular representations with embedded external knowledge. KCHML conceptualizes molecules through three distinct graph views-molecular, elemental, and pharmacological-enhanced by heterogeneous molecular graphs and a dual message-passing mechanism. This design offers a comprehensive representation for property prediction, as well as for downstream tasks such as drug-drug interaction (DDI) prediction. Extensive benchmarking demonstrates KCHML's superiority over state-of-the-art molecular property prediction models, underscoring its ability to capture intricate molecular features.
♻ ☆ Knowledge Graph Embeddings: A Comprehensive Survey on Capturing Relation Properties
Knowledge Graph Embedding (KGE) techniques play a pivotal role in transforming symbolic Knowledge Graphs (KGs) into numerical representations, thereby enhancing various deep learning models for knowledge-augmented applications. Unlike entities, relations in KGs are the carriers of semantic meaning, and their accurate modeling is crucial for the performance of KGE models. Firstly, we address the complex mapping properties inherent in relations, such as one-to-one, one-to-many, many-to-one, and many-to-many mappings. We provide a comprehensive summary of relation-aware mapping-based models, models that utilize specific representation spaces, tensor decomposition-based models, and neural network-based models. Next, focusing on capturing various relation patterns like symmetry, asymmetry, inversion, and composition, we review models that employ modified tensor decomposition, those based on modified relation-aware mappings, and those that leverage rotation operations. Subsequently, considering the implicit hierarchical relations among entities, we introduce models that incorporate auxiliary information, models based on hyperbolic spaces, and those that utilize the polar coordinate system. Finally, in response to more complex scenarios such as sparse and dynamic KGs, this paper discusses potential future research directions. We explore innovative ideas such as integrating multimodal information into KGE, enhancing relation pattern modeling with rules, and developing models to capture relation characteristics in dynamic KGE settings.
comment: 22 pages, 8 figures, 3 tables, this paper is a modified English version of our article already published in Computer Science journal (in Chinese), released to facilitate communication among international researchers in the relevant fields
♻ ☆ Prioritized Trajectory Replay: A Replay Memory for Data-driven Reinforcement Learning AAMAS 2024
In recent years, data-driven reinforcement learning (RL), also known as offline RL, have gained significant attention. However, the role of data sampling techniques in offline RL has been overlooked despite its potential to enhance online RL performance. Recent research suggests applying sampling techniques directly to state-transitions does not consistently improve performance in offline RL. Therefore, in this study, we propose a memory technique, (Prioritized) Trajectory Replay (TR/PTR), which extends the sampling perspective to trajectories for more comprehensive information extraction from limited data. TR enhances learning efficiency by backward sampling of trajectories that optimizes the use of subsequent state information. Building on TR, we build the weighted critic target to avoid sampling unseen actions in offline training, and Prioritized Trajectory Replay (PTR) that enables more efficient trajectory sampling, prioritized by various trajectory priority metrics. We demonstrate the benefits of integrating TR and PTR with existing offline RL algorithms on D4RL. In summary, our research emphasizes the significance of trajectory-based data sampling techniques in enhancing the efficiency and performance of offline RL algorithms.
comment: Accepted by AAMAS 2024, see https://dl.acm.org/doi/10.5555/3635637.3662980
♻ ☆ ChatBEV: A Visual Language Model that Understands BEV Maps
Traffic scene understanding is essential for intelligent transportation systems and autonomous driving, ensuring safe and efficient vehicle operation. While recent advancements in VLMs have shown promise for holistic scene understanding, the application of VLMs to traffic scenarios, particularly using BEV maps, remains under explored. Existing methods often suffer from limited task design and narrow data amount, hindering comprehensive scene understanding. To address these challenges, we introduce ChatBEV-QA, a novel BEV VQA benchmark contains over 137k questions, designed to encompass a wide range of scene understanding tasks, including global scene understanding, vehicle-lane interactions, and vehicle-vehicle interactions. This benchmark is constructed using an novel data collection pipeline that generates scalable and informative VQA data for BEV maps. We further fine-tune a specialized vision-language model ChatBEV, enabling it to interpret diverse question prompts and extract relevant context-aware information from BEV maps. Additionally, we propose a language-driven traffic scene generation pipeline, where ChatBEV facilitates map understanding and text-aligned navigation guidance, significantly enhancing the generation of realistic and consistent traffic scenarios. The dataset, code and the fine-tuned model will be released.
♻ ☆ Mirror Descent and Novel Exponentiated Gradient Algorithms Using Trace-Form Entropies and Deformed Logarithms
In this paper we propose and investigate a wide class of Mirror Descent updates (MD) and associated novel Generalized Exponentiated Gradient (GEG) algorithms by exploiting various trace-form entropies and associated deformed logarithms and their inverses - deformed (generalized) exponential functions. The proposed algorithms can be considered as extension of entropic MD and generalization of multiplicative updates. In the literature, there exist nowadays over fifty mathematically well defined generalized entropies, so impossible to exploit all of them in one research paper. So we focus on a few selected most popular entropies and associated logarithms like the Tsallis, Kaniadakis and Sharma-Taneja-Mittal and some of their extension like Tempesta or Kaniadakis-Scarfone entropies. The shape and properties of the deformed logarithms and their inverses are tuned by one or more hyperparameters. By learning these hyperparameters, we can adapt to distribution of training data, which can be designed to the specific geometry of the optimization problem, leading to potentially faster convergence and better performance. The using generalized entropies and associated deformed logarithms in the Bregman divergence, used as a regularization term, provides some new insight into exponentiated gradient descent updates.
comment: 22 pages, 9 figures
♻ ☆ HunyuanProver: A Scalable Data Synthesis Framework and Guided Tree Search for Automated Theorem Proving
We introduce HunyuanProver, an language model finetuned from the Hunyuan 7B for interactive automatic theorem proving with LEAN4. To alleviate the data sparsity issue, we design a scalable framework to iterative synthesize data with low cost. Besides, guided tree search algorithms are designed to enable effective ``system 2 thinking`` of the prover. HunyuanProver achieves state-of-the-art (SOTA) performances on major benchmarks. Specifically, it achieves a pass of 68.4% on the miniF2F-test compared to 65.9%, the current SOTA results. It proves 4 IMO statements (imo_1960_p2, imo_1962_p2}, imo_1964_p2 and imo_1983_p6) in miniF2F-test. To benefit the community, we will open-source a dataset of 30k synthesized instances, where each instance contains the original question in natural language, the converted statement by autoformalization, and the proof by HunyuanProver.
♻ ☆ KARMA: Augmenting Embodied AI Agents with Long-and-short Term Memory Systems
Embodied AI agents responsible for executing interconnected, long-sequence household tasks often face difficulties with in-context memory, leading to inefficiencies and errors in task execution. To address this issue, we introduce KARMA, an innovative memory system that integrates long-term and short-term memory modules, enhancing large language models (LLMs) for planning in embodied agents through memory-augmented prompting. KARMA distinguishes between long-term and short-term memory, with long-term memory capturing comprehensive 3D scene graphs as representations of the environment, while short-term memory dynamically records changes in objects' positions and states. This dual-memory structure allows agents to retrieve relevant past scene experiences, thereby improving the accuracy and efficiency of task planning. Short-term memory employs strategies for effective and adaptive memory replacement, ensuring the retention of critical information while discarding less pertinent data. Compared to state-of-the-art embodied agents enhanced with memory, our memory-augmented embodied AI agent improves success rates by 1.3x and 2.3x in Composite Tasks and Complex Tasks within the AI2-THOR simulator, respectively, and enhances task execution efficiency by 3.4x and 62.7x. Furthermore, we demonstrate that KARMA's plug-and-play capability allows for seamless deployment on real-world robotic systems, such as mobile manipulation platforms.Through this plug-and-play memory system, KARMA significantly enhances the ability of embodied agents to generate coherent and contextually appropriate plans, making the execution of complex household tasks more efficient. The experimental videos from the work can be found at https://youtu.be/4BT7fnw9ehs. Our code is available at https://github.com/WZX0Swarm0Robotics/KARMA/tree/master.
♻ ☆ GET: Unlocking the Multi-modal Potential of CLIP for Generalized Category Discovery CVPR 2025
Given unlabelled datasets containing both old and new categories, generalized category discovery (GCD) aims to accurately discover new classes while correctly classifying old classes. Current GCD methods only use a single visual modality of information, resulting in a poor classification of visually similar classes. As a different modality, text information can provide complementary discriminative information, which motivates us to introduce it into the GCD task. However, the lack of class names for unlabelled data makes it impractical to utilize text information. To tackle this challenging problem, in this paper, we propose a Text Embedding Synthesizer (TES) to generate pseudo text embeddings for unlabelled samples. Specifically, our TES leverages the property that CLIP can generate aligned vision-language features, converting visual embeddings into tokens of the CLIP's text encoder to generate pseudo text embeddings. Besides, we employ a dual-branch framework, through the joint learning and instance consistency of different modality branches, visual and semantic information mutually enhance each other, promoting the interaction and fusion of visual and text knowledge. Our method unlocks the multi-modal potentials of CLIP and outperforms the baseline methods by a large margin on all GCD benchmarks, achieving new state-of-the-art. Our code is available at: https://github.com/enguangW/GET.
comment: CVPR 2025
♻ ☆ A Comprehensive Survey of Time Series Forecasting: Architectural Diversity and Open Challenges
Time series forecasting is a critical task that provides key information for decision-making across various fields. Recently, various fundamental deep learning architectures such as MLPs, CNNs, RNNs, and GNNs have been developed and applied to solve time series forecasting problems. However, the structural limitations caused by the inductive biases of each deep learning architecture constrained their performance. Transformer models, which excel at handling long-term dependencies, have become significant architectural components for time series forecasting. However, recent research has shown that alternatives such as simple linear layers can outperform Transformers. These findings have opened up new possibilities for using diverse architectures. In this context of exploration into various models, the architectural modeling of time series forecasting has now entered a renaissance. This survey not only provides a historical context for time series forecasting but also offers comprehensive and timely analysis of the movement toward architectural diversification. By comparing and re-examining various deep learning models, we uncover new perspectives and presents the latest trends in time series forecasting, including the emergence of hybrid models, diffusion models, Mamba models, and foundation models. By focusing on the inherent characteristics of time series data, we also address open challenges that have gained attention in time series forecasting, such as channel dependency, distribution shift, causality, and feature extraction. This survey explores vital elements that can enhance forecasting performance through diverse approaches. These contributions lead to lowering the entry barriers for newcomers to the field of time series forecasting, while also offering seasoned researchers broad perspectives, new opportunities, and deep insights.
comment: Submitted to the Artificial Intelligence Review on October 10, 2024
♻ ☆ TCProF: Time-Complexity Prediction SSL Framework NAACL 2025
Time complexity is a theoretic measure to determine the amount of time the algorithm needs for its execution. In reality, developers write algorithms into code snippets within limited resources, making the calculation of a code's time complexity a fundamental task. However, determining the precise time complexity of a code is theoretically undecidable. In response, recent advancements have leaned toward deploying datasets for code time complexity prediction and initiating preliminary experiments for this challenge. We investigate the challenge in low-resource scenarios where only a few labeled instances are given for training. Remarkably, we are the first to introduce TCProF: a Time-Complexity Prediction SSL Framework as an effective solution for code time complexity prediction in low-resource settings. TCProF significantly boosts performance by integrating our augmentation, symbolic modules, and a co-training mechanism, achieving a more than 60% improvement over self-training approaches. We further provide an extensive comparative analysis between TCProF, ChatGPT, and Gemini-Pro, offering a detailed evaluation of our approach. Our code is at https://github.com/peer0/few-shot-tc.
comment: 26 pages, 13 figures, This paper has been accepted to NAACL 2025
♻ ☆ Moto: Latent Motion Token as the Bridging Language for Learning Robot Manipulation from Videos
Recent developments in Large Language Models pre-trained on extensive corpora have shown significant success in various natural language processing tasks with minimal fine-tuning. This success offers new promise for robotics, which has long been constrained by the high cost of action-labeled data. We ask: given the abundant video data containing interaction-related knowledge available as a rich "corpus", can a similar generative pre-training approach be effectively applied to enhance robot learning? The key challenge is to identify an effective representation for autoregressive pre-training that benefits robot manipulation tasks. Inspired by the way humans learn new skills through observing dynamic environments, we propose that effective robotic learning should emphasize motion-related knowledge, which is closely tied to low-level actions and is hardware-agnostic, facilitating the transfer of learned motions to actual robot actions. To this end, we introduce Moto, which converts video content into latent Motion Token sequences by a Latent Motion Tokenizer, learning a bridging "language" of motion from videos in an unsupervised manner. We pre-train Moto-GPT through motion token autoregression, enabling it to capture diverse visual motion knowledge. After pre-training, Moto-GPT demonstrates the promising ability to produce semantically interpretable motion tokens, predict plausible motion trajectories, and assess trajectory rationality through output likelihood. To transfer learned motion priors to real robot actions, we implement a co-fine-tuning strategy that seamlessly bridges latent motion token prediction and real robot control. Extensive experiments show that the fine-tuned Moto-GPT exhibits superior robustness and efficiency on robot manipulation benchmarks, underscoring its effectiveness in transferring knowledge from video data to downstream visual manipulation tasks.
comment: Project released at: https://chenyi99.github.io/moto/ Code released at: https://github.com/TencentARC/Moto Update: Added content related to real-world robot experiments and learning from human videos; Modified author information
♻ ☆ SPINE: Online Semantic Planning for Missions with Incomplete Natural Language Specifications in Unstructured Environments ICRA
As robots become increasingly capable, users will want to describe high-level missions and have robots infer the relevant details. Because pre-built maps are difficult to obtain in many realistic settings, accomplishing such missions will require the robot to map and plan online. While many semantic planning methods operate online, they are typically designed for well specified missions such as object search or exploration. Recently, Large Language Models (LLMs) have demonstrated powerful contextual reasoning abilities over a range of robotic tasks described in natural language. However, existing LLM-enabled planners typically do not consider online planning or complex missions; rather, relevant subtasks and semantics are provided by a pre-built map or a user. We address these limitations via SPINE, an online planner for missions with incomplete mission specifications provided in natural language. The planner uses an LLM to reason about subtasks implied by the mission specification and then realizes these subtasks in a receding horizon framework. Tasks are automatically validated for safety and refined online with new map observations. We evaluate SPINE in simulation and real-world settings with missions that require multiple steps of semantic reasoning and exploration in cluttered outdoor environments of over 20,000m$^2$. Compared to baselines that use existing LLM-enabled planning approaches, our method is over twice as efficient in terms of time and distance, requires less user interactions, and does not require a full map. Additional resources are provided at https://zacravichandran.github.io/SPINE.
comment: Accepted to the International Conference on Robotics and Automation (ICRA) 2025